├── .gitignore ├── README.md ├── data └── download_prepare_data.py ├── figures ├── Pipeline.png ├── Wave_Plot_N_Way_K_Shot.ipynb ├── features.png ├── query-left.png ├── signal.png ├── support-marvin-left-two.png ├── wave_plot_all.ipynb └── wave_plot_single.ipynb ├── protonets ├── __init__.py ├── data │ ├── FewShotSpeechData.py │ └── base.py ├── engine.py ├── models │ ├── __init__.py │ ├── encoder │ │ ├── GoogleKWS.py │ │ ├── TCResNet.py │ │ ├── __init__.py │ │ ├── baseUtil.py │ │ └── default.py │ ├── factory.py │ ├── few_shot.py │ └── utils.py └── utils │ ├── __init__.py │ ├── data.py │ ├── log.py │ └── model.py ├── requirements.txt ├── results ├── .ipynb_checkpoints │ └── Analyze Results-checkpoint.ipynb ├── Analyze Results.ipynb ├── C64.csv ├── Compare Results.ipynb ├── Compare_Core_Performance.ipynb ├── TCResNet8.csv ├── TCResNetDilated.csv ├── cnn_trad_fpool3.csv ├── create_table.ipynb ├── old │ ├── results_3CNN.csv │ ├── results_TCResNet8.csv │ ├── results_TCResNet8Dilated.csv │ ├── results_TCResNetDilated_1Stride.csv │ ├── results_cnn_trad_fpool3.csv │ └── results_cnn_trad_fpool3_simple.csv ├── performance-Core+Background.png ├── performance-Core+Unknown+Background+Silence.png ├── performance-Core+Unknown.png ├── performance-Core.png ├── results.tgn └── table.csv ├── scripts ├── debug_args.py ├── parse_results.py ├── plot.py ├── predict │ └── few_shot │ │ ├── eval.py │ │ ├── eval_results.py │ │ ├── qsub_eval.sh │ │ ├── run_eval.py │ │ └── submit.sh └── train │ └── few_shot │ ├── fewshotspeech │ ├── loop_qsub.sh │ ├── qsub_train.sh │ ├── train.sh │ └── train_fewshotspeech.sh │ ├── run_train.py │ ├── run_trainval.py │ ├── train.py │ └── trainval.py └── setup.py /.gitignore: -------------------------------------------------------------------------------- 1 | .ipynb_checkpoints 2 | *.pyc 3 | *.swp 4 | *.egg-info/ 5 | data/speech_commands 6 | scripts/train/few_shot/fewshotspeech/results 7 | -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 | Code Repository for the paper [Few-Shot Keyword Spotting with Prototypical Networks](https://arxiv.org/abs/2007.14463). 2 | 3 | ![Few-Shot Keyword Spotting Pipeline](figures/Pipeline.png) 4 | 5 | 6 | # Installation 7 | 8 | 1. Clone the repository: 9 | 10 | ``` 11 | git clone https://github.com/ArchitParnami/Few-Shot-KWS 12 | ``` 13 | 14 | 2. Create a conda environment: 15 | 16 | ``` 17 | conda create -n FS-KWS 18 | ``` 19 | 20 | 3. If pip not installed, install pip by: 21 | 22 | ``` 23 | conda install pip 24 | ``` 25 | 26 | 4. Install the required packages: 27 | 28 | ``` 29 | pip install -r requirements.text 30 | ``` 31 | 32 | 5. Install the protonets package: 33 | 34 | ``` 35 | cd Few-Shot-KWS 36 | python setup.py develop 37 | ``` 38 | 39 | # Download & Prepare Few-Shot Keyword Spotting Dataset 40 | 41 | ``` 42 | cd Few-Shot-KWS/data/ 43 | python download_prepare_data.py 44 | ``` 45 | # Train 46 | 47 | To train a simple 2-way 1-shot experiment. 48 | 49 | ``` 50 | cd Few-Shot-KWS/scripts/train/few-shot/fewshotspeech 51 | ./train.sh 2 1 0 mymodel 52 | ``` 53 | Specify arguments to train.sh in the following manner 54 | 55 | train.sh num_ways num_shots exp_type exp_id 56 | 57 | - num_ways 58 | + Number of classes 59 | + Eg. 2 or 4 60 | - num_shots 61 | + Number of samples per class. 62 | + Eg. 1,5 63 | - exp_type 64 | + Number indicating the type of experimental setup 65 | - 0 = Simple N-Way K-Shot Setup. No background, silence or unknown keywords. 66 | - 1 = Include Background 67 | - 2 = Include Silence 68 | - 3 = Include Unknown 69 | - 4 = Background + Silence 70 | - 5 = Background + Unkown 71 | - 6 = Unknown + Silence 72 | - 7 = Background + Silence + Unknown 73 | - exp_id 74 | + identifier = directory name 75 | + results are saved in `Few-Shot-KWS/scripts/train/few-shot/fewshotspeech/results/[exp_id]` 76 | 77 | 78 | # Evaluate 79 | 80 | ``` 81 | cd Few-Shot-KWS/scripts/predict/few-shot 82 | python eval_results.py ../../train/few_shot/fewshotspeech/results/ 83 | ``` 84 | 85 | The evaluation can be found in: 86 | ``` 87 | cat Few-Shot-KWS/scripts/train/few-shot/fewshotspeech/results/[exp-id]/[timestamp]/eval.txt 88 | ``` 89 | 90 | # Results 91 | 92 | Comaring test accuracy of different embedding networks on 4-way FS-KWS as we increase the number of support examples. The results are presented for four different cases discussed in the paper. 93 | 94 | ![](results/performance-Core.png) ![](results/performance-Core+Background.png) 95 | 96 | ![](results/performance-Core+Unknown.png) ![](results/performance-Core+Unknown+Background+Silence.png) 97 | 98 | 99 | # References 100 | 101 | The code in this repository has been adapted from: 102 | 103 | 1. https://github.com/jakesnell/prototypical-networks 104 | 2. https://github.com/hyperconnect/TC-ResNet 105 | 3. https://github.com/tensorflow/docs/blob/master/site/en/r1/tutorials/sequences/audio_recognition.md 106 | 107 | 108 | # Citation 109 | ``` 110 | @misc{parnami2020fewshot, 111 | title={Few-Shot Keyword Spotting With Prototypical Networks}, 112 | author={Archit Parnami and Minwoo Lee}, 113 | year={2020}, 114 | eprint={2007.14463}, 115 | archivePrefix={arXiv}, 116 | primaryClass={eess.AS} 117 | } 118 | ``` 119 | -------------------------------------------------------------------------------- /data/download_prepare_data.py: -------------------------------------------------------------------------------- 1 | import os 2 | import sys 3 | import urllib 4 | import tarfile 5 | import torchaudio 6 | import glob 7 | import random 8 | import shutil 9 | 10 | 11 | class FewShotSpeechDataDownloader(object): 12 | def __init__(self, data_url, dest_directory, sample_rate, 13 | clip_duration_ms, speaker_limit, core_split, 14 | unknown_split): 15 | self.data_url = data_url 16 | self.dataset_path = dest_directory 17 | self.desired_samples = int(sample_rate * clip_duration_ms / 1000) 18 | self.keywords = None 19 | self.speaker_limit = speaker_limit 20 | self.core_words_path = os.path.join(self.dataset_path, 'core_words.txt') 21 | self.other_words_path = os.path.join(self.dataset_path, 'other_words.txt') 22 | self.core_dataset_path = os.path.join(self.dataset_path, 'core') 23 | self.unknown_dataset_path = os.path.join(self.dataset_path, '_unknown_') 24 | self.num_core_speakers = None 25 | self.num_unknown_speakers = None 26 | self.core_split = core_split 27 | self.unknown_split = unknown_split 28 | 29 | def get_keywords(self): 30 | if self.keywords is None: 31 | keywords = [dir_name for dir_name in os.listdir(self.dataset_path) 32 | if os.path.isdir(os.path.join(self.dataset_path, dir_name)) and 33 | not dir_name.startswith('_')] 34 | self.keywords = keywords 35 | 36 | return self.keywords 37 | 38 | 39 | def load_audio(self, audio_file, desired_samples): 40 | sound, sr = torchaudio.load(filepath=audio_file, normalization=True, 41 | num_frames=desired_samples) 42 | return sound, sr 43 | 44 | def maybe_download_and_extract_dataset(self): 45 | if not self.data_url: 46 | return 47 | 48 | if os.path.exists(self.dataset_path): 49 | print('Found the dataset directory {}.'.format(self.dataset_path)) 50 | else: 51 | os.makedirs(self.dataset_path) 52 | 53 | filename = self.data_url.split('/')[-1] 54 | filepath = os.path.join(self.dataset_path, filename) 55 | 56 | if os.path.exists(filepath): 57 | print('Found the dataset {}. Skipping download'.format(filepath)) 58 | print('Resetting the state by removing old files.') 59 | for item in os.listdir(self.dataset_path): 60 | itempath = os.path.join(self.dataset_path, item) 61 | if itempath != filepath: 62 | if os.path.isdir(itempath): 63 | shutil.rmtree(itempath) 64 | else: 65 | os.unlink(itempath) 66 | else: 67 | 68 | def _progress(count, block_size, total_size): 69 | sys.stdout.write( 70 | '\r--> Downloading %s %.1f%%' % 71 | (filename, float(count * block_size) / float(total_size) * 100.0)) 72 | sys.stdout.flush() 73 | 74 | try: 75 | filepath, _ = urllib.request.urlretrieve(data_url, filepath, _progress) 76 | except: 77 | print( 78 | 'Failed to download URL: %s to folder: %s', self.data_url, filepath) 79 | print( 80 | 'Please make sure you have enough free space and' 81 | ' an internet connection') 82 | raise 83 | 84 | print() 85 | statinfo = os.stat(filepath) 86 | print('Successfully downloaded %s (%d bytes)', 87 | filename, statinfo.st_size) 88 | 89 | 90 | print('Extracting the dataset....') 91 | tarfile.open(filepath, 'r:gz').extractall(self.dataset_path) 92 | 93 | def delete_smaller_files(self): 94 | print("Scanning files..") 95 | keywords = self.get_keywords() 96 | count = 0 97 | total = 0 98 | for keyword in keywords: 99 | keyword_wavs = glob.glob(os.path.join(self.dataset_path, keyword, '*.wav')) 100 | total += len(keyword_wavs) 101 | for wav_file in sorted(keyword_wavs): 102 | sound, sr = self.load_audio(wav_file, self.desired_samples) 103 | if sound.shape[1] != self.desired_samples: 104 | os.unlink(wav_file) 105 | count += 1 106 | print("{} samples with audio < 1 second removed".format(count)) 107 | print("{} samples retained".format(total-count)) 108 | 109 | def group_by_author(self): 110 | print("Grouping audio samples by speakers ..") 111 | path = self.dataset_path 112 | items = os.listdir(path) 113 | for item in items: 114 | item_path = os.path.join(path, item) 115 | if os.path.isdir(item_path): 116 | wav_files = os.listdir(item_path) 117 | ids = {} 118 | for wav_file in wav_files: 119 | pos = wav_file.find('_nohash_') 120 | if pos != -1: 121 | id = wav_file[:pos] 122 | if id not in ids: 123 | ids[id] = [] 124 | ids[id].append(wav_file) 125 | for id in ids: 126 | dir_path = os.path.join(item_path, id) 127 | os.makedirs(dir_path) 128 | for wav_file in ids[id]: 129 | wav_file_path = os.path.join(item_path, wav_file) 130 | new_wave_file_path = os.path.join(dir_path, wav_file) 131 | os.rename(wav_file_path, new_wave_file_path) 132 | 133 | def keyword_stats(self): 134 | stats = [] 135 | for word in self.keywords: 136 | word_path = os.path.join(self.dataset_path, word) 137 | speakers = os.listdir(word_path) 138 | speaker_samples = [] 139 | for speaker in speakers: 140 | speaker_path = os.path.join(word_path, speaker) 141 | samples = os.listdir(speaker_path) 142 | speaker_samples.append(len(samples)) 143 | 144 | num_speakers = len(speakers) 145 | min_samples = min(speaker_samples) 146 | max_samples = max(speaker_samples) 147 | mean_samples = sum(speaker_samples) / num_speakers 148 | stats.append((word, num_speakers, min_samples, max_samples, mean_samples)) 149 | return stats 150 | 151 | def analyze_data(self): 152 | stats = self.keyword_stats() 153 | stats.sort(key=lambda stat: stat[1], reverse=True) 154 | print() 155 | print('=' * 30) 156 | header = '{:8s} {:7s} {:3s} {:3s} {:3s}'.format("Word", "Speakers", "Min", "Max", "Mean") 157 | print(header) 158 | print('=' * 30) 159 | print() 160 | 161 | 162 | core = []; other = [] 163 | min_core = float('inf') 164 | min_other = float('inf') 165 | 166 | for stat in stats: 167 | word, num_speakers, min_samples, max_samples, mean_samples = stat 168 | if num_speakers >= self.speaker_limit: 169 | core.append(word) 170 | if num_speakers < min_core: 171 | min_core = num_speakers 172 | else: 173 | other.append(word) 174 | if num_speakers < min_other: 175 | min_other = num_speakers 176 | 177 | out = '{:8s} {:7d} {:3d} {:3d} {:.2f}'.format(word, num_speakers, min_samples, max_samples, mean_samples) 178 | print(out) 179 | 180 | print() 181 | with open(self.core_words_path, 'w') as wf: 182 | wf.writelines([word + '\n' for word in core]) 183 | #print('Words with speakers >= {} saved to file {}'.format(speaker_limit, self.core_words_path)) 184 | #print('Min core word speakers = ', min_core) 185 | self.num_core_speakers = min_core 186 | 187 | with open(self.other_words_path, 'w') as wf: 188 | wf.writelines([word + '\n' for word in other]) 189 | #print('Other saved to {}'.format(self.other_words_path)) 190 | #print('Min other word speakers = ', min_other) 191 | self.num_unknown_speakers = min_other 192 | 193 | print('Core Words ({}): {}\n'.format(len(core), core)) 194 | print('Unknown Words ({}): {}'.format(len(other), other)) 195 | 196 | 197 | def filter_dataset(self, path, new_path, words, num_speakers): 198 | os.makedirs(new_path) 199 | for word in words: 200 | item_path = os.path.join(path, word) 201 | new_item_path = os.path.join(new_path, word) 202 | os.makedirs(new_item_path) 203 | speakers = os.listdir(item_path) 204 | random.shuffle(speakers) 205 | speakers = speakers[:num_speakers] 206 | for speaker in speakers: 207 | speaker_path = os.path.join(item_path, speaker) 208 | wav_files = os.listdir(speaker_path) 209 | random_wave_file = wav_files[random.randrange(len(wav_files))] 210 | shutil.move(os.path.join(speaker_path, random_wave_file), 211 | os.path.join(new_item_path, speaker + '.wav')) 212 | shutil.rmtree(item_path) 213 | 214 | def uniform_data(self): 215 | print('Making data uniform with respect to number of samples per keyword..') 216 | with open(self.core_words_path, 'r') as rf: 217 | core_words = [line.strip('\n') for line in rf.readlines()] 218 | with open(self.other_words_path, 'r') as rf: 219 | unknown_words = [line.strip('\n') for line in rf.readlines()] 220 | self.filter_dataset(self.dataset_path, self.core_dataset_path, 221 | core_words, self.num_core_speakers) 222 | self.filter_dataset(self.dataset_path, self.unknown_dataset_path, 223 | unknown_words, self.num_unknown_speakers) 224 | 225 | def write_file(self, path, classes): 226 | with open(path, 'w') as wf: 227 | for word in classes: 228 | wf.write(word + '\n') 229 | 230 | def write_splits(self, train, val, test, dataset_path): 231 | self.write_file(os.path.join(dataset_path, 'train.txt'), train) 232 | self.write_file(os.path.join(dataset_path, 'test.txt'), test) 233 | self.write_file(os.path.join(dataset_path, 'val.txt'), val) 234 | 235 | def generate_core_split(self): 236 | classes = os.listdir(self.core_dataset_path) 237 | random.shuffle(classes) 238 | tr, vl, te = self.core_split 239 | train = classes[:-(vl+te)] 240 | testval = classes[-(vl+te):] 241 | test = testval[:te] 242 | val = testval[te:] 243 | self.write_splits(train, val, test, self.core_dataset_path) 244 | 245 | def generate_unknown_split(self): 246 | words = sorted([word for word in os.listdir(self.unknown_dataset_path) 247 | if os.path.isdir(os.path.join(self.unknown_dataset_path, word))]) 248 | d = {} 249 | for word in words: 250 | d[word] = sorted(os.listdir(os.path.join(self.unknown_dataset_path, word))) 251 | random.shuffle(d[word]) 252 | d[word] = d[word][:self.num_unknown_speakers] 253 | tr, vl , te = self.unknown_split 254 | train_count = int(tr * self.num_unknown_speakers / 100) 255 | test_count = int(te * self.num_unknown_speakers / 100) 256 | val_count = self.num_unknown_speakers - (train_count + test_count) 257 | train = []; test = []; val = [] 258 | for i in range(self.num_unknown_speakers): 259 | for word in words: 260 | if i < train_count: 261 | train.append(word + '/' + d[word][i]) 262 | elif i < train_count + val_count: 263 | val.append(word + '/' + d[word][i]) 264 | else: 265 | test.append(word + '/' + d[word][i]) 266 | self.write_splits(train, val, test, self.unknown_dataset_path) 267 | 268 | def generate_splits(self): 269 | print('Generating train, val and test splits') 270 | self.generate_core_split() 271 | self.generate_unknown_split() 272 | 273 | def remove_unwanted_files(self): 274 | #print('Cleaning up.') 275 | files = ['LICENSE', 'README.md', 'testing_list.txt', 276 | 'validation_list.txt', 'core_words.txt', 'other_words.txt'] 277 | for file_name in files: 278 | file_path = os.path.join(self.dataset_path, file_name) 279 | if os.path.exists(file_path): 280 | os.unlink(file_path) 281 | #print("You may delete the tar file now.") 282 | 283 | def run(self): 284 | 285 | print("\n>> Downloading Google Speech Command Dataset") 286 | 287 | self.maybe_download_and_extract_dataset() 288 | 289 | print("\n>> Preparing Few-Shot Speech Command Dataset") 290 | 291 | print("\n--> 1/6 Filtering") 292 | self.delete_smaller_files() 293 | 294 | print("\n--> 2/6 Grouping") 295 | self.group_by_author() 296 | 297 | print("\n--> 3/6 Analyzing") 298 | self.analyze_data() 299 | 300 | print("\n--> 4/6 Balancing") 301 | self.uniform_data() 302 | 303 | print("\n--> 5/6 Splitting") 304 | self.generate_splits() 305 | 306 | print("\n--> 6/6 Cleaning up") 307 | self.remove_unwanted_files() 308 | print('\nAll done.') 309 | 310 | 311 | if __name__ == '__main__': 312 | data_url = 'http://storage.googleapis.com/download.tensorflow.org/data/speech_commands_v0.02.tar.gz' 313 | dataset_path = os.path.join(os.curdir, 'speech_commands') 314 | sample_rate = 16000 315 | clip_duration_ms = 1000 316 | speaker_limit = 1000 317 | core_split = (24,5,5) # (train, val, test) 318 | unknown_split = (60, 20, 20) # in percentage 319 | random.seed(42) 320 | downloader = FewShotSpeechDataDownloader(data_url, dataset_path, 321 | sample_rate, clip_duration_ms, speaker_limit, core_split, unknown_split) 322 | downloader.run() 323 | 324 | 325 | -------------------------------------------------------------------------------- /figures/Pipeline.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ArchitParnami/Few-Shot-KWS/04823543110be9d186b3a5ecbc695548a6cec097/figures/Pipeline.png -------------------------------------------------------------------------------- /figures/features.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ArchitParnami/Few-Shot-KWS/04823543110be9d186b3a5ecbc695548a6cec097/figures/features.png -------------------------------------------------------------------------------- /figures/query-left.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ArchitParnami/Few-Shot-KWS/04823543110be9d186b3a5ecbc695548a6cec097/figures/query-left.png -------------------------------------------------------------------------------- /figures/signal.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ArchitParnami/Few-Shot-KWS/04823543110be9d186b3a5ecbc695548a6cec097/figures/signal.png -------------------------------------------------------------------------------- /figures/support-marvin-left-two.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ArchitParnami/Few-Shot-KWS/04823543110be9d186b3a5ecbc695548a6cec097/figures/support-marvin-left-two.png -------------------------------------------------------------------------------- /protonets/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ArchitParnami/Few-Shot-KWS/04823543110be9d186b3a5ecbc695548a6cec097/protonets/__init__.py -------------------------------------------------------------------------------- /protonets/data/FewShotSpeechData.py: -------------------------------------------------------------------------------- 1 | import os 2 | from functools import partial 3 | import torch 4 | import numpy as np 5 | import glob 6 | 7 | from torch.utils.data import Dataset 8 | from torchvision import transforms 9 | from torchnet.transform import compose 10 | from torchnet.dataset import ListDataset, TransformDataset 11 | import torchaudio 12 | import torch.nn.functional as F 13 | from protonets.data.base import convert_dict, CudaTransform 14 | 15 | 16 | class FewShotSpeechDataset(TransformDataset): 17 | 18 | def __init__(self, data_dir, class_file, n_support, n_query, cuda, args): 19 | self.sample_rate = args['sample_rate'] 20 | self.clip_duration_ms = args['clip_duration'] 21 | self.window_size_ms = args['window_size'] 22 | self.window_stride_ms = args['window_stride'] 23 | self.feature_bin_count = args['num_features'] 24 | self.foreground_volume = args['foreground_volume'] 25 | self.time_shift_ms = args['time_shift'] 26 | self.use_background = args['include_background'] 27 | self.background_volume = args['bg_volume'] 28 | self.background_frequency= args['bg_frequency'] 29 | self.desired_samples = int(self.sample_rate * self.clip_duration_ms / 1000) 30 | self.silence = args['include_silence'] 31 | self.silence_num_samples = args['num_silence'] 32 | self.unknown = args['include_unknown'] 33 | self.data_cache = {} 34 | self.data_dir = data_dir 35 | self.class_file = class_file 36 | self.n_support = n_support 37 | self.n_query = n_query 38 | self.background_data = self.load_background_data() 39 | self.mfcc = self.build_mfcc_extractor() 40 | self.transforms = [partial(convert_dict, 'class'), 41 | self.load_class_samples, 42 | self.extract_episode] 43 | if cuda: 44 | self.transforms.append(CudaTransform()) 45 | self.class_names = self.read() 46 | transforms = compose(self.transforms) 47 | super().__init__(ListDataset(self.class_names), transforms) 48 | 49 | def load_background_data(self): 50 | background_path = os.path.join(self.data_dir, '..' , '_background_noise_', '*.wav') 51 | background_data = [] 52 | if self.use_background or self.silence: 53 | for wav_path in glob.glob(background_path): 54 | bg_sound, bg_sr = torchaudio.load(wav_path) 55 | background_data.append(bg_sound.flatten()) 56 | return background_data 57 | 58 | def build_mfcc_extractor(self): 59 | frame_len = self.window_size_ms / 1000 60 | stride = self.window_stride_ms / 1000 61 | mfcc = torchaudio.transforms.MFCC(self.sample_rate, 62 | n_mfcc=self.feature_bin_count, 63 | melkwargs={ 64 | 'hop_length' : int(stride*self.sample_rate), 65 | 'n_fft' : int(frame_len*self.sample_rate)}) 66 | return mfcc 67 | 68 | def read(self): 69 | class_names = [] 70 | with open(self.class_file, 'r') as f: 71 | class_names = list(map(lambda x: x.rstrip('\n'), f.readlines())) 72 | if self.silence: 73 | class_names.append('_silence_') 74 | if self.unknown: 75 | class_names.append('_unknown_') 76 | return class_names 77 | 78 | def load_audio(self, key, out_field, d): 79 | sound, _ = torchaudio.load(filepath=d[key], normalization=True, 80 | num_frames=self.desired_samples) 81 | d[out_field] = sound 82 | return d 83 | 84 | def adjust_volume(self, key, d): 85 | d[key] = d[key] * self.foreground_volume 86 | return d 87 | 88 | def shift_and_pad(self, key, d): 89 | audio = d[key] 90 | time_shift = int((self.time_shift_ms * self.sample_rate) / 1000) 91 | if time_shift > 0: 92 | time_shift_amount = np.random.randint(-time_shift, time_shift) 93 | else: 94 | time_shift_amount = 0 95 | 96 | if time_shift_amount > 0: 97 | time_shift_padding = (time_shift_amount, 0) 98 | time_shift_offset = 0 99 | else: 100 | time_shift_padding = (0, -time_shift_amount) 101 | time_shift_offset = -time_shift_amount 102 | 103 | padded_foreground = F.pad(audio, time_shift_padding, 'constant', 0) 104 | sliced_foreground = torch.narrow(padded_foreground, 1, time_shift_offset, self.desired_samples) 105 | d[key] = sliced_foreground 106 | return d 107 | 108 | def mix_background(self, use_background, k, d): 109 | foreground = d[k] 110 | if use_background: 111 | background_index = np.random.randint(len(self.background_data)) 112 | background_samples = self.background_data[background_index] 113 | if len(background_samples) <= self.desired_samples: 114 | raise ValueError( 115 | 'Background sample is too short! Need more than %d' 116 | ' samples but only %d were found' % 117 | (self.desired_samples, len(background_samples))) 118 | background_offset = np.random.randint( 119 | 0, len(background_samples) - self.desired_samples) 120 | background_clipped = background_samples[background_offset:( 121 | background_offset + self.desired_samples)] 122 | background_reshaped = background_clipped.reshape([1, self.desired_samples]) 123 | 124 | if np.random.uniform(0, 1) < self.background_frequency: 125 | bg_vol = np.random.uniform(0, self.background_volume) 126 | else: 127 | bg_vol = 0 128 | else: 129 | background_reshaped = torch.zeros(1, self.desired_samples) 130 | bg_vol = 0 131 | 132 | background_mul = background_reshaped * bg_vol 133 | background_add = background_mul + foreground 134 | background_clamped = torch.clamp(background_add, -1.0, 1.0) 135 | d[k] = background_clamped 136 | return d 137 | 138 | def extract_features(self, k, d): 139 | features = self.mfcc(d[k])[0] # just one channel 140 | features = features.T # f x t -> t x f 141 | d[k] = torch.unsqueeze(features,0) 142 | return d 143 | 144 | def load_class_samples(self, d): 145 | if d['class'] not in self.data_cache: 146 | if d['class'] == '_silence_': 147 | samples = torch.zeros(self.silence_num_samples, 1, self.desired_samples) 148 | sample_ds = TransformDataset(ListDataset(samples), 149 | compose([ 150 | partial(convert_dict, 'data'), 151 | partial(self.mix_background, True, 'data'), 152 | partial(self.extract_features, 'data') 153 | ])) 154 | 155 | else: 156 | samples = [] 157 | 158 | if d['class'] == '_unknown_': 159 | unknown_dir = os.path.join(self.data_dir, '..', '_unknown_') 160 | split = os.path.basename(self.class_file) 161 | unknown_wavs = os.path.join(unknown_dir, split) 162 | with open(unknown_wavs, 'r') as rf: 163 | samples = [os.path.join(unknown_dir, wav_file.strip('\n')) for wav_file in rf.readlines()] 164 | else: 165 | keyword_dir = os.path.join(self.data_dir, d['class']) 166 | samples = glob.glob(os.path.join(keyword_dir, '*.wav')) 167 | 168 | if len(samples) == 0: 169 | raise Exception("No Samples found for GoogleSpeechCommand {} at {}".format(d['class'], keyword_dir)) 170 | 171 | sample_ds = TransformDataset(ListDataset(samples), 172 | compose([ 173 | partial(convert_dict, 'file_name'), 174 | partial(self.load_audio, 'file_name', 'data'), 175 | partial(self.adjust_volume, 'data'), 176 | partial(self.shift_and_pad, 'data'), 177 | partial(self.mix_background, self.use_background,'data'), 178 | partial(self.extract_features, 'data') 179 | ])) 180 | 181 | loader = torch.utils.data.DataLoader(sample_ds, batch_size=len(sample_ds), shuffle=False) 182 | 183 | for sample in loader: 184 | self.data_cache[d['class']] = sample['data'] 185 | break # only need one sample because batch size equal to dataset length 186 | 187 | return { 'class': d['class'], 'data': self.data_cache[d['class']] } 188 | 189 | def extract_episode(self, d): 190 | # data: N x C x H x W 191 | n_examples = d['data'].size(0) 192 | 193 | if self.n_query == -1: 194 | self.n_query = n_examples - self.n_support 195 | 196 | if d['class'] == '_unknown_': 197 | start_index = torch.randint(0, n_examples - (self.n_support + self.n_query), (1,)).data[0] 198 | example_inds = torch.arange(start_index, start_index + (self.n_support + self.n_query)) 199 | else: 200 | example_inds = torch.randperm(n_examples)[:(self.n_support + self.n_query)] 201 | 202 | support_inds = example_inds[:self.n_support] 203 | query_inds = example_inds[self.n_support:] 204 | 205 | xs = d['data'][support_inds] 206 | xq = d['data'][query_inds] 207 | 208 | return { 209 | 'class': d['class'], 210 | 'xs': xs, 211 | 'xq': xq 212 | } 213 | 214 | 215 | 216 | 217 | 218 | 219 | 220 | -------------------------------------------------------------------------------- /protonets/data/base.py: -------------------------------------------------------------------------------- 1 | import torch 2 | 3 | def convert_dict(k, v): 4 | return { k: v } 5 | 6 | class CudaTransform(object): 7 | def __init__(self): 8 | pass 9 | 10 | def __call__(self, data): 11 | for k,v in data.items(): 12 | if hasattr(v, 'cuda'): 13 | data[k] = v.cuda() 14 | 15 | return data 16 | 17 | class SequentialBatchSampler(object): 18 | def __init__(self, n_classes): 19 | self.n_classes = n_classes 20 | 21 | def __len__(self): 22 | return self.n_classes 23 | 24 | def __iter__(self): 25 | for i in range(self.n_classes): 26 | yield torch.LongTensor([i]) 27 | 28 | class EpisodicBatchSampler(object): 29 | def __init__(self, n_classes, n_way, n_episodes): 30 | self.n_classes = n_classes 31 | self.n_way = n_way 32 | self.n_episodes = n_episodes 33 | 34 | def __len__(self): 35 | return self.n_episodes 36 | 37 | def __iter__(self): 38 | for i in range(self.n_episodes): 39 | yield torch.randperm(self.n_classes)[:self.n_way] 40 | 41 | class EpisodicSpeechBatchSampler(object): 42 | def __init__(self, n_classes, n_way, n_episodes, 43 | include_silence=False, include_unknown=False): 44 | self.n_classes = n_classes 45 | self.n_episodes = n_episodes 46 | self.n_way = n_way 47 | self.include_silence = include_silence 48 | self.include_unknown = include_unknown 49 | self.skip = 0 50 | if include_silence: 51 | self.skip += 1 52 | if include_unknown: 53 | self.skip += 1 54 | 55 | def __len__(self): 56 | return self.n_episodes 57 | 58 | def __iter__(self): 59 | for i in range(self.n_episodes): 60 | selected = torch.randperm(self.n_classes - self.skip)[:self.n_way] 61 | 62 | if self.include_silence: 63 | silence_class = torch.tensor([self.n_classes - 2]) 64 | selected = torch.cat((selected, silence_class)) 65 | if self.include_unknown: 66 | unknown_class = torch.tensor([self.n_classes - 1]) 67 | selected = torch.cat((selected, unknown_class)) 68 | 69 | yield selected[torch.randperm(self.n_way + self.skip)] 70 | -------------------------------------------------------------------------------- /protonets/engine.py: -------------------------------------------------------------------------------- 1 | from tqdm import tqdm 2 | 3 | class Engine(object): 4 | def __init__(self): 5 | hook_names = ['on_start', 'on_start_epoch', 'on_sample', 'on_forward', 6 | 'on_backward', 'on_end_epoch', 'on_update', 'on_end'] 7 | 8 | self.hooks = { } 9 | for hook_name in hook_names: 10 | self.hooks[hook_name] = lambda state: None 11 | 12 | def train(self, **kwargs): 13 | state = { 14 | 'model': kwargs['model'], 15 | 'loader': kwargs['loader'], 16 | 'optim_method': kwargs['optim_method'], 17 | 'optim_config': kwargs['optim_config'], 18 | 'max_epoch': kwargs['max_epoch'], 19 | 'epoch': 0, # epochs done so far 20 | 't': 0, # samples seen so far 21 | 'batch': 0, # samples seen in current epoch 22 | 'stop': False 23 | } 24 | 25 | state['optimizer'] = state['optim_method'](state['model'].parameters(), **state['optim_config']) 26 | 27 | self.hooks['on_start'](state) 28 | while state['epoch'] < state['max_epoch'] and not state['stop']: 29 | state['model'].train() 30 | 31 | self.hooks['on_start_epoch'](state) 32 | 33 | state['epoch_size'] = len(state['loader']) 34 | 35 | for sample in tqdm(state['loader'], desc="Epoch {:d} train".format(state['epoch'] + 1)): 36 | state['sample'] = sample 37 | self.hooks['on_sample'](state) 38 | 39 | state['optimizer'].zero_grad() 40 | loss, state['output'] = state['model'].loss(state['sample']) 41 | self.hooks['on_forward'](state) 42 | 43 | loss.backward() 44 | self.hooks['on_backward'](state) 45 | 46 | state['optimizer'].step() 47 | 48 | state['t'] += 1 49 | state['batch'] += 1 50 | self.hooks['on_update'](state) 51 | 52 | state['epoch'] += 1 53 | state['batch'] = 0 54 | self.hooks['on_end_epoch'](state) 55 | 56 | self.hooks['on_end'](state) 57 | -------------------------------------------------------------------------------- /protonets/models/__init__.py: -------------------------------------------------------------------------------- 1 | from protonets.models.factory import get_model, register_model 2 | 3 | import protonets.models.few_shot 4 | -------------------------------------------------------------------------------- /protonets/models/encoder/GoogleKWS.py: -------------------------------------------------------------------------------- 1 | import torch 2 | import torch.nn as nn 3 | from protonets.models.encoder.baseUtil import Flatten 4 | 5 | class cnn_trad_fpool3(nn.Module): 6 | def __init__(self, in_channels, hid_channels, out_channels): 7 | super(cnn_trad_fpool3, self).__init__() 8 | conv1 = nn.Sequential( 9 | nn.Conv2d(in_channels=in_channels, out_channels=hid_channels, 10 | kernel_size=(20,8),stride=1), 11 | nn.BatchNorm2d(out_channels), 12 | nn.ReLU(), 13 | nn.MaxPool2d(kernel_size=(1,3))) 14 | 15 | conv2 = nn.Sequential( 16 | nn.Conv2d(in_channels=hid_channels, out_channels=hid_channels, 17 | kernel_size=(10,4), stride=1), 18 | nn.BatchNorm2d(out_channels), 19 | nn.ReLU()) 20 | 21 | lin = nn.Sequential( 22 | nn.Linear(11776, 32), 23 | nn.ReLU() 24 | ) 25 | 26 | dnn = nn.Sequential( 27 | nn.Linear(32, 128), 28 | nn.Sigmoid() 29 | ) 30 | 31 | self.encoder = nn.Sequential( 32 | conv1, 33 | conv2, 34 | Flatten(), 35 | lin, 36 | dnn 37 | ) 38 | 39 | def forward(self, x): 40 | out = self.encoder(x) 41 | return out 42 | 43 | 44 | class cnn_trad_fpool3_simple(nn.Module): 45 | def __init__(self, in_channels, hid_channels, out_channels): 46 | super(cnn_trad_fpool3_simple, self).__init__() 47 | conv1 = nn.Sequential( 48 | nn.Conv2d(in_channels=in_channels, out_channels=hid_channels, 49 | kernel_size=(20,8),stride=1), 50 | nn.BatchNorm2d(out_channels), 51 | nn.ReLU(), 52 | nn.MaxPool2d(kernel_size=(1,3))) 53 | 54 | conv2 = nn.Sequential( 55 | nn.Conv2d(in_channels=hid_channels, out_channels=hid_channels, 56 | kernel_size=(10,4), stride=1), 57 | nn.BatchNorm2d(out_channels), 58 | nn.ReLU()) 59 | 60 | self.encoder = nn.Sequential( 61 | conv1, 62 | conv2, 63 | Flatten(), 64 | ) 65 | 66 | def forward(self, x): 67 | out = self.encoder(x) 68 | return out -------------------------------------------------------------------------------- /protonets/models/encoder/TCResNet.py: -------------------------------------------------------------------------------- 1 | import torch 2 | import torch.nn as nn 3 | import torch.nn.functional as F 4 | from protonets.models.encoder.baseUtil import Flatten, get_padding 5 | from collections import OrderedDict 6 | 7 | 8 | class TC(nn.Module): 9 | def __init__(self): 10 | super(TC, self).__init__() 11 | 12 | def forward(self, x): 13 | return torch.transpose(x, 1, 3) 14 | 15 | 16 | class ResidualBlock(nn.Module): 17 | def __init__(self, in_channels, in_height, in_width, out_channels, kernel, dilation, stride=2): 18 | super(ResidualBlock, self).__init__() 19 | self.k_size = kernel 20 | self.stride2 = 1 21 | self.k_1D = (1,1) 22 | 23 | if in_channels != out_channels: 24 | self.stride1 = stride 25 | self.upsample = True 26 | else: 27 | self.stride1 = 1 28 | self.upsample = False 29 | 30 | self.conv1 = nn.Conv2d(in_channels, out_channels, self.k_size, stride=self.stride1, 31 | bias=False, padding=get_padding(in_height, in_width, self.k_size[0], 32 | self.k_size[1], self.stride1,d_h=dilation), dilation=dilation) 33 | self.bn1 = nn.BatchNorm2d(out_channels) 34 | self.relu1 = nn.ReLU() 35 | 36 | self.conv2 = nn.Conv2d(out_channels, out_channels, self.k_size, stride=self.stride2, 37 | bias=False, padding=get_padding(in_height, in_width, self.k_size[0], 38 | self.k_size[1], self.stride2, d_h=dilation), dilation=dilation) 39 | self.bn2 = nn.BatchNorm2d(out_channels) 40 | 41 | self.conv3 = nn.Conv2d(in_channels, out_channels, self.k_1D, stride=self.stride1, 42 | bias=False, padding=get_padding(in_height, in_width, self.k_1D[0], 43 | self.k_1D[1], self.stride1)) 44 | self.bn3 = nn.BatchNorm2d(out_channels) 45 | self.relu3 = nn.ReLU() 46 | 47 | self.encoder = nn.Sequential( 48 | self.conv1, self.bn1, self.relu1, 49 | self.conv2, self.bn2 50 | ) 51 | 52 | self.upsampler = nn.Sequential( 53 | self.conv3, self.bn3, self.relu3 54 | ) 55 | 56 | self.relu = nn.ReLU() 57 | 58 | 59 | def forward(self, x): 60 | a = self.encoder(x) 61 | if self.upsample: 62 | b = self.upsampler(x) 63 | else: 64 | b = x 65 | out = a + b 66 | out = self.relu(out) 67 | return out 68 | 69 | 70 | class TCResNet(nn.Module): 71 | def __init__(self, in_channels, in_height, in_width, n_blocks, n_channels, 72 | conv_kernel, res_kernel, dilation, stride=2): 73 | super(TCResNet, self).__init__() 74 | self.conv_k_size = conv_kernel 75 | self.conv_stride = 1 76 | self.conv_channels = n_channels[0] 77 | 78 | self.tc = TC() 79 | self.conv1 = nn.Conv2d(in_channels, self.conv_channels, self.conv_k_size, 80 | stride=self.conv_stride, bias=False, padding= get_padding(in_height, 81 | in_width, self.conv_k_size[0], self.conv_k_size[1], self.conv_stride, 82 | d_h=dilation[0]), dilation = dilation[0]) 83 | 84 | self.resnet = self.build_resnet(in_height, in_width, n_blocks, n_channels[1:], 85 | dilation[1:], res_kernel, stride) 86 | self.avg_pool = nn.AvgPool2d((in_height, in_width)) 87 | self.flatten = Flatten() 88 | 89 | def build_resnet(self,in_height, in_width, n_blocks, n_channels, dilation, res_kernel, stride): 90 | res_blocks = [] 91 | for i in range(n_blocks): 92 | input_channels = self.conv_channels if i == 0 else n_channels[i-1] 93 | output_channels = n_channels[i] 94 | res_blocks.append(('Res_{}'.format(i+1), 95 | ResidualBlock(input_channels, in_height, in_width, 96 | output_channels, res_kernel, dilation[i], stride))) 97 | return nn.Sequential(OrderedDict(res_blocks)) 98 | 99 | def forward(self, x): 100 | out = self.tc(x) 101 | out = self.conv1(out) 102 | out = self.resnet(out) 103 | out = self.avg_pool(out) 104 | out = self.flatten(out) 105 | return out 106 | 107 | 108 | def TCResNet8(in_c, in_h, in_w, width_multiplier=1.0): 109 | n_blocks = 3 110 | n_channels = [16, 24, 32, 48] 111 | conv_kernel = (3,1) 112 | res_kernel = (9,1) 113 | dilation = [1] * 4 114 | n_channels = [int(x * width_multiplier) for x in n_channels] 115 | 116 | return TCResNet(in_w, in_h, in_c, n_blocks, n_channels, conv_kernel, res_kernel, dilation) 117 | 118 | 119 | def TCResNet14(in_c, in_h, in_w, width_multiplier=1.0): 120 | n_blocks = 6 121 | n_channels = [16, 24, 24, 32, 32, 48, 48] 122 | conv_kernel = (3,1) 123 | res_kernel = (9,1) 124 | dilation = [1] * 4 125 | n_channels = [int(x * width_multiplier) for x in n_channels] 126 | 127 | return TCResNet(in_w, in_h, in_c, n_blocks, n_channels, conv_kernel, res_kernel, dilation) 128 | 129 | def TCResNet8Dilated(in_c, in_h, in_w, width_multiplier=1.0): 130 | n_blocks = 3 131 | n_channels = [16, 24, 32, 48] 132 | n_channels = [int(x * width_multiplier) for x in n_channels] 133 | conv_kernel = (3,1) 134 | res_kernel = (7,1) 135 | dilation = [1,1,2,4] 136 | return TCResNet(in_w, in_h, in_c, n_blocks, n_channels, conv_kernel, res_kernel, dilation, stride=1) -------------------------------------------------------------------------------- /protonets/models/encoder/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ArchitParnami/Few-Shot-KWS/04823543110be9d186b3a5ecbc695548a6cec097/protonets/models/encoder/__init__.py -------------------------------------------------------------------------------- /protonets/models/encoder/baseUtil.py: -------------------------------------------------------------------------------- 1 | import torch.nn as nn 2 | import torch 3 | 4 | 5 | def same_padding(n, f, s, d=1): 6 | p = int(((s-1)*(n-1) + d*(f-1)) / 2) 7 | return p 8 | 9 | def get_padding(in_h, in_w, f_h, f_w, s, d_h=1, d_w=1): 10 | p_h = same_padding(in_h, f_h, s, d_h) 11 | p_w = same_padding(in_w, f_w, s, d_w) 12 | return (p_h, p_w) 13 | 14 | class Flatten(nn.Module): 15 | def __init__(self): 16 | super(Flatten, self).__init__() 17 | 18 | def forward(self, x): 19 | return x.view(x.size(0), -1) 20 | 21 | class TC(nn.Module): 22 | def __init__(self): 23 | super(TC, self).__init__() 24 | 25 | def forward(self, x): 26 | return torch.transpose(x, 1, 3) -------------------------------------------------------------------------------- /protonets/models/encoder/default.py: -------------------------------------------------------------------------------- 1 | import torch 2 | import torch.nn as nn 3 | from protonets.models.encoder.baseUtil import * 4 | 5 | class C64(nn.Module): 6 | def __init__(self, in_channels, hid_channels, out_channels): 7 | super(C64, self).__init__() 8 | self.encoder = self.build_encoder(in_channels, hid_channels, out_channels) 9 | 10 | def build_encoder(self, in_dim, hid_dim, z_dim): 11 | encoder = nn.Sequential( 12 | self.conv_block(in_dim, hid_dim), 13 | self.conv_block(hid_dim, hid_dim), 14 | self.conv_block(hid_dim, hid_dim), 15 | self.conv_block(hid_dim, z_dim), 16 | Flatten() 17 | ) 18 | return encoder 19 | 20 | def conv_block(self, in_channels, out_channels): 21 | return nn.Sequential( 22 | nn.Conv2d(in_channels, out_channels, 3, padding=1), 23 | nn.BatchNorm2d(out_channels), 24 | nn.ReLU(), 25 | nn.MaxPool2d(2) 26 | ) 27 | 28 | def forward(self, x): 29 | out = self.encoder(x) 30 | return out -------------------------------------------------------------------------------- /protonets/models/factory.py: -------------------------------------------------------------------------------- 1 | MODEL_REGISTRY = {} 2 | 3 | def register_model(model_name): 4 | def decorator(f): 5 | MODEL_REGISTRY[model_name] = f 6 | return f 7 | 8 | return decorator 9 | 10 | def get_model(model_name, model_opt): 11 | if model_name in MODEL_REGISTRY: 12 | return MODEL_REGISTRY[model_name](**model_opt) 13 | else: 14 | raise ValueError("Unknown model {:s}".format(model_name)) 15 | -------------------------------------------------------------------------------- /protonets/models/few_shot.py: -------------------------------------------------------------------------------- 1 | import torch 2 | import torch.nn as nn 3 | import torch.nn.functional as F 4 | 5 | from torch.autograd import Variable 6 | 7 | from protonets.models import register_model 8 | from protonets.models.encoder.default import C64 9 | from protonets.models.encoder.GoogleKWS import cnn_trad_fpool3 10 | from protonets.models.encoder.TCResNet import TCResNet8, TCResNet8Dilated 11 | #from torch.utils.tensorboard import SummaryWriter 12 | 13 | from .utils import euclidean_dist 14 | 15 | class Protonet(nn.Module): 16 | def __init__(self, encoder, encoding): 17 | super(Protonet, self).__init__() 18 | #self.encoding = encoding 19 | self.encoder = encoder 20 | #self.write = False 21 | 22 | def loss(self, sample): 23 | xs = Variable(sample['xs']) # support 24 | xq = Variable(sample['xq']) # query 25 | 26 | n_class = xs.size(0) 27 | assert xq.size(0) == n_class 28 | n_support = xs.size(1) 29 | n_query = xq.size(1) 30 | 31 | target_inds = torch.arange(0, n_class).view(n_class, 1, 1).expand(n_class, n_query, 1).long() 32 | target_inds = Variable(target_inds, requires_grad=False) 33 | 34 | if xq.is_cuda: 35 | target_inds = target_inds.cuda() 36 | 37 | x = torch.cat([xs.view(n_class * n_support, *xs.size()[2:]), 38 | xq.view(n_class * n_query, *xq.size()[2:])], 0) 39 | 40 | #if not self.write: 41 | #writer = SummaryWriter('runs/{}'.format(self.encoding)) 42 | #writer.add_graph(self.encoder, x) 43 | #writer.close() 44 | #self.write = True 45 | 46 | z = self.encoder.forward(x) 47 | z_dim = z.size(-1) 48 | 49 | z_proto = z[:n_class*n_support].view(n_class, n_support, z_dim).mean(1) 50 | zq = z[n_class*n_support:] 51 | 52 | dists = euclidean_dist(zq, z_proto) 53 | 54 | log_p_y = F.log_softmax(-dists, dim=1).view(n_class, n_query, -1) 55 | 56 | loss_val = -log_p_y.gather(2, target_inds).squeeze().view(-1).mean() 57 | 58 | _, y_hat = log_p_y.max(2) 59 | acc_val = torch.eq(y_hat, target_inds.squeeze()).float().mean() 60 | 61 | return loss_val, { 62 | 'loss': loss_val.item(), 63 | 'acc': acc_val.item() 64 | } 65 | 66 | 67 | def get_enocder(encoding, x_dim, hid_dim, out_dim): 68 | if encoding == 'C64': 69 | return C64(x_dim[0], hid_dim, out_dim) 70 | elif encoding == 'cnn-trad-fpool3': 71 | return cnn_trad_fpool3(x_dim[0], hid_dim, out_dim) 72 | elif encoding == 'TCResNet8': 73 | return TCResNet8(x_dim[0], x_dim[1], x_dim[2]) 74 | elif encoding == 'TCResNet8Dilated': 75 | return TCResNet8Dilated(x_dim[0], x_dim[1], x_dim[2]) 76 | 77 | @register_model('protonet_conv') 78 | def load_protonet_conv(**kwargs): 79 | x_dim = kwargs['x_dim'] 80 | hid_dim = kwargs['hid_dim'] 81 | z_dim = kwargs['z_dim'] 82 | encoding = kwargs['encoding'] 83 | encoder = get_enocder(encoding, x_dim, hid_dim, z_dim) 84 | return Protonet(encoder, encoding) 85 | -------------------------------------------------------------------------------- /protonets/models/utils.py: -------------------------------------------------------------------------------- 1 | import torch 2 | 3 | def euclidean_dist(x, y): 4 | # x: N x D 5 | # y: M x D 6 | n = x.size(0) 7 | m = y.size(0) 8 | d = x.size(1) 9 | assert d == y.size(1) 10 | 11 | x = x.unsqueeze(1).expand(n, m, d) 12 | y = y.unsqueeze(0).expand(n, m, d) 13 | 14 | return torch.pow(x - y, 2).sum(2) 15 | -------------------------------------------------------------------------------- /protonets/utils/__init__.py: -------------------------------------------------------------------------------- 1 | def filter_opt(opt, tag): 2 | ret = { } 3 | 4 | for k,v in opt.items(): 5 | tokens = k.split('.') 6 | if tokens[0] == tag: 7 | ret['.'.join(tokens[1:])] = v 8 | 9 | return ret 10 | 11 | def format_opts(d): 12 | ret = [] 13 | for k,v in d.items(): 14 | if isinstance(v, bool) and v == True: 15 | ret = ret + ["--" + k] 16 | elif isinstance(v, bool) and v == False: 17 | pass 18 | else: 19 | ret = ret + ["--" + k, str(v)] 20 | return ret 21 | 22 | def merge_dict(x, y): 23 | ret = x.copy() 24 | 25 | for k,v in y.items(): 26 | ret[k] = v 27 | 28 | return ret 29 | -------------------------------------------------------------------------------- /protonets/utils/data.py: -------------------------------------------------------------------------------- 1 | import protonets.data 2 | import os 3 | from protonets.data.FewShotSpeechData import FewShotSpeechDataset 4 | import torch 5 | from protonets.utils import filter_opt 6 | from protonets.data.base import EpisodicBatchSampler, SequentialBatchSampler, EpisodicSpeechBatchSampler 7 | 8 | def load(opt, splits): 9 | if opt['data.dataset'] in ['googlespeech']: 10 | ds = loader(opt, splits) 11 | else: 12 | raise ValueError("Unknown dataset: {:s}".format(opt['data.dataset'])) 13 | 14 | return ds 15 | 16 | def loader(opt, splits): 17 | 18 | ret = { } 19 | for split in splits: 20 | if split in ['val', 'test'] and opt['data.test_way'] != 0: 21 | n_way = opt['data.test_way'] 22 | else: 23 | n_way = opt['data.way'] 24 | 25 | if split in ['val', 'test'] and opt['data.test_shot'] != 0: 26 | n_support = opt['data.test_shot'] 27 | else: 28 | n_support = opt['data.shot'] 29 | 30 | if split in ['val', 'test'] and opt['data.test_query'] != 0: 31 | n_query = opt['data.test_query'] 32 | else: 33 | n_query = opt['data.query'] 34 | 35 | if split in ['val', 'test']: 36 | n_episodes = opt['data.test_episodes'] 37 | else: 38 | n_episodes = opt['data.train_episodes'] 39 | 40 | if opt['data.dataset'] == 'googlespeech': 41 | speech_args = filter_opt(opt, 'speech') 42 | data_dir = os.path.join(os.path.dirname(__file__), '../../data/speech_commands/core') 43 | class_file = os.path.join(os.path.dirname(__file__), '../../data/speech_commands/core', split + '.txt') 44 | ds = FewShotSpeechDataset(data_dir, class_file, n_support, n_query, opt['data.cuda'], speech_args) 45 | 46 | if opt['data.sequential']: 47 | sampler = SequentialBatchSampler(len(ds)) 48 | else: 49 | 50 | sampler = EpisodicSpeechBatchSampler(len(ds), n_way, n_episodes, 51 | include_silence=opt['speech.include_silence'], 52 | include_unknown=opt['speech.include_unknown']) 53 | 54 | # use num_workers=0, otherwise may receive duplicate episodes 55 | ret[split] = torch.utils.data.DataLoader(ds, batch_sampler=sampler, num_workers=0) 56 | 57 | return ret 58 | -------------------------------------------------------------------------------- /protonets/utils/log.py: -------------------------------------------------------------------------------- 1 | import json 2 | 3 | import numpy as np 4 | 5 | def extract_meter_values(meters): 6 | ret = { } 7 | 8 | for split in meters.keys(): 9 | ret[split] = { } 10 | for field,meter in meters[split].items(): 11 | ret[split][field] = meter.value()[0] 12 | 13 | return ret 14 | 15 | def render_meter_values(meter_values): 16 | field_info = [] 17 | for split in meter_values.keys(): 18 | for field,val in meter_values[split].items(): 19 | field_info.append("{:s} {:s} = {:0.6f}".format(split, field, val)) 20 | 21 | return ', '.join(field_info) 22 | 23 | def convert_array(d): 24 | ret = { } 25 | for k,v in d.items(): 26 | if isinstance(v, dict): 27 | ret[k] = { } 28 | for kk,vv in v.items(): 29 | ret[k][kk] = np.array(vv) 30 | else: 31 | ret[k] = np.array(v) 32 | 33 | return ret 34 | 35 | def load_trace(trace_file): 36 | ret = { } 37 | 38 | with open(trace_file, 'r') as f: 39 | for i,line in enumerate(f): 40 | vals = json.loads(line.rstrip('\n')) 41 | 42 | if i == 0: 43 | for k,v in vals.items(): 44 | if isinstance(v, dict): 45 | ret[k] = { } 46 | for kk in v.keys(): 47 | ret[k][kk] = [] 48 | else: 49 | ret[k] = [] 50 | 51 | for k,v in vals.items(): 52 | if isinstance(v, dict): 53 | for kk,vv in v.items(): 54 | ret[k][kk].append(vv) 55 | else: 56 | ret[k].append(v) 57 | 58 | return convert_array(ret) 59 | -------------------------------------------------------------------------------- /protonets/utils/model.py: -------------------------------------------------------------------------------- 1 | from tqdm import tqdm 2 | 3 | from protonets.utils import filter_opt 4 | from protonets.models import get_model 5 | 6 | def load(opt): 7 | model_opt = filter_opt(opt, 'model') 8 | model_name = model_opt['model_name'] 9 | 10 | del model_opt['model_name'] 11 | 12 | return get_model(model_name, model_opt) 13 | 14 | def evaluate(model, data_loader, meters, desc=None): 15 | model.eval() 16 | 17 | for field,meter in meters.items(): 18 | meter.reset() 19 | 20 | if desc is not None: 21 | data_loader = tqdm(data_loader, desc=desc) 22 | 23 | for sample in data_loader: 24 | _, output = model.loss(sample) 25 | for field, meter in meters.items(): 26 | meter.add(output[field]) 27 | 28 | return meters 29 | -------------------------------------------------------------------------------- /requirements.txt: -------------------------------------------------------------------------------- 1 | torch 2 | torchnet 3 | torchaudio 4 | torchvision -------------------------------------------------------------------------------- /results/create_table.ipynb: -------------------------------------------------------------------------------- 1 | { 2 | "cells": [ 3 | { 4 | "cell_type": "code", 5 | "execution_count": 1, 6 | "metadata": {}, 7 | "outputs": [], 8 | "source": [ 9 | "%matplotlib inline\n", 10 | "import matplotlib.pyplot as plt\n", 11 | "import pandas as pd\n", 12 | "import numpy as np" 13 | ] 14 | }, 15 | { 16 | "cell_type": "code", 17 | "execution_count": 2, 18 | "metadata": {}, 19 | "outputs": [], 20 | "source": [ 21 | "files = ['cnn_trad_fpool3.csv', 'C64.csv','TCResNet8.csv', 'TCResNetDilated.csv']" 22 | ] 23 | }, 24 | { 25 | "cell_type": "code", 26 | "execution_count": 3, 27 | "metadata": {}, 28 | "outputs": [], 29 | "source": [ 30 | "labels = [file[:-4] for file in files]\n", 31 | "results = [pd.read_csv(file) for file in files]" 32 | ] 33 | }, 34 | { 35 | "cell_type": "code", 36 | "execution_count": 4, 37 | "metadata": {}, 38 | "outputs": [], 39 | "source": [ 40 | "def get_cases():\n", 41 | " cases = []\n", 42 | " for i in [0,1,3,7]:\n", 43 | "\n", 44 | " include_background = False\n", 45 | " include_silence = False\n", 46 | " include_unknown = False\n", 47 | " title = 'core'\n", 48 | "\n", 49 | " if i == 1:\n", 50 | " include_background = True\n", 51 | " title = 'core + background'\n", 52 | " elif i == 2:\n", 53 | " include_silence = True\n", 54 | " title = 'core + silence'\n", 55 | " elif i == 3:\n", 56 | " include_unknown = True\n", 57 | " title = 'core + unknown'\n", 58 | " elif i == 4:\n", 59 | " include_background = True\n", 60 | " include_silence = True\n", 61 | " title = 'core + background + silence'\n", 62 | " elif i == 5:\n", 63 | " include_background = True\n", 64 | " include_unknown = True\n", 65 | " title = 'core + background + unknown'\n", 66 | " elif i == 6:\n", 67 | " include_unknown = True\n", 68 | " include_silence = True\n", 69 | " title = 'core + silence + unknown'\n", 70 | " elif i == 7:\n", 71 | " include_background = True\n", 72 | " include_unknown = True\n", 73 | " include_silence = True\n", 74 | " title = 'core + background + silence + unknown'\n", 75 | " \n", 76 | " cases.append([title, (include_background, include_silence, include_unknown)])\n", 77 | " return cases" 78 | ] 79 | }, 80 | { 81 | "cell_type": "code", 82 | "execution_count": 5, 83 | "metadata": {}, 84 | "outputs": [], 85 | "source": [ 86 | "def get_way_shots():\n", 87 | " ways = [2,4]\n", 88 | " shots = [1,5]\n", 89 | " cols = []\n", 90 | " for way in ways:\n", 91 | " for shot in shots:\n", 92 | " cols.append((way, shot))\n", 93 | " return cols" 94 | ] 95 | }, 96 | { 97 | "cell_type": "code", 98 | "execution_count": 6, 99 | "metadata": {}, 100 | "outputs": [], 101 | "source": [ 102 | "rows = get_cases()" 103 | ] 104 | }, 105 | { 106 | "cell_type": "code", 107 | "execution_count": 7, 108 | "metadata": {}, 109 | "outputs": [], 110 | "source": [ 111 | "cols = get_way_shots()" 112 | ] 113 | }, 114 | { 115 | "cell_type": "code", 116 | "execution_count": 8, 117 | "metadata": {}, 118 | "outputs": [], 119 | "source": [ 120 | "table = []\n", 121 | "for case in rows:\n", 122 | " name = case[0]\n", 123 | " background, silence, unknown = case[1]\n", 124 | " for model_name, df in zip(labels, results):\n", 125 | " output = []\n", 126 | " for way, shot in cols:\n", 127 | " x = df[(df['test.way'] == way) &\n", 128 | " (df['test.shot'] == shot) &\n", 129 | " (df['background'] == background) & \n", 130 | " (df['silence'] == silence) & \n", 131 | " (df['unknown'] == unknown)][['test.acc.mean', 'test.acc.confidence']].values[0]\n", 132 | " acc = round(x[0]*100,2)\n", 133 | " cnf = round(x[1], 2)\n", 134 | " out = '{} $\\pm$ {}'.format(acc, cnf)\n", 135 | " output.append(out)\n", 136 | " row = name + ',' + model_name + ',' + ','.join(output)\n", 137 | " table.append(row)" 138 | ] 139 | }, 140 | { 141 | "cell_type": "code", 142 | "execution_count": 9, 143 | "metadata": {}, 144 | "outputs": [], 145 | "source": [ 146 | "file = 'table.csv'\n", 147 | "with open(file, 'w') as wf:\n", 148 | " for row in table:\n", 149 | " wf.write(row)\n", 150 | " wf.write('\\n')" 151 | ] 152 | }, 153 | { 154 | "cell_type": "code", 155 | "execution_count": null, 156 | "metadata": {}, 157 | "outputs": [], 158 | "source": [] 159 | } 160 | ], 161 | "metadata": { 162 | "kernelspec": { 163 | "display_name": "Python 3", 164 | "language": "python", 165 | "name": "python3" 166 | }, 167 | "language_info": { 168 | "codemirror_mode": { 169 | "name": "ipython", 170 | "version": 3 171 | }, 172 | "file_extension": ".py", 173 | "mimetype": "text/x-python", 174 | "name": "python", 175 | "nbconvert_exporter": "python", 176 | "pygments_lexer": "ipython3", 177 | "version": "3.7.3" 178 | } 179 | }, 180 | "nbformat": 4, 181 | "nbformat_minor": 2 182 | } 183 | -------------------------------------------------------------------------------- /results/old/results_3CNN.csv: -------------------------------------------------------------------------------- 1 | train.way,test.way,train.shot,test.shot,train.query,test.query,train.episodes,test.episodes,background,silence,unknown,epochs,lr,wd,train.acc,val.acc,timestamp 2 | 2,2,1,1,5,15,200,100,False,False,False,200,0.001,0.0,0.8065000096708534,0.7653333750367164,2019-10-28 18:36:06 3 | 2,2,5,5,5,15,200,100,True,False,False,200,0.001,0.0,0.9135000132024288,0.8353333783149721,2019-10-28 18:36:15 4 | 4,4,10,10,5,15,200,100,False,False,True,200,0.001,0.0,0.9281999787688252,0.7733333480358123,2019-10-28 19:27:53 5 | 4,4,10,10,5,15,200,100,True,True,False,200,0.001,0.0,0.8681999772787095,0.6394666779041291,2019-10-28 19:28:21 6 | 4,4,10,10,5,15,200,100,True,False,True,200,0.001,0.0,0.9079999780654905,0.7097333472967147,2019-10-28 19:28:59 7 | 4,4,10,10,5,15,200,100,False,True,True,200,0.001,0.0,0.9558333647251127,0.8231111365556718,2019-10-28 19:29:04 8 | 4,4,10,10,5,15,200,100,True,True,True,200,0.001,0.0,0.8766667154431341,0.7330000215768815,2019-10-28 19:29:57 9 | 4,4,15,15,5,15,200,100,False,False,False,200,0.001,0.0,0.9562500023841859,0.8273333805799481,2019-10-28 19:30:08 10 | 4,4,15,15,5,15,200,100,True,False,False,200,0.001,0.0,0.9482500037550925,0.8098333770036699,2019-10-28 19:30:08 11 | 4,4,15,15,5,15,200,100,False,True,False,200,0.001,0.0,0.9193999812006948,0.6854666811227799,2019-10-28 19:30:12 12 | 4,4,15,15,5,15,200,100,False,False,True,200,0.001,0.0,0.9513999816775318,0.8025333535671235,2019-10-28 19:31:20 13 | 4,4,15,15,5,15,200,100,True,True,False,200,0.001,0.0,0.8971999788284301,0.6561333486437798,2019-10-28 19:33:29 14 | 2,2,5,5,5,15,200,100,False,True,False,200,0.001,0.0,0.9110000288486478,0.7391111251711844,2019-10-28 18:36:15 15 | 4,4,15,15,5,15,200,100,True,False,True,200,0.001,0.0,0.8245999753475188,0.7117333477735518,2019-10-28 19:33:45 16 | 4,4,15,15,5,15,200,100,False,True,True,200,0.001,0.0,0.9445000353455549,0.837777802944183,2019-10-28 19:34:17 17 | 4,4,15,15,5,15,200,100,True,True,True,200,0.001,0.0,0.8965000468492508,0.7587778019905088,2019-10-28 19:34:56 18 | 4,4,20,20,5,15,200,100,False,False,False,200,0.001,0.0,0.9625000020861622,0.8510000437498091,2019-10-28 19:34:56 19 | 4,4,20,20,5,15,200,100,True,False,False,200,0.001,0.0,0.9457500028610228,0.8078333759307861,2019-10-28 19:36:57 20 | 4,4,20,20,5,15,200,100,False,True,False,200,0.001,0.0,0.9275999844074247,0.7177333486080169,2019-10-28 19:36:57 21 | 4,4,20,20,5,15,200,100,False,False,True,200,0.001,0.0,0.9539999839663504,0.8266666865348814,2019-10-28 19:37:31 22 | 4,4,20,20,5,15,200,100,True,True,False,200,0.001,0.0,0.9087999829649928,0.6856000119447708,2019-10-28 19:38:26 23 | 4,4,20,20,5,15,200,100,True,False,True,200,0.001,0.0,0.8985999760031701,0.751333349943161,2019-10-28 19:38:29 24 | 4,4,20,20,5,15,200,100,False,True,True,200,0.001,0.0,0.9698333579301837,0.8542222476005555,2019-10-28 19:39:13 25 | 2,2,5,5,5,15,200,100,False,False,True,200,0.001,0.0,0.918666707277298,0.8240000104904172,2019-10-28 18:36:15 26 | 4,4,20,20,5,15,200,100,True,True,True,200,0.001,0.0,0.9033333772420883,0.7735555809736254,2019-10-28 19:40:33 27 | 5,5,1,1,5,15,200,100,False,False,False,200,0.001,0.0,0.6943999797105784,0.49640000969171516,2019-10-28 19:40:57 28 | 5,5,1,1,5,15,200,100,True,False,False,200,0.001,0.0,0.5881999859213829,0.4236000120639801,2019-10-28 19:41:11 29 | 5,5,1,1,5,15,200,100,False,True,False,200,0.001,0.0,0.5628333646059035,0.41488889530301093,2019-10-28 19:43:30 30 | 5,5,1,1,5,15,200,100,False,False,True,200,0.001,0.0,0.6038333679735661,0.4650000074505808,2019-10-28 19:44:02 31 | 5,5,1,1,5,15,200,100,True,True,False,200,0.001,0.0,0.5311666965484622,0.37666667446494084,2019-10-28 19:44:12 32 | 5,5,1,1,5,15,200,100,True,False,True,200,0.001,0.0,0.5393333633244038,0.3865555624663828,2019-10-28 19:47:03 33 | 5,5,1,1,5,15,200,100,False,True,True,200,0.001,0.0,0.6948571604490281,0.5637143102288247,2019-10-28 19:47:37 34 | 5,5,1,1,5,15,200,100,True,True,True,200,0.001,0.0,0.5751428762078282,0.4380952546000481,2019-10-28 19:47:59 35 | 5,5,5,5,5,15,200,100,False,False,False,200,0.001,0.0,0.9225999760627744,0.7210666811466213,2019-10-28 19:48:05 36 | 2,2,5,5,5,15,200,100,True,True,False,200,0.001,0.0,0.8843333750963214,0.6846666836738584,2019-10-28 18:36:15 37 | 5,5,5,5,5,15,200,100,True,False,False,200,0.001,0.0,0.8861999756097796,0.6638666796684265,2019-10-28 19:48:20 38 | 5,5,5,5,5,15,200,100,False,True,False,200,0.001,0.0,0.8195000472664835,0.5930000093579295,2019-10-28 19:49:24 39 | 5,5,5,5,5,15,200,100,False,False,True,200,0.001,0.0,0.8581667134165765,0.6754444563388826,2019-10-28 19:49:28 40 | 5,5,5,5,5,15,200,100,True,True,False,200,0.001,0.0,0.7306667058169841,0.5517777866125106,2019-10-28 19:49:44 41 | 5,5,5,5,5,15,200,100,True,False,True,200,0.001,0.0,0.7961667126417163,0.6116666752099988,2019-10-28 19:49:52 42 | 5,5,5,5,5,15,200,100,False,True,True,200,0.001,0.0,0.9154285764694212,0.7316190636157991,2019-10-28 19:50:49 43 | 5,5,5,5,5,15,200,100,True,True,True,200,0.001,0.0,0.7987142975628377,0.6448571640253069,2019-10-28 19:51:01 44 | 5,5,10,10,5,15,200,100,False,False,False,200,0.001,0.0,0.9415999808907513,0.7860000181198122,2019-10-28 19:52:15 45 | 5,5,10,10,5,15,200,100,True,False,False,200,0.001,0.0,0.9271999776363374,0.7405333489179611,2019-10-28 19:53:46 46 | 5,5,10,10,5,15,200,100,False,True,False,200,0.001,0.0,0.9138333711028099,0.6442222315073016,2019-10-28 19:54:33 47 | 2,2,5,5,5,15,200,100,True,False,True,200,0.001,0.0,0.8696667155623437,0.7195555713772772,2019-10-28 18:41:38 48 | 5,5,10,10,5,15,200,100,False,False,True,200,0.001,0.0,0.8925000473856927,0.7460000240802765,2019-10-28 19:54:36 49 | 5,5,10,10,5,15,200,100,True,True,False,200,0.001,0.0,0.8846667116880419,0.6200000077486038,2019-10-28 19:55:07 50 | 5,5,10,10,5,15,200,100,True,False,True,200,0.001,0.0,0.8648333826661109,0.6888889014720917,2019-10-28 19:55:17 51 | 5,5,10,10,5,15,200,100,False,True,True,200,0.001,0.0,0.9181428620219226,0.7772381085157392,2019-10-28 19:55:43 52 | 5,5,10,10,5,15,200,100,True,True,True,200,0.001,0.0,0.9115714338421825,0.7098095411062242,2019-10-28 19:56:09 53 | 5,5,15,15,5,15,200,100,False,False,False,200,0.001,0.0,0.963399984240532,0.8374666863679888,2019-10-28 19:57:26 54 | 5,5,15,15,5,15,200,100,True,False,False,200,0.001,0.0,0.9353999796509742,0.7710666811466215,2019-10-28 19:57:32 55 | 5,5,15,15,5,15,200,100,False,True,False,200,0.001,0.0,0.9163333693146707,0.6612222313880923,2019-10-28 19:57:36 56 | 5,5,15,15,5,15,200,100,False,False,True,200,0.001,0.0,0.9261667087674139,0.7693333601951601,2019-10-28 19:57:39 57 | 5,5,15,15,5,15,200,100,True,True,False,200,0.001,0.0,0.883000042438507,0.6417777872085573,2019-10-28 19:59:12 58 | 2,2,5,5,5,15,200,100,False,True,True,200,0.001,0.0,0.9445000025629995,0.8523333817720409,2019-10-28 18:41:38 59 | 5,5,15,15,5,15,200,100,True,False,True,200,0.001,0.0,0.8496667152643202,0.683222236633301,2019-10-28 19:59:28 60 | 5,5,15,15,5,15,200,100,False,True,True,200,0.001,0.0,0.9302857184410097,0.7972381079196931,2019-10-28 20:01:09 61 | 5,5,15,15,5,15,200,100,True,True,True,200,0.001,0.0,0.8781428644061087,0.744380967617035,2019-10-28 20:03:56 62 | 5,5,20,20,5,15,200,100,False,False,False,200,0.001,0.0,0.9701999875903131,0.8361333549022675,2019-10-28 20:05:10 63 | 5,5,20,20,5,15,200,100,True,False,False,200,0.001,0.0,0.9211999788880348,0.8029333537817,2019-10-28 20:05:13 64 | 5,5,20,20,5,15,200,100,False,True,False,200,0.001,0.0,0.9008333736658096,0.6820000106096267,2019-10-28 20:05:46 65 | 5,5,20,20,5,15,200,100,False,False,True,200,0.001,0.0,0.933666706383228,0.8047778052091594,2019-10-28 20:05:51 66 | 5,5,20,20,5,15,200,100,True,True,False,200,0.001,0.0,0.8841667109727864,0.639888897538185,2019-10-28 20:06:49 67 | 5,5,20,20,5,15,200,100,True,False,True,200,0.001,0.0,0.8733333820104601,0.7166666871309282,2019-10-28 20:07:15 68 | 5,5,20,20,5,15,200,100,False,True,True,200,0.001,0.0,0.9430000033974647,0.8357142966985703,2019-10-28 20:07:20 69 | 2,2,5,5,5,15,200,100,True,True,True,200,0.001,0.0,0.8625000146031382,0.7700000429153442,2019-10-28 18:41:48 70 | 5,5,20,20,5,15,200,100,True,True,True,200,0.001,0.0,0.8955714347958564,0.7397143012285236,2019-10-28 20:08:17 71 | 2,2,10,10,5,15,200,100,False,False,False,200,0.001,0.0,0.9615000066161156,0.9026667100191118,2019-10-28 18:41:54 72 | 2,2,10,10,5,15,200,100,True,False,False,200,0.001,0.0,0.9315000118315218,0.8886667132377626,2019-10-28 18:42:07 73 | 2,2,10,10,5,15,200,100,False,True,False,200,0.001,0.0,0.9220000299811362,0.7633333450555797,2019-10-28 18:43:04 74 | 2,2,1,1,5,15,200,100,True,False,False,200,0.001,0.0,0.7890000136196611,0.7100000420212744,2019-10-28 18:36:06 75 | 2,2,10,10,5,15,200,100,False,False,True,200,0.001,0.0,0.9343333756923669,0.8477777868509295,2019-10-28 18:44:56 76 | 2,2,10,10,5,15,200,100,True,True,False,200,0.001,0.0,0.927333362400532,0.7211111262440681,2019-10-28 18:45:10 77 | 2,2,10,10,5,15,200,100,True,False,True,200,0.001,0.0,0.9086667120456695,0.806000011265278,2019-10-28 18:45:14 78 | 2,2,10,10,5,15,200,100,False,True,True,200,0.001,0.0,0.9735000005364417,0.8931667089462285,2019-10-28 18:45:18 79 | 2,2,10,10,5,15,200,100,True,True,True,200,0.001,0.0,0.957500001490116,0.8290000438690187,2019-10-28 18:45:29 80 | 2,2,15,15,5,15,200,100,False,False,False,200,0.001,0.0,0.9700000077486037,0.923666707277298,2019-10-28 18:45:35 81 | 2,2,15,15,5,15,200,100,True,False,False,200,0.001,0.0,0.9445000103116032,0.8906667155027387,2019-10-28 18:45:50 82 | 2,2,15,15,5,15,200,100,False,True,False,200,0.001,0.0,0.9423333525657657,0.7764444571733474,2019-10-28 18:47:26 83 | 2,2,15,15,5,15,200,100,False,False,True,200,0.001,0.0,0.9393333715200421,0.8671111190319062,2019-10-28 18:47:35 84 | 2,2,15,15,5,15,200,100,True,True,False,200,0.001,0.0,0.9333333575725552,0.7457777917385102,2019-10-28 18:47:38 85 | 2,2,1,1,5,15,200,100,False,True,False,200,0.001,0.0,0.7626667089760294,0.5804444618523122,2019-10-28 18:36:06 86 | 2,2,15,15,5,15,200,100,True,False,True,200,0.001,0.0,0.8686667162179943,0.8097777891159058,2019-10-28 18:48:04 87 | 2,2,15,15,5,15,200,100,False,True,True,200,0.001,0.0,0.9724999999999998,0.8905000442266463,2019-10-28 18:48:12 88 | 2,2,15,15,5,15,200,100,True,True,True,200,0.001,0.0,0.9340000072121618,0.8403333795070647,2019-10-28 18:50:18 89 | 2,2,20,20,5,15,200,100,False,False,False,200,0.001,0.0,0.96750000923872,0.9266667121648786,2019-10-28 18:50:26 90 | 2,2,20,20,5,15,200,100,True,False,False,200,0.001,0.0,0.9560000085830689,0.9076667112112047,2019-10-28 18:51:29 91 | 2,2,20,20,5,15,200,100,False,True,False,200,0.001,0.0,0.9446666860580443,0.7988889008760454,2019-10-28 18:51:36 92 | 2,2,20,20,5,15,200,100,False,False,True,200,0.001,0.0,0.971666687428951,0.8802222293615339,2019-10-28 18:51:36 93 | 2,2,20,20,5,15,200,100,True,True,False,200,0.001,0.0,0.9253333607316016,0.7528889030218124,2019-10-28 18:51:39 94 | 2,2,20,20,5,15,200,100,True,False,True,200,0.001,0.0,0.961000027656555,0.8242222326993939,2019-10-28 18:52:40 95 | 2,2,20,20,5,15,200,100,False,True,True,200,0.001,0.0,0.9724999997019764,0.9070000463724139,2019-10-28 18:52:51 96 | 2,2,1,1,5,15,200,100,False,False,True,200,0.001,0.0,0.7330000381916764,0.6448889051377774,2019-10-28 18:36:06 97 | 2,2,20,20,5,15,200,100,True,True,True,200,0.001,0.0,0.9317500090599062,0.854833378791809,2019-10-28 18:53:14 98 | 3,3,1,1,5,15,200,100,False,False,False,200,0.001,0.0,0.7726667103171355,0.6533333507180219,2019-10-28 18:54:23 99 | 3,3,1,1,5,15,200,100,True,False,False,200,0.001,0.0,0.6716666999831794,0.5686666826903818,2019-10-28 18:54:23 100 | 3,3,1,1,5,15,200,100,False,True,False,200,0.001,0.0,0.7160000124573705,0.509166693985462,2019-10-28 18:54:25 101 | 3,3,1,1,5,15,200,100,False,False,True,200,0.001,0.0,0.700000012367963,0.5868333655595774,2019-10-28 18:54:41 102 | 3,3,1,1,5,15,200,100,True,True,False,200,0.001,0.0,0.5685000103712081,0.4205000227689742,2019-10-28 18:56:27 103 | 3,3,1,1,5,15,200,100,True,False,True,200,0.001,0.0,0.6410000117868185,0.504333358108997,2019-10-28 18:57:15 104 | 3,3,1,1,5,15,200,100,False,True,True,200,0.001,0.0,0.769199978560209,0.6682666781544684,2019-10-28 18:57:39 105 | 3,3,1,1,5,15,200,100,True,True,True,200,0.001,0.0,0.6623999832570551,0.5800000131130214,2019-10-28 18:59:24 106 | 3,3,5,5,5,15,200,100,False,False,False,200,0.001,0.0,0.9113333722949029,0.8091111224889757,2019-10-28 18:59:44 107 | 2,2,1,1,5,15,200,100,True,True,False,200,0.001,0.0,0.6246666977554556,0.5175555725395676,2019-10-28 18:36:06 108 | 3,3,5,5,5,15,200,100,True,False,False,200,0.001,0.0,0.8896667090058326,0.7637777918577195,2019-10-28 18:59:44 109 | 3,3,5,5,5,15,200,100,False,True,False,200,0.001,0.0,0.8790000075101853,0.6463333693146704,2019-10-28 19:00:08 110 | 3,3,5,5,5,15,200,100,False,False,True,200,0.001,0.0,0.8542500114440915,0.7436667090654373,2019-10-28 19:00:17 111 | 3,3,5,5,5,15,200,100,True,True,False,200,0.001,0.0,0.8417500120401382,0.6156667020916938,2019-10-28 19:00:20 112 | 3,3,5,5,5,15,200,100,True,False,True,200,0.001,0.0,0.8885000121593479,0.6906667065620421,2019-10-28 19:00:46 113 | 3,3,5,5,5,15,200,100,False,True,True,200,0.001,0.0,0.9359999805688859,0.8080000209808347,2019-10-28 19:00:52 114 | 3,3,5,5,5,15,200,100,True,True,True,200,0.001,0.0,0.88819997638464,0.7425333499908445,2019-10-28 19:01:12 115 | 3,3,10,10,5,15,200,100,False,False,False,200,0.001,0.0,0.9553333622217179,0.8337777876853943,2019-10-28 19:02:48 116 | 3,3,10,10,5,15,200,100,True,False,False,200,0.001,0.0,0.9266667082905772,0.8184444552659987,2019-10-28 19:02:51 117 | 3,3,10,10,5,15,200,100,False,True,False,200,0.001,0.0,0.9097500070929527,0.7041667041182518,2019-10-28 19:03:23 118 | 2,2,1,1,5,15,200,100,True,False,True,200,0.001,0.0,0.6983333714306352,0.5848889072239397,2019-10-28 18:36:06 119 | 3,3,10,10,5,15,200,100,False,False,True,200,0.001,0.0,0.8825000146031383,0.797666711807251,2019-10-28 19:04:23 120 | 3,3,10,10,5,15,200,100,True,True,False,200,0.001,0.0,0.859000010341406,0.6578333702683445,2019-10-28 19:04:44 121 | 3,3,10,10,5,15,200,100,True,False,True,200,0.001,0.0,0.9287500086426739,0.7301667070388793,2019-10-28 19:06:18 122 | 3,3,10,10,5,15,200,100,False,True,True,200,0.001,0.0,0.9515999844670295,0.847866686582565,2019-10-28 19:06:28 123 | 3,3,10,10,5,15,200,100,True,True,True,200,0.001,0.0,0.9133999782800677,0.7801333504915238,2019-10-28 19:06:29 124 | 3,3,15,15,5,15,200,100,False,False,False,200,0.001,0.0,0.9563333630561831,0.8688888967037199,2019-10-28 19:06:40 125 | 3,3,15,15,5,15,200,100,True,False,False,200,0.001,0.0,0.9466666996479032,0.8626666748523714,2019-10-28 19:07:08 126 | 3,3,15,15,5,15,200,100,False,True,False,200,0.001,0.0,0.9322500011324885,0.7326667067408563,2019-10-28 19:07:10 127 | 3,3,15,15,5,15,200,100,False,False,True,200,0.001,0.0,0.9560000023245815,0.8313333767652512,2019-10-28 19:07:37 128 | 3,3,15,15,5,15,200,100,True,True,False,200,0.001,0.0,0.9147500035166742,0.7060000383853914,2019-10-28 19:07:40 129 | 2,2,1,1,5,15,200,100,False,True,True,200,0.001,0.0,0.7987500104308124,0.7435000401735302,2019-10-28 18:36:15 130 | 3,3,15,15,5,15,200,100,True,False,True,200,0.001,0.0,0.8887500125169755,0.7555000430345535,2019-10-28 19:09:11 131 | 3,3,15,15,5,15,200,100,False,True,True,200,0.001,0.0,0.9537999835610389,0.852133351564407,2019-10-28 19:09:39 132 | 3,3,15,15,5,15,200,100,True,True,True,200,0.001,0.0,0.9111999770998952,0.8029333513975142,2019-10-28 19:09:44 133 | 3,3,20,20,5,15,200,100,False,False,False,200,0.001,0.0,0.9706666871905328,0.8908888953924181,2019-10-28 19:11:01 134 | 3,3,20,20,5,15,200,100,True,False,False,200,0.001,0.0,0.9490000307559965,0.8451111203432082,2019-10-28 19:12:13 135 | 3,3,20,20,5,15,200,100,False,True,False,200,0.001,0.0,0.9357500010728838,0.7501667076349259,2019-10-28 19:12:16 136 | 3,3,20,20,5,15,200,100,False,False,True,200,0.001,0.0,0.9255000078678131,0.8343333798646928,2019-10-28 19:13:24 137 | 3,3,20,20,5,15,200,100,True,True,False,200,0.001,0.0,0.9062500059604642,0.7153333729505541,2019-10-28 19:13:40 138 | 3,3,20,20,5,15,200,100,True,False,True,200,0.001,0.0,0.9347500064969065,0.7796667128801343,2019-10-28 19:14:40 139 | 3,3,20,20,5,15,200,100,False,True,True,200,0.001,0.0,0.9743999883532524,0.8814666873216628,2019-10-28 19:15:10 140 | 2,2,1,1,5,15,200,100,True,True,True,200,0.001,0.0,0.7117500115931035,0.6468333685398104,2019-10-28 18:36:15 141 | 3,3,20,20,5,15,200,100,True,True,True,200,0.001,0.0,0.9463999822735785,0.8229333519935605,2019-10-28 19:15:40 142 | 4,4,1,1,5,15,200,100,False,False,False,200,0.001,0.0,0.7577500127255913,0.5635000303387643,2019-10-28 19:15:54 143 | 4,4,1,1,5,15,200,100,True,False,False,200,0.001,0.0,0.6575000117719173,0.512333359122276,2019-10-28 19:15:59 144 | 4,4,1,1,5,15,200,100,False,True,False,200,0.001,0.0,0.6419999849051236,0.42066667541861547,2019-10-28 19:16:19 145 | 4,4,1,1,5,15,200,100,False,False,True,200,0.001,0.0,0.672599981278181,0.5046666750311851,2019-10-28 19:17:32 146 | 4,4,1,1,5,15,200,100,True,True,False,200,0.001,0.0,0.5249999868124721,0.39173334404826166,2019-10-28 19:19:19 147 | 4,4,1,1,5,15,200,100,True,False,True,200,0.001,0.0,0.5763999870419502,0.4342666763067248,2019-10-28 19:20:47 148 | 4,4,1,1,5,15,200,100,False,True,True,200,0.001,0.0,0.7308333760499954,0.599333344101906,2019-10-28 19:20:57 149 | 4,4,1,1,5,15,200,100,True,True,True,200,0.001,0.0,0.5996666988730431,0.5105555611848831,2019-10-28 19:21:32 150 | 4,4,5,5,5,15,200,100,False,False,False,200,0.001,0.0,0.9120000082254407,0.7600000435113907,2019-10-28 19:22:12 151 | 2,2,5,5,5,15,200,100,False,False,False,200,0.001,0.0,0.9505000123381617,0.8746667110919952,2019-10-28 18:36:15 152 | 4,4,5,5,5,15,200,100,True,False,False,200,0.001,0.0,0.8892500102519988,0.7001667082309723,2019-10-28 19:22:16 153 | 4,4,5,5,5,15,200,100,False,True,False,200,0.001,0.0,0.8669999787211423,0.6100000119209291,2019-10-28 19:22:32 154 | 4,4,5,5,5,15,200,100,False,False,True,200,0.001,0.0,0.8653999757766729,0.7334666866064073,2019-10-28 19:22:59 155 | 4,4,5,5,5,15,200,100,True,True,False,200,0.001,0.0,0.8507999759912491,0.5717333465814589,2019-10-28 19:23:11 156 | 4,4,5,5,5,15,200,100,True,False,True,200,0.001,0.0,0.8113999745249745,0.6253333458304406,2019-10-28 19:24:14 157 | 4,4,5,5,5,15,200,100,False,True,True,200,0.001,0.0,0.8956667143106457,0.7628889155387878,2019-10-28 19:24:26 158 | 4,4,5,5,5,15,200,100,True,True,True,200,0.001,0.0,0.7971667137742042,0.654000009894371,2019-10-28 19:25:19 159 | 4,4,10,10,5,15,200,100,False,False,False,200,0.001,0.0,0.9525000032782551,0.8255000466108321,2019-10-28 19:26:36 160 | 4,4,10,10,5,15,200,100,True,False,False,200,0.001,0.0,0.9322500047087672,0.7793333798646926,2019-10-28 19:26:47 161 | 4,4,10,10,5,15,200,100,False,True,False,200,0.001,0.0,0.8971999776363373,0.6740000104904177,2019-10-28 19:26:51 162 | -------------------------------------------------------------------------------- /results/old/results_TCResNet8.csv: -------------------------------------------------------------------------------- 1 | train.way,test.way,train.shot,test.shot,train.query,test.query,train.episodes,test.episodes,background,silence,unknown,epochs,lr,wd,train.acc,val.acc,timestamp 2 | 2,2,1,1,5,15,200,100,False,False,False,200,0.001,0.0,0.9325000095367431,0.892333371639252,2019-11-06 15:40:21 3 | 2,2,5,5,5,15,200,100,True,False,False,200,0.001,0.0,0.9690000054240228,0.936666703224182,2019-11-06 15:51:45 4 | 4,4,10,10,5,15,200,100,False,False,True,200,0.001,0.0,0.964799984395504,0.87386668920517,2019-11-06 18:50:25 5 | 4,4,10,10,5,15,200,100,True,True,False,200,0.001,0.0,0.9293999853730198,0.6752000147104265,2019-11-06 18:50:49 6 | 4,4,10,10,5,15,200,100,True,False,True,200,0.001,0.0,0.9139999765157701,0.8172000169754032,2019-11-06 18:51:27 7 | 4,4,10,10,5,15,200,100,False,True,True,200,0.001,0.0,0.9593333616852759,0.8738889110088348,2019-11-06 18:56:07 8 | 4,4,10,10,5,15,200,100,True,True,True,200,0.001,0.0,0.9420000365376475,0.8575555747747425,2019-11-06 18:56:41 9 | 4,4,15,15,5,15,200,100,False,False,False,200,0.001,0.0,0.9582500016689302,0.9056667137145993,2019-11-06 18:56:57 10 | 4,4,15,15,5,15,200,100,True,False,False,200,0.001,0.0,0.9714999997615816,0.888166708946228,2019-11-06 18:58:08 11 | 4,4,15,15,5,15,200,100,False,True,False,200,0.001,0.0,0.9439999863505361,0.7584000170230868,2019-11-06 19:00:18 12 | 4,4,15,15,5,15,200,100,False,False,True,200,0.001,0.0,0.9575999829173085,0.8689333540201187,2019-11-06 19:03:49 13 | 4,4,15,15,5,15,200,100,True,True,False,200,0.001,0.0,0.9333999848365785,0.7213333487510682,2019-11-06 19:03:57 14 | 2,2,5,5,5,15,200,100,False,True,False,200,0.001,0.0,0.9593333458900446,0.8133333444595336,2019-11-06 15:51:45 15 | 4,4,15,15,5,15,200,100,True,False,True,200,0.001,0.0,0.9411999794840814,0.8349333542585374,2019-11-06 19:04:21 16 | 4,4,15,15,5,15,200,100,False,True,True,200,0.001,0.0,0.9696666911244394,0.8887777990102768,2019-11-06 19:17:02 17 | 4,4,15,15,5,15,200,100,True,True,True,200,0.001,0.0,0.9496666994690885,0.8400000238418579,2019-11-06 19:17:25 18 | 4,4,20,20,5,15,200,100,False,False,False,200,0.001,0.0,0.9600000011920925,0.8806667119264604,2019-11-06 19:25:15 19 | 4,4,20,20,5,15,200,100,True,False,False,200,0.001,0.0,0.9325000074505808,0.8786667132377627,2019-11-06 19:28:48 20 | 4,4,20,20,5,15,200,100,False,True,False,200,0.001,0.0,0.943199988603592,0.743600018620491,2019-11-06 19:29:42 21 | 4,4,20,20,5,15,200,100,False,False,True,200,0.001,0.0,0.9663999867439269,0.8822666865587235,2019-11-06 19:31:17 22 | 4,4,20,20,5,15,200,100,True,True,False,200,0.001,0.0,0.9307999843358998,0.7046666777133941,2019-11-06 19:31:30 23 | 4,4,20,20,5,15,200,100,True,False,True,200,0.001,0.0,0.9503999808430672,0.8402666866779327,2019-11-06 19:31:42 24 | 4,4,20,20,5,15,200,100,False,True,True,200,0.001,0.0,0.9588333627581596,0.9021111315488816,2019-11-06 19:42:29 25 | 2,2,5,5,5,15,200,100,False,False,True,200,0.001,0.0,0.9206667092442512,0.8635555636882779,2019-11-06 15:51:45 26 | 4,4,20,20,5,15,200,100,True,True,True,200,0.001,0.0,0.9358333715796472,0.8661111360788345,2019-11-06 19:44:15 27 | 5,5,1,1,5,15,200,100,False,False,False,200,0.001,0.0,0.9277999797463415,0.7045333474874497,2019-11-06 19:44:23 28 | 5,5,1,1,5,15,200,100,True,False,False,200,0.001,0.0,0.9267999807000156,0.6465333473682405,2019-11-06 19:44:31 29 | 5,5,1,1,5,15,200,100,False,True,False,200,0.001,0.0,0.8203333762288095,0.5823333400487902,2019-11-06 19:44:42 30 | 5,5,1,1,5,15,200,100,False,False,True,200,0.001,0.0,0.8951667091250417,0.6580000153183938,2019-11-06 19:45:26 31 | 5,5,1,1,5,15,200,100,True,True,False,200,0.001,0.0,0.5293333628773689,0.44944444984197607,2019-11-06 19:52:04 32 | 5,5,1,1,5,15,200,100,True,False,True,200,0.001,0.0,0.8538333790004254,0.5656666761636733,2019-11-06 19:55:50 33 | 5,5,1,1,5,15,200,100,False,True,True,200,0.001,0.0,0.9041428628563879,0.7100952553749083,2019-11-06 19:55:52 34 | 5,5,1,1,5,15,200,100,True,True,True,200,0.001,0.0,0.8975714346766468,0.6267619267106057,2019-11-06 19:56:44 35 | 5,5,5,5,5,15,200,100,False,False,False,200,0.001,0.0,0.9527999836206439,0.8569333535432815,2019-11-06 19:57:08 36 | 2,2,5,5,5,15,200,100,True,True,False,200,0.001,0.0,0.9593333464860917,0.7684444579482075,2019-11-06 15:51:48 37 | 5,5,5,5,5,15,200,100,True,False,False,200,0.001,0.0,0.9475999802351,0.8076000195741652,2019-11-06 19:58:20 38 | 5,5,5,5,5,15,200,100,False,True,False,200,0.001,0.0,0.9246667003631588,0.693333346247673,2019-11-06 19:59:52 39 | 5,5,5,5,5,15,200,100,False,False,True,200,0.001,0.0,0.8896667149662973,0.8102222472429276,2019-11-06 20:00:46 40 | 5,5,5,5,5,15,200,100,True,True,False,200,0.001,0.0,0.9153333714604378,0.6306666719913484,2019-11-06 20:07:50 41 | 5,5,5,5,5,15,200,100,True,False,True,200,0.001,0.0,0.8931667119264602,0.773666691184044,2019-11-06 20:11:57 42 | 5,5,5,5,5,15,200,100,False,True,True,200,0.001,0.0,0.9515714314579965,0.859142870903015,2019-11-06 20:14:58 43 | 5,5,5,5,5,15,200,100,True,True,True,200,0.001,0.0,0.9494285744428635,0.830857154130936,2019-11-06 20:17:19 44 | 5,5,10,10,5,15,200,100,False,False,False,200,0.001,0.0,0.9025999778509137,0.8270666873455049,2019-11-06 20:17:33 45 | 5,5,10,10,5,15,200,100,True,False,False,200,0.001,0.0,0.9187999787926676,0.8240000212192536,2019-11-06 20:17:43 46 | 5,5,10,10,5,15,200,100,False,True,False,200,0.001,0.0,0.8803333771228791,0.7123333513736726,2019-11-06 20:19:24 47 | 2,2,5,5,5,15,200,100,True,False,True,200,0.001,0.0,0.9450000324845311,0.8366666764020919,2019-11-06 15:52:02 48 | 5,5,10,10,5,15,200,100,False,False,True,200,0.001,0.0,0.939333372116089,0.8377778011560438,2019-11-06 20:20:35 49 | 5,5,10,10,5,15,200,100,True,True,False,200,0.001,0.0,0.9058333715796467,0.7176666867733003,2019-11-06 20:27:42 50 | 5,5,10,10,5,15,200,100,True,False,True,200,0.001,0.0,0.9330000433325767,0.7894444710016247,2019-11-06 20:28:48 51 | 5,5,10,10,5,15,200,100,False,True,True,200,0.001,0.0,0.947428574562073,0.8856190657615661,2019-11-06 20:30:26 52 | 5,5,10,10,5,15,200,100,True,True,True,200,0.001,0.0,0.9608571451902391,0.8180952489376069,2019-11-06 20:31:03 53 | 5,5,15,15,5,15,200,100,False,False,False,200,0.001,0.0,0.9177999767661098,0.8498666870594024,2019-11-06 20:32:21 54 | 5,5,15,15,5,15,200,100,True,False,False,200,0.001,0.0,0.9527999821305273,0.8146666836738585,2019-11-06 20:32:24 55 | 5,5,15,15,5,15,200,100,False,True,False,200,0.001,0.0,0.9565000191330905,0.7355555796623233,2019-11-06 20:48:02 56 | 5,5,15,15,5,15,200,100,False,False,True,200,0.001,0.0,0.9216667097806928,0.8167778027057648,2019-11-06 20:49:30 57 | 5,5,15,15,5,15,200,100,True,True,False,200,0.001,0.0,0.9248333713412289,0.6800000095367432,2019-11-06 20:56:53 58 | 2,2,5,5,5,15,200,100,False,True,True,200,0.001,0.0,0.967750001847744,0.9243333733081817,2019-11-06 15:52:31 59 | 5,5,15,15,5,15,200,100,True,False,True,200,0.001,0.0,0.9488333654403688,0.8103333562612534,2019-11-06 20:57:59 60 | 5,5,15,15,5,15,200,100,False,True,True,200,0.001,0.0,0.9632857164740559,0.8920000225305553,2019-11-06 20:59:25 61 | 5,5,15,15,5,15,200,100,True,True,True,200,0.001,0.0,0.9361428609490398,0.8222857248783109,2019-11-06 21:00:04 62 | 5,5,20,20,5,15,200,100,False,False,False,200,0.001,0.0,0.920399979054928,0.8512000215053561,2019-11-06 21:00:41 63 | 5,5,20,20,5,15,200,100,True,False,False,200,0.001,0.0,0.9065999776124952,0.8374666863679884,2019-11-06 21:01:50 64 | 5,5,20,20,5,15,200,100,False,True,False,200,0.001,0.0,0.8835000398755072,0.689222233891487,2019-11-06 21:13:38 65 | 5,5,20,20,5,15,200,100,False,False,True,200,0.001,0.0,0.9230000448226932,0.858333355784416,2019-11-06 21:13:51 66 | 5,5,20,20,5,15,200,100,True,True,False,200,0.001,0.0,0.9018333718180661,0.6974444615840913,2019-11-06 21:14:26 67 | 5,5,20,20,5,15,200,100,True,False,True,200,0.001,0.0,0.9083333793282512,0.8151111376285554,2019-11-06 21:15:33 68 | 5,5,20,20,5,15,200,100,False,True,True,200,0.001,0.0,0.9658571448922156,0.8899047875404357,2019-11-06 21:15:33 69 | 2,2,5,5,5,15,200,100,True,True,True,200,0.001,0.0,0.9600000017881394,0.871000046133995,2019-11-06 15:53:47 70 | 5,5,20,20,5,15,200,100,True,True,True,200,0.001,0.0,0.9247142902016638,0.8516190576553344,2019-11-06 21:17:41 71 | 2,2,10,10,5,15,200,100,False,False,False,200,0.001,0.0,0.9700000059604648,0.9516666978597642,2019-11-06 16:01:08 72 | 2,2,10,10,5,15,200,100,True,False,False,200,0.001,0.0,0.9775000053644183,0.9286667078733443,2019-11-06 16:01:14 73 | 2,2,10,10,5,15,200,100,False,True,False,200,0.001,0.0,0.9590000090003014,0.7951111227273939,2019-11-06 16:02:07 74 | 2,2,1,1,5,15,200,100,True,False,False,200,0.001,0.0,0.932500009536743,0.8546667170524598,2019-11-06 15:40:21 75 | 2,2,10,10,5,15,200,100,False,False,True,200,0.001,0.0,0.9490000322461128,0.907333338856697,2019-11-06 16:02:11 76 | 2,2,10,10,5,15,200,100,True,True,False,200,0.001,0.0,0.9573333483934401,0.8000000107288364,2019-11-06 16:02:22 77 | 2,2,10,10,5,15,200,100,True,False,True,200,0.001,0.0,0.9426667028665549,0.8697777855396269,2019-11-06 16:02:34 78 | 2,2,10,10,5,15,200,100,False,True,True,200,0.001,0.0,0.981499999165535,0.9345000463724138,2019-11-06 16:03:50 79 | 2,2,10,10,5,15,200,100,True,True,True,200,0.001,0.0,0.9570000016689302,0.8743333792686462,2019-11-06 16:06:58 80 | 2,2,15,15,5,15,200,100,False,False,False,200,0.001,0.0,0.9785000061988831,0.9596666944026948,2019-11-06 16:11:38 81 | 2,2,15,15,5,15,200,100,True,False,False,200,0.001,0.0,0.9690000078082084,0.9313333708047866,2019-11-06 16:11:48 82 | 2,2,15,15,5,15,200,100,False,True,False,200,0.001,0.0,0.9660000118613238,0.8202222326397897,2019-11-06 16:12:58 83 | 2,2,15,15,5,15,200,100,False,False,True,200,0.001,0.0,0.9540000304579733,0.916888893842697,2019-11-06 16:13:07 84 | 2,2,15,15,5,15,200,100,True,True,False,200,0.001,0.0,0.9620000126957895,0.7940000122785569,2019-11-06 16:13:11 85 | 2,2,1,1,5,15,200,100,False,True,False,200,0.001,0.0,0.8566667038202284,0.65844446092844,2019-11-06 15:40:21 86 | 2,2,15,15,5,15,200,100,True,False,True,200,0.001,0.0,0.9310000383853914,0.8613333415985103,2019-11-06 16:13:36 87 | 2,2,15,15,5,15,200,100,False,True,True,200,0.001,0.0,0.9714999985694883,0.9273333781957624,2019-11-06 16:22:07 88 | 2,2,15,15,5,15,200,100,True,True,True,200,0.001,0.0,0.9602500033378605,0.8955000454187394,2019-11-06 16:22:07 89 | 2,2,20,20,5,15,200,100,False,False,False,200,0.001,0.0,0.9815000054240225,0.954333360791206,2019-11-06 16:23:23 90 | 2,2,20,20,5,15,200,100,True,False,False,200,0.001,0.0,0.9740000066161156,0.9136667108535765,2019-11-06 16:27:59 91 | 2,2,20,20,5,15,200,100,False,True,False,200,0.001,0.0,0.9546666809916504,0.8097777888178826,2019-11-06 16:32:13 92 | 2,2,20,20,5,15,200,100,False,False,True,200,0.001,0.0,0.9660000249743463,0.9140000051259995,2019-11-06 16:32:50 93 | 2,2,20,20,5,15,200,100,True,True,False,200,0.001,0.0,0.957333346009255,0.7773333463072776,2019-11-06 16:33:08 94 | 2,2,20,20,5,15,200,100,True,False,True,200,0.001,0.0,0.9523333615064623,0.8397777873277662,2019-11-06 16:34:55 95 | 2,2,20,20,5,15,200,100,False,True,True,200,0.001,0.0,0.9775000008940697,0.936333379149437,2019-11-06 16:35:03 96 | 2,2,1,1,5,15,200,100,False,False,True,200,0.001,0.0,0.9070000404119493,0.8168888998031616,2019-11-06 15:40:21 97 | 2,2,20,20,5,15,200,100,True,True,True,200,0.001,0.0,0.9675000017881399,0.9123333781957628,2019-11-06 16:41:35 98 | 3,3,1,1,5,15,200,100,False,False,False,200,0.001,0.0,0.9523333579301834,0.8251111212372778,2019-11-06 16:42:13 99 | 3,3,1,1,5,15,200,100,True,False,False,200,0.001,0.0,0.9163333693146709,0.7768889012932778,2019-11-06 16:42:24 100 | 3,3,1,1,5,15,200,100,False,True,False,200,0.001,0.0,0.9070000046491621,0.6428333681821826,2019-11-06 16:51:38 101 | 3,3,1,1,5,15,200,100,False,False,True,200,0.001,0.0,0.8930000101029878,0.7668333780765532,2019-11-06 16:52:03 102 | 3,3,1,1,5,15,200,100,True,True,False,200,0.001,0.0,0.7692500112950806,0.586166698485613,2019-11-06 16:52:18 103 | 3,3,1,1,5,15,200,100,True,False,True,200,0.001,0.0,0.8190000128746033,0.6675000390410419,2019-11-06 16:52:30 104 | 3,3,1,1,5,15,200,100,False,True,True,200,0.001,0.0,0.9317999812960623,0.8064000201225278,2019-11-06 16:52:44 105 | 3,3,1,1,5,15,200,100,True,True,True,200,0.001,0.0,0.7717999777197837,0.6772000145912171,2019-11-06 16:56:46 106 | 3,3,5,5,5,15,200,100,False,False,False,200,0.001,0.0,0.96366669267416,0.8968888950347899,2019-11-06 17:01:51 107 | 2,2,1,1,5,15,200,100,True,True,False,200,0.001,0.0,0.7960000391304495,0.6528889063000679,2019-11-06 15:40:21 108 | 3,3,5,5,5,15,200,100,True,False,False,200,0.001,0.0,0.9460000327229504,0.8517777866125109,2019-11-06 17:02:18 109 | 3,3,5,5,5,15,200,100,False,True,False,200,0.001,0.0,0.9385000008344649,0.7480000394582749,2019-11-06 17:02:40 110 | 3,3,5,5,5,15,200,100,False,False,True,200,0.001,0.0,0.9527500021457675,0.8735000473260883,2019-11-06 17:02:53 111 | 3,3,5,5,5,15,200,100,True,True,False,200,0.001,0.0,0.9285000026226039,0.7138333734869958,2019-11-06 17:03:07 112 | 3,3,5,5,5,15,200,100,True,False,True,200,0.001,0.0,0.9197500091791153,0.8146667116880416,2019-11-06 17:04:03 113 | 3,3,5,5,5,15,200,100,False,True,True,200,0.001,0.0,0.9599999833106996,0.9086666923761367,2019-11-06 17:08:04 114 | 3,3,5,5,5,15,200,100,True,True,True,200,0.001,0.0,0.9641999852657318,0.8542666852474213,2019-11-06 17:09:38 115 | 3,3,10,10,5,15,200,100,False,False,False,200,0.001,0.0,0.9763333517313008,0.9188888937234881,2019-11-06 17:12:13 116 | 3,3,10,10,5,15,200,100,True,False,False,200,0.001,0.0,0.9656666919589045,0.8717777854204176,2019-11-06 17:12:54 117 | 3,3,10,10,5,15,200,100,False,True,False,200,0.001,0.0,0.9482500016689301,0.7818333739042279,2019-11-06 17:13:53 118 | 2,2,1,1,5,15,200,100,True,False,True,200,0.001,0.0,0.8780000382661822,0.7317777928709982,2019-11-06 15:40:21 119 | 3,3,10,10,5,15,200,100,False,False,True,200,0.001,0.0,0.9575000017881391,0.9020000439882276,2019-11-06 17:13:56 120 | 3,3,10,10,5,15,200,100,True,True,False,200,0.001,0.0,0.9342500025033948,0.7383333724737168,2019-11-06 17:14:18 121 | 3,3,10,10,5,15,200,100,True,False,True,200,0.001,0.0,0.9040000119805333,0.8160000443458556,2019-11-06 17:15:16 122 | 3,3,10,10,5,15,200,100,False,True,True,200,0.001,0.0,0.9641999843716625,0.9276000267267227,2019-11-06 17:23:02 123 | 3,3,10,10,5,15,200,100,True,True,True,200,0.001,0.0,0.9567999848723412,0.8649333512783053,2019-11-06 17:23:59 124 | 3,3,15,15,5,15,200,100,False,False,False,200,0.001,0.0,0.9513333624601364,0.898666672706604,2019-11-06 17:29:14 125 | 3,3,15,15,5,15,200,100,True,False,False,200,0.001,0.0,0.9720000201463701,0.9055555611848829,2019-11-06 17:30:38 126 | 3,3,15,15,5,15,200,100,False,True,False,200,0.001,0.0,0.9415000012516975,0.7703333741426469,2019-11-06 17:33:02 127 | 3,3,15,15,5,15,200,100,False,False,True,200,0.001,0.0,0.966000001132488,0.8810000461339952,2019-11-06 17:33:04 128 | 3,3,15,15,5,15,200,100,True,True,False,200,0.001,0.0,0.931500002145767,0.7380000430345536,2019-11-06 17:33:23 129 | 2,2,1,1,5,15,200,100,False,True,True,200,0.001,0.0,0.9382500043511389,0.8686667102575302,2019-11-06 15:40:21 130 | 3,3,15,15,5,15,200,100,True,False,True,200,0.001,0.0,0.952000003159046,0.8498333781957624,2019-11-06 17:34:37 131 | 3,3,15,15,5,15,200,100,False,True,True,200,0.001,0.0,0.9397999784350394,0.9158666950464248,2019-11-06 17:35:56 132 | 3,3,15,15,5,15,200,100,True,True,True,200,0.001,0.0,0.9557999822497366,0.8677333521842957,2019-11-06 17:37:06 133 | 3,3,20,20,5,15,200,100,False,False,False,200,0.001,0.0,0.9790000164508821,0.9024444502592088,2019-11-06 17:51:17 134 | 3,3,20,20,5,15,200,100,True,False,False,200,0.001,0.0,0.9710000208020212,0.8997777837514875,2019-11-06 17:51:29 135 | 3,3,20,20,5,15,200,100,False,True,False,200,0.001,0.0,0.9432500010728838,0.7815000402927399,2019-11-06 17:52:56 136 | 3,3,20,20,5,15,200,100,False,False,True,200,0.001,0.0,0.9482500037550924,0.9040000450611114,2019-11-06 17:53:01 137 | 3,3,20,20,5,15,200,100,True,True,False,200,0.001,0.0,0.9370000001788138,0.7268333762884138,2019-11-06 17:53:06 138 | 3,3,20,20,5,15,200,100,True,False,True,200,0.001,0.0,0.9395000055432321,0.8503333801031112,2019-11-06 17:54:36 139 | 3,3,20,20,5,15,200,100,False,True,True,200,0.001,0.0,0.9605999851226805,0.9160000205039979,2019-11-06 18:02:50 140 | 2,2,1,1,5,15,200,100,True,True,True,200,0.001,0.0,0.919000006318092,0.8023333775997161,2019-11-06 15:40:21 141 | 3,3,20,20,5,15,200,100,True,True,True,200,0.001,0.0,0.9533999812602997,0.8902666872739792,2019-11-06 18:04:57 142 | 4,4,1,1,5,15,200,100,False,False,False,200,0.001,0.0,0.9095000046491627,0.7533333748579024,2019-11-06 18:12:50 143 | 4,4,1,1,5,15,200,100,True,False,False,200,0.001,0.0,0.8872500081360336,0.6543333682417868,2019-11-06 18:13:15 144 | 4,4,1,1,5,15,200,100,False,True,False,200,0.001,0.0,0.9141999819874758,0.6354666781425474,2019-11-06 18:15:44 145 | 4,4,1,1,5,15,200,100,False,False,True,200,0.001,0.0,0.8819999769330021,0.717733346223831,2019-11-06 18:18:12 146 | 4,4,1,1,5,15,200,100,True,True,False,200,0.001,0.0,0.8545999781787399,0.5769333446025848,2019-11-06 18:20:19 147 | 4,4,1,1,5,15,200,100,True,False,True,200,0.001,0.0,0.8479999788105488,0.6373333454132081,2019-11-06 18:20:22 148 | 4,4,1,1,5,15,200,100,False,True,True,200,0.001,0.0,0.9225000393390657,0.7592222428321839,2019-11-06 18:20:37 149 | 4,4,1,1,5,15,200,100,True,True,True,200,0.001,0.0,0.8645000457763667,0.6700000160932539,2019-11-06 18:21:51 150 | 4,4,5,5,5,15,200,100,False,False,False,200,0.001,0.0,0.9700000005960467,0.8845000481605528,2019-11-06 18:23:48 151 | 2,2,5,5,5,15,200,100,False,False,False,200,0.001,0.0,0.9765000063180923,0.9470000302791596,2019-11-06 15:51:11 152 | 4,4,5,5,5,15,200,100,True,False,False,200,0.001,0.0,0.9322500056028369,0.8095000445842744,2019-11-06 18:25:53 153 | 4,4,5,5,5,15,200,100,False,True,False,200,0.001,0.0,0.8869999793171881,0.7120000153779985,2019-11-06 18:26:58 154 | 4,4,5,5,5,15,200,100,False,False,True,200,0.001,0.0,0.9499999812245368,0.8410666894912718,2019-11-06 18:30:38 155 | 4,4,5,5,5,15,200,100,True,True,False,200,0.001,0.0,0.8951999774575236,0.6761333465576168,2019-11-06 18:31:51 156 | 4,4,5,5,5,15,200,100,True,False,True,200,0.001,0.0,0.8711999759078026,0.7893333524465563,2019-11-06 18:32:03 157 | 4,4,5,5,5,15,200,100,False,True,True,200,0.001,0.0,0.9526666978001596,0.8774444669485093,2019-11-06 18:34:57 158 | 4,4,5,5,5,15,200,100,True,True,True,200,0.001,0.0,0.9391667038202288,0.840777804851532,2019-11-06 18:36:21 159 | 4,4,10,10,5,15,200,100,False,False,False,200,0.001,0.0,0.9260000076889994,0.8905000454187394,2019-11-06 18:37:38 160 | 4,4,10,10,5,15,200,100,True,False,False,200,0.001,0.0,0.9442500057816511,0.8213333767652511,2019-11-06 18:39:28 161 | 4,4,10,10,5,15,200,100,False,True,False,200,0.001,0.0,0.8791999772191046,0.7406666845083236,2019-11-06 18:45:26 162 | -------------------------------------------------------------------------------- /results/old/results_TCResNet8Dilated.csv: -------------------------------------------------------------------------------- 1 | train.way,test.way,train.shot,test.shot,train.query,test.query,train.episodes,test.episodes,background,silence,unknown,epochs,lr,wd,train.acc,val.acc,timestamp 2 | 2,2,1,1,5,15,200,100,False,False,False,200,0.001,0.0,0.9530000070482489,0.925000033378601,2019-11-12 10:33:17 3 | 2,2,5,5,5,15,200,100,True,False,False,200,0.001,0.0,0.962500007748604,0.9443333625793454,2019-11-12 10:33:23 4 | 4,4,10,10,5,15,200,100,False,False,True,200,0.001,0.0,0.9391999796032905,0.8572000241279599,2019-11-12 11:55:52 5 | 4,4,10,10,5,15,200,100,True,True,False,200,0.001,0.0,0.9133999809622764,0.6942666828632355,2019-11-12 11:56:19 6 | 4,4,10,10,5,15,200,100,True,False,True,200,0.001,0.0,0.9385999807715414,0.8016000205278399,2019-11-12 11:56:37 7 | 4,4,10,10,5,15,200,100,False,True,True,200,0.001,0.0,0.9598333624005313,0.8908889102935789,2019-11-12 11:56:45 8 | 4,4,10,10,5,15,200,100,True,True,True,200,0.001,0.0,0.9375000375509257,0.8477778023481369,2019-11-12 11:57:07 9 | 4,4,15,15,5,15,200,100,False,False,False,200,0.001,0.0,0.9560000023245813,0.9110000401735305,2019-11-12 11:58:02 10 | 4,4,15,15,5,15,200,100,True,False,False,200,0.001,0.0,0.9650000014901163,0.8773333793878554,2019-11-12 12:01:03 11 | 4,4,15,15,5,15,200,100,False,True,False,200,0.001,0.0,0.93279998242855,0.785600019097328,2019-11-12 12:04:46 12 | 4,4,15,15,5,15,200,100,False,False,True,200,0.001,0.0,0.9595999845862385,0.8752000206708908,2019-11-12 12:05:25 13 | 4,4,15,15,5,15,200,100,True,True,False,200,0.001,0.0,0.9183999803662298,0.7076000159978866,2019-11-12 12:06:56 14 | 2,2,5,5,5,15,200,100,False,True,False,200,0.001,0.0,0.9420000134408476,0.7966666784882548,2019-11-12 10:33:23 15 | 4,4,15,15,5,15,200,100,True,False,True,200,0.001,0.0,0.9159999787807468,0.8177333539724351,2019-11-12 12:06:56 16 | 4,4,15,15,5,15,200,100,False,True,True,200,0.001,0.0,0.9710000240802765,0.9052222430706026,2019-11-12 12:07:36 17 | 4,4,15,15,5,15,200,100,True,True,True,200,0.001,0.0,0.9461667013168331,0.8495555812120438,2019-11-12 12:07:42 18 | 4,4,20,20,5,15,200,100,False,False,False,200,0.001,0.0,0.9652499997615814,0.9046667098999022,2019-11-12 12:09:47 19 | 4,4,20,20,5,15,200,100,True,False,False,200,0.001,0.0,0.978000001013279,0.8811667120456698,2019-11-12 12:11:21 20 | 4,4,20,20,5,15,200,100,False,True,False,200,0.001,0.0,0.9413999897241592,0.7653333479166031,2019-11-12 12:12:25 21 | 4,4,20,20,5,15,200,100,False,False,True,200,0.001,0.0,0.9511999830603601,0.8900000226497653,2019-11-12 12:13:13 22 | 4,4,20,20,5,15,200,100,True,True,False,200,0.001,0.0,0.9099999809265132,0.7252000153064727,2019-11-12 12:14:11 23 | 4,4,20,20,5,15,200,100,True,False,True,200,0.001,0.0,0.9125999775528907,0.8325333511829375,2019-11-12 12:14:53 24 | 4,4,20,20,5,15,200,100,False,True,True,200,0.001,0.0,0.9591666963696478,0.9011111301183701,2019-11-12 12:14:59 25 | 2,2,5,5,5,15,200,100,False,False,True,200,0.001,0.0,0.9593333607912062,0.8848888957500458,2019-11-12 10:33:23 26 | 4,4,20,20,5,15,200,100,True,True,True,200,0.001,0.0,0.9441667029261587,0.8651111364364623,2019-11-12 12:15:46 27 | 5,5,1,1,5,15,200,100,False,False,False,200,0.001,0.0,0.8921999797224999,0.6969333472847936,2019-11-12 12:17:38 28 | 5,5,1,1,5,15,200,100,True,False,False,200,0.001,0.0,0.8621999783813954,0.6193333485722543,2019-11-12 12:21:13 29 | 5,5,1,1,5,15,200,100,False,True,False,200,0.001,0.0,0.85850004196167,0.5757777875661849,2019-11-12 12:27:29 30 | 5,5,1,1,5,15,200,100,False,False,True,200,0.001,0.0,0.8928333780169484,0.6571111249923705,2019-11-12 12:27:40 31 | 5,5,1,1,5,15,200,100,True,True,False,200,0.001,0.0,0.82616671025753,0.5272222307324407,2019-11-12 12:28:30 32 | 5,5,1,1,5,15,200,100,True,False,True,200,0.001,0.0,0.6961667072772985,0.5581111228466032,2019-11-12 12:29:04 33 | 5,5,1,1,5,15,200,100,False,True,True,200,0.001,0.0,0.8810000070929526,0.7209523975849152,2019-11-12 12:29:21 34 | 5,5,1,1,5,15,200,100,True,True,True,200,0.001,0.0,0.7584285856783389,0.5904762130975725,2019-11-12 12:31:18 35 | 5,5,5,5,5,15,200,100,False,False,False,200,0.001,0.0,0.9119999778270718,0.8817333525419239,2019-11-12 12:31:44 36 | 2,2,5,5,5,15,200,100,True,True,False,200,0.001,0.0,0.9443333524465561,0.7804444566369058,2019-11-12 10:33:23 37 | 5,5,5,5,5,15,200,100,True,False,False,200,0.001,0.0,0.9199999773502351,0.803066684603691,2019-11-12 12:32:06 38 | 5,5,5,5,5,15,200,100,False,True,False,200,0.001,0.0,0.9075000366568565,0.7033333504199981,2019-11-12 12:32:44 39 | 5,5,5,5,5,15,200,100,False,False,True,200,0.001,0.0,0.9521667003631592,0.8506666898727417,2019-11-12 12:35:42 40 | 5,5,5,5,5,15,200,100,True,True,False,200,0.001,0.0,0.8930000445246694,0.6544444519281387,2019-11-12 12:36:25 41 | 5,5,5,5,5,15,200,100,True,False,True,200,0.001,0.0,0.8650000482797628,0.743777800798416,2019-11-12 12:36:56 42 | 5,5,5,5,5,15,200,100,False,True,True,200,0.001,0.0,0.9428571462631219,0.8596190601587296,2019-11-12 12:37:19 43 | 5,5,5,5,5,15,200,100,True,True,True,200,0.001,0.0,0.925428575873375,0.7911428695917131,2019-11-12 12:37:22 44 | 5,5,10,10,5,15,200,100,False,False,False,200,0.001,0.0,0.934599980413914,0.8757333558797834,2019-11-12 12:37:48 45 | 5,5,10,10,5,15,200,100,True,False,False,200,0.001,0.0,0.9411999812722207,0.842933350801468,2019-11-12 12:40:37 46 | 5,5,10,10,5,15,200,100,False,True,False,200,0.001,0.0,0.9238333684206012,0.7468889170885086,2019-11-12 12:42:28 47 | 2,2,5,5,5,15,200,100,True,False,True,200,0.001,0.0,0.9296667057275771,0.8248888990283008,2019-11-12 10:33:23 48 | 5,5,10,10,5,15,200,100,False,False,True,200,0.001,0.0,0.943833371698856,0.8727778005599975,2019-11-12 12:42:40 49 | 5,5,10,10,5,15,200,100,True,True,False,200,0.001,0.0,0.9063333693146705,0.6775555676221846,2019-11-12 12:43:20 50 | 5,5,10,10,5,15,200,100,True,False,True,200,0.001,0.0,0.9215000441670421,0.7856666892766956,2019-11-12 12:43:52 51 | 5,5,10,10,5,15,200,100,False,True,True,200,0.001,0.0,0.9514285743236537,0.883047633767128,2019-11-12 12:48:30 52 | 5,5,10,10,5,15,200,100,True,True,True,200,0.001,0.0,0.9242857187986375,0.8153333443403246,2019-11-12 12:48:33 53 | 5,5,15,15,5,15,200,100,False,False,False,200,0.001,0.0,0.9549999830126762,0.8674666851758955,2019-11-12 12:48:58 54 | 5,5,15,15,5,15,200,100,True,False,False,200,0.001,0.0,0.9189999788999553,0.8209333562850955,2019-11-12 12:50:17 55 | 5,5,15,15,5,15,200,100,False,True,False,200,0.001,0.0,0.8920000419020649,0.7254444670677188,2019-11-12 12:50:19 56 | 5,5,15,15,5,15,200,100,False,False,True,200,0.001,0.0,0.9446667039394379,0.8646666920185091,2019-11-12 12:51:05 57 | 5,5,15,15,5,15,200,100,True,True,False,200,0.001,0.0,0.8696667119860649,0.6983333450555803,2019-11-12 12:51:09 58 | 2,2,5,5,5,15,200,100,False,True,True,200,0.001,0.0,0.9662500008940695,0.9211667144298556,2019-11-12 10:33:23 59 | 5,5,15,15,5,15,200,100,True,False,True,200,0.001,0.0,0.9175000432133672,0.8143333613872527,2019-11-12 12:51:28 60 | 5,5,15,15,5,15,200,100,False,True,True,200,0.001,0.0,0.9392857179045677,0.8881905025243758,2019-11-12 12:51:54 61 | 5,5,15,15,5,15,200,100,True,True,True,200,0.001,0.0,0.9311428612470624,0.8492381066083909,2019-11-12 12:55:20 62 | 5,5,20,20,5,15,200,100,False,False,False,200,0.001,0.0,0.9527999812364578,0.8688000166416168,2019-11-12 12:55:22 63 | 5,5,20,20,5,15,200,100,True,False,False,200,0.001,0.0,0.9327999788522718,0.8521333551406863,2019-11-12 12:55:24 64 | 5,5,20,20,5,15,200,100,False,True,False,200,0.001,0.0,0.9186667013168331,0.7373333555459977,2019-11-12 13:00:39 65 | 5,5,20,20,5,15,200,100,False,False,True,200,0.001,0.0,0.9465000355243682,0.8577777999639508,2019-11-12 13:02:00 66 | 5,5,20,20,5,15,200,100,True,True,False,200,0.001,0.0,0.9178333699703214,0.6964444595575332,2019-11-12 13:04:10 67 | 5,5,20,20,5,15,200,100,True,False,True,200,0.001,0.0,0.9508333674073217,0.8292222464084624,2019-11-12 13:04:23 68 | 5,5,20,20,5,15,200,100,False,True,True,200,0.001,0.0,0.9577142882347105,0.8880000221729277,2019-11-12 13:07:37 69 | 2,2,5,5,5,15,200,100,True,True,True,200,0.001,0.0,0.9455000054836268,0.8711667114496229,2019-11-12 10:33:23 70 | 5,5,20,20,5,15,200,100,True,True,True,200,0.001,0.0,0.9302857184410095,0.8296190577745435,2019-11-12 13:09:34 71 | 2,2,10,10,5,15,200,100,False,False,False,200,0.001,0.0,0.9710000061988828,0.9543333607912062,2019-11-12 10:43:42 72 | 2,2,10,10,5,15,200,100,True,False,False,200,0.001,0.0,0.9655000071227553,0.9290000355243684,2019-11-12 10:43:44 73 | 2,2,10,10,5,15,200,100,False,True,False,200,0.001,0.0,0.9630000120401383,0.8175555652379989,2019-11-12 10:43:47 74 | 2,2,1,1,5,15,200,100,True,False,False,200,0.001,0.0,0.8780000095814464,0.8170000416040422,2019-11-12 10:33:16 75 | 2,2,10,10,5,15,200,100,False,False,True,200,0.001,0.0,0.9500000327825546,0.9086666721105575,2019-11-12 10:43:52 76 | 2,2,10,10,5,15,200,100,True,True,False,200,0.001,0.0,0.9593333441019057,0.7913333448767663,2019-11-12 10:43:55 77 | 2,2,10,10,5,15,200,100,True,False,True,200,0.001,0.0,0.9463333657383916,0.8555555641651156,2019-11-12 10:43:55 78 | 2,2,10,10,5,15,200,100,False,True,True,200,0.001,0.0,0.9627500018477445,0.9335000443458554,2019-11-12 10:44:29 79 | 2,2,10,10,5,15,200,100,True,True,True,200,0.001,0.0,0.9587500026822093,0.8931667131185531,2019-11-12 10:44:35 80 | 2,2,15,15,5,15,200,100,False,False,False,200,0.001,0.0,0.978000006079674,0.9646666878461836,2019-11-12 10:44:35 81 | 2,2,15,15,5,15,200,100,True,False,False,200,0.001,0.0,0.9735000067949295,0.9416667008399966,2019-11-12 10:45:12 82 | 2,2,15,15,5,15,200,100,False,True,False,200,0.001,0.0,0.9600000101327892,0.8291111209988595,2019-11-12 10:45:17 83 | 2,2,15,15,5,15,200,100,False,False,True,200,0.001,0.0,0.958333359956742,0.9008888947963712,2019-11-12 10:45:24 84 | 2,2,15,15,5,15,200,100,True,True,False,200,0.001,0.0,0.95533334851265,0.7931111225485803,2019-11-12 10:45:33 85 | 2,2,1,1,5,15,200,100,False,True,False,200,0.001,0.0,0.8563333719968793,0.7193333479762075,2019-11-12 10:33:16 86 | 2,2,15,15,5,15,200,100,True,False,True,200,0.001,0.0,0.952333368957043,0.8631111192703246,2019-11-12 10:45:42 87 | 2,2,15,15,5,15,200,100,False,True,True,200,0.001,0.0,0.9540000015497206,0.9300000429153444,2019-11-12 10:46:22 88 | 2,2,15,15,5,15,200,100,True,True,True,200,0.001,0.0,0.9525000035762787,0.8930000424385074,2019-11-12 10:47:11 89 | 2,2,20,20,5,15,200,100,False,False,False,200,0.001,0.0,0.9820000037550926,0.969000023007393,2019-11-12 10:53:21 90 | 2,2,20,20,5,15,200,100,True,False,False,200,0.001,0.0,0.9690000066161161,0.9393333697319031,2019-11-12 10:53:23 91 | 2,2,20,20,5,15,200,100,False,True,False,200,0.001,0.0,0.9503333470225336,0.8224444550275802,2019-11-12 10:53:34 92 | 2,2,20,20,5,15,200,100,False,False,True,200,0.001,0.0,0.9676666915416717,0.9182222270965577,2019-11-12 10:53:43 93 | 2,2,20,20,5,15,200,100,True,True,False,200,0.001,0.0,0.954666682481766,0.7906666785478594,2019-11-12 10:53:47 94 | 2,2,20,20,5,15,200,100,True,False,True,200,0.001,0.0,0.949333366155624,0.8417777872085572,2019-11-12 10:53:49 95 | 2,2,20,20,5,15,200,100,False,True,True,200,0.001,0.0,0.9630000001192092,0.9290000408887863,2019-11-12 10:53:53 96 | 2,2,1,1,5,15,200,100,False,False,True,200,0.001,0.0,0.9363333624601365,0.8093333441019059,2019-11-12 10:33:16 97 | 2,2,20,20,5,15,200,100,True,True,True,200,0.001,0.0,0.972999999523163,0.9093333745002745,2019-11-12 10:54:17 98 | 3,3,1,1,5,15,200,100,False,False,False,200,0.001,0.0,0.9330000299215315,0.8320000094175339,2019-11-12 10:58:40 99 | 3,3,1,1,5,15,200,100,True,False,False,200,0.001,0.0,0.9376666989922525,0.7611111241579058,2019-11-12 10:58:47 100 | 3,3,1,1,5,15,200,100,False,True,False,200,0.001,0.0,0.876000007688999,0.6596667033433912,2019-11-12 10:58:59 101 | 3,3,1,1,5,15,200,100,False,False,True,200,0.001,0.0,0.890000008046627,0.7466667082905768,2019-11-12 10:59:07 102 | 3,3,1,1,5,15,200,100,True,True,False,200,0.001,0.0,0.7742500113695864,0.5855000317096712,2019-11-12 11:02:46 103 | 3,3,1,1,5,15,200,100,True,False,True,200,0.001,0.0,0.8325000122189519,0.649333369731903,2019-11-12 11:02:48 104 | 3,3,1,1,5,15,200,100,False,True,True,200,0.001,0.0,0.9005999800562859,0.8042666870355607,2019-11-12 11:02:58 105 | 3,3,1,1,5,15,200,100,True,True,True,200,0.001,0.0,0.8635999771952628,0.7118666821718214,2019-11-12 11:03:23 106 | 3,3,5,5,5,15,200,100,False,False,False,200,0.001,0.0,0.9720000210404404,0.9157777827978133,2019-11-12 11:06:39 107 | 2,2,1,1,5,15,200,100,True,True,False,200,0.001,0.0,0.7136667056381705,0.6320000180602074,2019-11-12 10:33:17 108 | 3,3,5,5,5,15,200,100,True,False,False,200,0.001,0.0,0.9470000338554385,0.8684444522857667,2019-11-12 11:07:32 109 | 3,3,5,5,5,15,200,100,False,True,False,200,0.001,0.0,0.9182500049471856,0.758833374381065,2019-11-12 11:07:35 110 | 3,3,5,5,5,15,200,100,False,False,True,200,0.001,0.0,0.9442500045895579,0.8520000439882278,2019-11-12 11:08:04 111 | 3,3,5,5,5,15,200,100,True,True,False,200,0.001,0.0,0.9077500048279763,0.7298333731293676,2019-11-12 11:08:11 112 | 3,3,5,5,5,15,200,100,True,False,True,200,0.001,0.0,0.927500008046627,0.7705000424385072,2019-11-12 11:08:28 113 | 3,3,5,5,5,15,200,100,False,True,True,200,0.001,0.0,0.9479999828338626,0.9060000228881836,2019-11-12 11:11:08 114 | 3,3,5,5,5,15,200,100,True,True,True,200,0.001,0.0,0.9459999826550483,0.8370666855573653,2019-11-12 11:11:40 115 | 3,3,10,10,5,15,200,100,False,False,False,200,0.001,0.0,0.9720000192522997,0.9288888931274415,2019-11-12 11:12:01 116 | 3,3,10,10,5,15,200,100,True,False,False,200,0.001,0.0,0.9653333559632297,0.9111111164093021,2019-11-12 11:12:38 117 | 3,3,10,10,5,15,200,100,False,True,False,200,0.001,0.0,0.9507500010728834,0.8005000418424609,2019-11-12 11:12:40 118 | 2,2,1,1,5,15,200,100,True,False,True,200,0.001,0.0,0.8490000425279144,0.6942222374677658,2019-11-12 10:33:16 119 | 3,3,10,10,5,15,200,100,False,False,True,200,0.001,0.0,0.9482500028610227,0.8851667135953902,2019-11-12 11:13:09 120 | 3,3,10,10,5,15,200,100,True,True,False,200,0.001,0.0,0.9385000020265576,0.7500000429153438,2019-11-12 11:13:14 121 | 3,3,10,10,5,15,200,100,True,False,True,200,0.001,0.0,0.9275000068545343,0.8225000447034838,2019-11-12 11:14:05 122 | 3,3,10,10,5,15,200,100,False,True,True,200,0.001,0.0,0.9667999860644343,0.9164000248908996,2019-11-12 11:16:50 123 | 3,3,10,10,5,15,200,100,True,True,True,200,0.001,0.0,0.943199980556965,0.8556000185012818,2019-11-12 11:16:55 124 | 3,3,15,15,5,15,200,100,False,False,False,200,0.001,0.0,0.9740000194311141,0.9415555590391161,2019-11-12 11:17:21 125 | 3,3,15,15,5,15,200,100,True,False,False,200,0.001,0.0,0.9560000270605088,0.9124444496631623,2019-11-12 11:17:25 126 | 3,3,15,15,5,15,200,100,False,True,False,200,0.001,0.0,0.951000000536442,0.8068333750963212,2019-11-12 11:18:56 127 | 3,3,15,15,5,15,200,100,False,False,True,200,0.001,0.0,0.9744999998807908,0.8976667076349261,2019-11-12 11:19:53 128 | 3,3,15,15,5,15,200,100,True,True,False,200,0.001,0.0,0.9407500010728838,0.7615000396966936,2019-11-12 11:21:56 129 | 2,2,1,1,5,15,200,100,False,True,True,200,0.001,0.0,0.9170000061392787,0.8421667087078093,2019-11-12 10:33:16 130 | 3,3,15,15,5,15,200,100,True,False,True,200,0.001,0.0,0.929750007092953,0.8248333775997162,2019-11-12 11:22:03 131 | 3,3,15,15,5,15,200,100,False,True,True,200,0.001,0.0,0.9561999815702439,0.9177333575487135,2019-11-12 11:23:01 132 | 3,3,15,15,5,15,200,100,True,True,True,200,0.001,0.0,0.940199980437756,0.8760000181198122,2019-11-12 11:24:22 133 | 3,3,20,20,5,15,200,100,False,False,False,200,0.001,0.0,0.9730000182986257,0.9295555597543721,2019-11-12 11:27:40 134 | 3,3,20,20,5,15,200,100,True,False,False,200,0.001,0.0,0.9690000233054159,0.8928888952732087,2019-11-12 11:27:42 135 | 3,3,20,20,5,15,200,100,False,True,False,200,0.001,0.0,0.9515000015497208,0.7926667088270188,2019-11-12 11:29:31 136 | 3,3,20,20,5,15,200,100,False,False,True,200,0.001,0.0,0.9462500050663952,0.8970000433921818,2019-11-12 11:29:39 137 | 3,3,20,20,5,15,200,100,True,True,False,200,0.001,0.0,0.9107500049471856,0.7351667088270188,2019-11-12 11:30:22 138 | 3,3,20,20,5,15,200,100,True,False,True,200,0.001,0.0,0.9340000042319295,0.8218333804607391,2019-11-12 11:30:31 139 | 3,3,20,20,5,15,200,100,False,True,True,200,0.001,0.0,0.9629999852180481,0.9110666948556901,2019-11-12 11:30:43 140 | 2,2,1,1,5,15,200,100,True,True,True,200,0.001,0.0,0.8417500123381612,0.7633333751559258,2019-11-12 10:33:16 141 | 3,3,20,20,5,15,200,100,True,True,True,200,0.001,0.0,0.9525999829173086,0.871866688132286,2019-11-12 11:31:47 142 | 4,4,1,1,5,15,200,100,False,False,False,200,0.001,0.0,0.9495000019669535,0.7881667089462279,2019-11-12 11:32:34 143 | 4,4,1,1,5,15,200,100,True,False,False,200,0.001,0.0,0.8495000103116034,0.6896667075157167,2019-11-12 11:33:03 144 | 4,4,1,1,5,15,200,100,False,True,False,200,0.001,0.0,0.8147999791800973,0.6436000141501428,2019-11-12 11:39:07 145 | 4,4,1,1,5,15,200,100,False,False,True,200,0.001,0.0,0.8921999788284298,0.7066666841506959,2019-11-12 11:40:48 146 | 4,4,1,1,5,15,200,100,True,True,False,200,0.001,0.0,0.847199978083372,0.5649333447217944,2019-11-12 11:41:20 147 | 4,4,1,1,5,15,200,100,True,False,True,200,0.001,0.0,0.8531999757885932,0.5934666773676873,2019-11-12 11:41:52 148 | 4,4,1,1,5,15,200,100,False,True,True,200,0.001,0.0,0.897333376705647,0.75466668844223,2019-11-12 11:42:32 149 | 4,4,1,1,5,15,200,100,True,True,True,200,0.001,0.0,0.8660000449419019,0.6492222321033482,2019-11-12 11:42:35 150 | 4,4,5,5,5,15,200,100,False,False,False,200,0.001,0.0,0.9400000035762788,0.8990000456571577,2019-11-12 11:45:58 151 | 2,2,5,5,5,15,200,100,False,False,False,200,0.001,0.0,0.9720000067353254,0.9626666903495786,2019-11-12 10:33:23 152 | 4,4,5,5,5,15,200,100,True,False,False,200,0.001,0.0,0.9629999998211863,0.8383333790302275,2019-11-12 11:46:02 153 | 4,4,5,5,5,15,200,100,False,True,False,200,0.001,0.0,0.9059999811649321,0.7224000126123429,2019-11-12 11:46:06 154 | 4,4,5,5,5,15,200,100,False,False,True,200,0.001,0.0,0.938599979579449,0.8453333544731139,2019-11-12 11:46:30 155 | 4,4,5,5,5,15,200,100,True,True,False,200,0.001,0.0,0.9225999817252158,0.7148000168800355,2019-11-12 11:51:32 156 | 4,4,5,5,5,15,200,100,True,False,True,200,0.001,0.0,0.8707999745011328,0.7782666867971422,2019-11-12 11:51:50 157 | 4,4,5,5,5,15,200,100,False,True,True,200,0.001,0.0,0.9375000396370884,0.88922224342823,2019-11-12 11:52:01 158 | 4,4,5,5,5,15,200,100,True,True,True,200,0.001,0.0,0.9316667062044142,0.8301111352443693,2019-11-12 11:53:10 159 | 4,4,10,10,5,15,200,100,False,False,False,200,0.001,0.0,0.9507500049471859,0.9003333801031116,2019-11-12 11:53:45 160 | 4,4,10,10,5,15,200,100,True,False,False,200,0.001,0.0,0.9617500010132792,0.8460000461339949,2019-11-12 11:55:17 161 | 4,4,10,10,5,15,200,100,False,True,False,200,0.001,0.0,0.9121999830007553,0.7552000141143801,2019-11-12 11:55:45 162 | -------------------------------------------------------------------------------- /results/old/results_TCResNetDilated_1Stride.csv: -------------------------------------------------------------------------------- 1 | train.way,test.way,train.shot,test.shot,train.query,test.query,train.episodes,test.episodes,background,silence,unknown,epochs,lr,wd,train.acc,val.acc,timestamp 2 | 2,2,1,1,5,15,200,100,False,False,False,200,0.001,0.0,0.962500005811453,0.9273333680629732,2019-11-18 13:06:22 3 | 2,2,5,5,5,15,200,100,True,False,False,200,0.001,0.0,0.971000008583069,0.9613333588838577,2019-11-18 13:06:25 4 | 4,4,10,10,5,15,200,100,False,False,True,200,0.001,0.0,0.9629999846220018,0.8774666869640354,2019-11-18 14:36:17 5 | 4,4,10,10,5,15,200,100,True,True,False,200,0.001,0.0,0.9121999830007554,0.7544000160694123,2019-11-18 14:37:56 6 | 4,4,10,10,5,15,200,100,True,False,True,200,0.001,0.0,0.9555999818444254,0.865066688656807,2019-11-18 14:39:22 7 | 4,4,10,10,5,15,200,100,False,True,True,200,0.001,0.0,0.9675000271201131,0.9280000203847885,2019-11-18 14:40:04 8 | 4,4,10,10,5,15,200,100,True,True,True,200,0.001,0.0,0.9650000262260442,0.9071111315488816,2019-11-18 14:40:46 9 | 4,4,15,15,5,15,200,100,False,False,False,200,0.001,0.0,0.9864999997615815,0.9341667145490645,2019-11-18 14:41:37 10 | 4,4,15,15,5,15,200,100,True,False,False,200,0.001,0.0,0.9797500002384187,0.9065000474452973,2019-11-18 14:45:59 11 | 4,4,15,15,5,15,200,100,False,True,False,200,0.001,0.0,0.9343999880552291,0.8018666857481,2019-11-18 14:48:19 12 | 4,4,15,15,5,15,200,100,False,False,True,200,0.001,0.0,0.9531999817490576,0.9270666921138764,2019-11-18 14:48:22 13 | 4,4,15,15,5,15,200,100,True,True,False,200,0.001,0.0,0.9339999878406526,0.7706666827201842,2019-11-18 14:48:22 14 | 2,2,5,5,5,15,200,100,False,True,False,200,0.001,0.0,0.9663333418965339,0.8324444535374641,2019-11-18 13:06:25 15 | 4,4,15,15,5,15,200,100,True,False,True,200,0.001,0.0,0.9433999806642531,0.8842666894197464,2019-11-18 14:48:26 16 | 4,4,15,15,5,15,200,100,False,True,True,200,0.001,0.0,0.9650000250339507,0.9280000191926956,2019-11-18 14:49:46 17 | 4,4,15,15,5,15,200,100,True,True,True,200,0.001,0.0,0.9560000309348107,0.9011111313104632,2019-11-18 14:49:55 18 | 4,4,20,20,5,15,200,100,False,False,False,200,0.001,0.0,0.9889999994635578,0.9481667119264603,2019-11-18 14:50:33 19 | 4,4,20,20,5,15,200,100,True,False,False,200,0.001,0.0,0.9864999979734419,0.913500044345856,2019-11-18 14:51:06 20 | 4,4,20,20,5,15,200,100,False,True,False,200,0.001,0.0,0.9175999823212624,0.7890666842460636,2019-11-18 14:53:33 21 | 4,4,20,20,5,15,200,100,False,False,True,200,0.001,0.0,0.9601999837160107,0.9006666880846023,2019-11-18 14:54:20 22 | 4,4,20,20,5,15,200,100,True,True,False,200,0.001,0.0,0.933199987113476,0.782266684174538,2019-11-18 14:54:20 23 | 4,4,20,20,5,15,200,100,True,False,True,200,0.001,0.0,0.953199981749058,0.884533352255821,2019-11-18 15:01:35 24 | 4,4,20,20,5,15,200,100,False,True,True,200,0.001,0.0,0.9740000221133236,0.9266666859388355,2019-11-18 15:02:05 25 | 2,2,5,5,5,15,200,100,False,False,True,200,0.001,0.0,0.9583333599567415,0.927555559873581,2019-11-18 13:06:25 26 | 4,4,20,20,5,15,200,100,True,True,True,200,0.001,0.0,0.934666707515717,0.8891111350059508,2019-11-18 15:03:03 27 | 5,5,1,1,5,15,200,100,False,False,False,200,0.001,0.0,0.9463999825716018,0.7344000157713888,2019-11-18 15:06:48 28 | 5,5,1,1,5,15,200,100,True,False,False,200,0.001,0.0,0.8295999769866467,0.7086666795611382,2019-11-18 15:07:29 29 | 5,5,1,1,5,15,200,100,False,True,False,200,0.001,0.0,0.8836667077243329,0.6236666780710218,2019-11-18 15:09:20 30 | 5,5,1,1,5,15,200,100,False,False,True,200,0.001,0.0,0.8280000473558907,0.6910000178217888,2019-11-18 15:09:31 31 | 5,5,1,1,5,15,200,100,True,True,False,200,0.001,0.0,0.8693333765864371,0.5967777866125108,2019-11-18 15:11:15 32 | 5,5,1,1,5,15,200,100,True,False,True,200,0.001,0.0,0.8850000455975536,0.6467777910828588,2019-11-18 15:16:46 33 | 5,5,1,1,5,15,200,100,False,True,True,200,0.001,0.0,0.9530000028014181,0.781619062423706,2019-11-18 15:17:48 34 | 5,5,1,1,5,15,200,100,True,True,True,200,0.001,0.0,0.8865714353322982,0.7159047788381576,2019-11-18 15:18:27 35 | 5,5,5,5,5,15,200,100,False,False,False,200,0.001,0.0,0.9641999858617786,0.8817333549261093,2019-11-18 15:18:31 36 | 2,2,5,5,5,15,200,100,True,True,False,200,0.001,0.0,0.9633333432674406,0.818000010550022,2019-11-18 13:06:25 37 | 5,5,5,5,5,15,200,100,True,False,False,200,0.001,0.0,0.9501999834179878,0.8546666872501372,2019-11-18 15:18:32 38 | 5,5,5,5,5,15,200,100,False,True,False,200,0.001,0.0,0.9308333662152289,0.7500000244379043,2019-11-18 15:19:26 39 | 5,5,5,5,5,15,200,100,False,False,True,200,0.001,0.0,0.9468333700299267,0.8834444653987886,2019-11-18 15:19:26 40 | 5,5,5,5,5,15,200,100,True,True,False,200,0.001,0.0,0.9293333640694619,0.7182222396135329,2019-11-18 15:19:33 41 | 5,5,5,5,5,15,200,100,True,False,True,200,0.001,0.0,0.9233333760499954,0.8174444711208345,2019-11-18 15:19:38 42 | 5,5,5,5,5,15,200,100,False,True,True,200,0.001,0.0,0.9225714331865312,0.8779047787189485,2019-11-18 15:20:19 43 | 5,5,5,5,5,15,200,100,True,True,True,200,0.001,0.0,0.934571432471275,0.8510476303100587,2019-11-18 15:20:26 44 | 5,5,10,10,5,15,200,100,False,False,False,200,0.001,0.0,0.964599986076355,0.9112000274658206,2019-11-18 15:21:12 45 | 5,5,10,10,5,15,200,100,True,False,False,200,0.001,0.0,0.9717999887466428,0.8910666877031328,2019-11-18 15:22:28 46 | 5,5,10,10,5,15,200,100,False,True,False,200,0.001,0.0,0.8990000411868099,0.7624444723129272,2019-11-18 15:23:23 47 | 2,2,5,5,5,15,200,100,True,False,True,200,0.001,0.0,0.9576666945219041,0.8851111179590225,2019-11-18 13:06:25 48 | 5,5,10,10,5,15,200,100,False,False,True,200,0.001,0.0,0.9648333597183224,0.9120000219345095,2019-11-18 15:23:43 49 | 5,5,10,10,5,15,200,100,True,True,False,200,0.001,0.0,0.915666700899601,0.7343333536386493,2019-11-18 15:25:55 50 | 5,5,10,10,5,15,200,100,True,False,True,200,0.001,0.0,0.9540000334382055,0.8691111320257188,2019-11-18 15:31:20 51 | 5,5,10,10,5,15,200,100,False,True,True,200,0.001,0.0,0.9612857165932657,0.9050476533174514,2019-11-18 15:32:01 52 | 5,5,10,10,5,15,200,100,True,True,True,200,0.001,0.0,0.9490000030398368,0.8704762035608289,2019-11-18 15:32:24 53 | 5,5,15,15,5,15,200,100,False,False,False,200,0.001,0.0,0.97839998960495,0.9380000263452533,2019-11-18 15:32:48 54 | 5,5,15,15,5,15,200,100,True,False,False,200,0.001,0.0,0.9485999807715413,0.8874666881561281,2019-11-18 15:35:39 55 | 5,5,15,15,5,15,200,100,False,True,False,200,0.001,0.0,0.9425000235438351,0.7604444694519042,2019-11-18 15:36:55 56 | 5,5,15,15,5,15,200,100,False,False,True,200,0.001,0.0,0.9523333665728573,0.900888911485672,2019-11-18 15:37:04 57 | 5,5,15,15,5,15,200,100,True,True,False,200,0.001,0.0,0.9166667011380197,0.7428889143466952,2019-11-18 15:37:18 58 | 2,2,5,5,5,15,200,100,False,True,True,200,0.001,0.0,0.965250001549721,0.9448333770036696,2019-11-18 13:06:25 59 | 5,5,15,15,5,15,200,100,True,False,True,200,0.001,0.0,0.9693333572149276,0.861888911128044,2019-11-18 15:37:40 60 | 5,5,15,15,5,15,200,100,False,True,True,200,0.001,0.0,0.9670000019669532,0.9009524071216584,2019-11-18 15:38:12 61 | 5,5,15,15,5,15,200,100,True,True,True,200,0.001,0.0,0.9594285738468171,0.8866666859388355,2019-11-18 15:38:33 62 | 5,5,20,20,5,15,200,100,False,False,False,200,0.001,0.0,0.9863999930024148,0.9241333574056624,2019-11-18 15:38:48 63 | 5,5,20,20,5,15,200,100,True,False,False,200,0.001,0.0,0.9705999869108201,0.8994666868448256,2019-11-18 15:39:11 64 | 5,5,20,20,5,15,200,100,False,True,False,200,0.001,0.0,0.9446666917204855,0.7656666946411131,2019-11-18 15:42:43 65 | 5,5,20,20,5,15,200,100,False,False,True,200,0.001,0.0,0.9600000303983691,0.904888907074928,2019-11-18 15:42:58 66 | 5,5,20,20,5,15,200,100,True,True,False,200,0.001,0.0,0.9276667010784151,0.7510000252723698,2019-11-18 15:47:34 67 | 5,5,20,20,5,15,200,100,True,False,True,200,0.001,0.0,0.9686666911840444,0.8548889100551604,2019-11-18 15:52:41 68 | 5,5,20,20,5,15,200,100,False,True,True,200,0.001,0.0,0.966714287698269,0.9102857500314712,2019-11-18 15:54:38 69 | 2,2,5,5,5,15,200,100,True,True,True,200,0.001,0.0,0.9687499994039537,0.9163333773612976,2019-11-18 13:16:54 70 | 5,5,20,20,5,15,200,100,True,True,True,200,0.001,0.0,0.942285717725754,0.8937143099308016,2019-11-18 15:55:11 71 | 2,2,10,10,5,15,200,100,False,False,False,200,0.001,0.0,0.988500003516674,0.9666666913032529,2019-11-18 13:17:11 72 | 2,2,10,10,5,15,200,100,True,False,False,200,0.001,0.0,0.9855000045895579,0.9656666928529744,2019-11-18 13:17:16 73 | 2,2,10,10,5,15,200,100,False,True,False,200,0.001,0.0,0.9553333443403242,0.8502222308516503,2019-11-18 13:17:19 74 | 2,2,1,1,5,15,200,100,True,False,False,200,0.001,0.0,0.9595000082254413,0.8986667090654377,2019-11-18 13:06:22 75 | 2,2,10,10,5,15,200,100,False,False,True,200,0.001,0.0,0.9720000228285789,0.9348888927698137,2019-11-18 13:17:20 76 | 2,2,10,10,5,15,200,100,True,True,False,200,0.001,0.0,0.9483333545923232,0.8262222319841386,2019-11-18 13:17:23 77 | 2,2,10,10,5,15,200,100,True,False,True,200,0.001,0.0,0.9550000327825547,0.915111116170883,2019-11-18 13:17:29 78 | 2,2,10,10,5,15,200,100,False,True,True,200,0.001,0.0,0.9744999992847443,0.9526667112112044,2019-11-18 13:17:35 79 | 2,2,10,10,5,15,200,100,True,True,True,200,0.001,0.0,0.9592500019073489,0.9220000427961349,2019-11-18 13:17:36 80 | 2,2,15,15,5,15,200,100,False,False,False,200,0.001,0.0,0.9865000045299529,0.9756666862964629,2019-11-18 13:17:36 81 | 2,2,15,15,5,15,200,100,True,False,False,200,0.001,0.0,0.9815000048279767,0.9693333554267882,2019-11-18 13:17:38 82 | 2,2,15,15,5,15,200,100,False,True,False,200,0.001,0.0,0.9523333442211155,0.8446666759252548,2019-11-18 13:18:00 83 | 2,2,15,15,5,15,200,100,False,False,True,200,0.001,0.0,0.962333360910416,0.9340000039339067,2019-11-18 13:18:10 84 | 2,2,15,15,5,15,200,100,True,True,False,200,0.001,0.0,0.9460000181198122,0.8188888996839524,2019-11-18 13:19:42 85 | 2,2,1,1,5,15,200,100,False,True,False,200,0.001,0.0,0.8766667032241818,0.7597777906060217,2019-11-18 13:06:22 86 | 2,2,15,15,5,15,200,100,True,False,True,200,0.001,0.0,0.951333364844322,0.9217777824401855,2019-11-18 13:20:03 87 | 2,2,15,15,5,15,200,100,False,True,True,200,0.001,0.0,0.9692500004172327,0.9498333775997164,2019-11-18 13:26:05 88 | 2,2,15,15,5,15,200,100,True,True,True,200,0.001,0.0,0.9685000014305111,0.9263333761692047,2019-11-18 13:26:23 89 | 2,2,20,20,5,15,200,100,False,False,False,200,0.001,0.0,0.985000005066395,0.9750000190734864,2019-11-18 13:26:23 90 | 2,2,20,20,5,15,200,100,True,False,False,200,0.001,0.0,0.9810000044107436,0.9730000227689739,2019-11-18 13:26:29 91 | 2,2,20,20,5,15,200,100,False,True,False,200,0.001,0.0,0.9640000092983245,0.8362222307920455,2019-11-18 13:26:29 92 | 2,2,20,20,5,15,200,100,False,False,True,200,0.001,0.0,0.9686666882038111,0.9406666702032089,2019-11-18 13:26:33 93 | 2,2,20,20,5,15,200,100,True,True,False,200,0.001,0.0,0.957666681110859,0.8482222312688831,2019-11-18 13:26:43 94 | 2,2,20,20,5,15,200,100,True,False,True,200,0.001,0.0,0.9656666931509972,0.8968888950347899,2019-11-18 13:27:32 95 | 2,2,20,20,5,15,200,100,False,True,True,200,0.001,0.0,0.9655000019073485,0.9543333798646926,2019-11-18 13:27:32 96 | 2,2,1,1,5,15,200,100,False,False,True,200,0.001,0.0,0.9116667060554028,0.8440000084042547,2019-11-18 13:06:22 97 | 2,2,20,20,5,15,200,100,True,True,True,200,0.001,0.0,0.9642500007152558,0.9293333798646927,2019-11-18 13:32:26 98 | 3,3,1,1,5,15,200,100,False,False,False,200,0.001,0.0,0.9726666846871375,0.8722222298383715,2019-11-18 13:32:45 99 | 3,3,1,1,5,15,200,100,True,False,False,200,0.001,0.0,0.9343333664536474,0.7884444567561151,2019-11-18 13:35:50 100 | 3,3,1,1,5,15,200,100,False,True,False,200,0.001,0.0,0.9030000036954878,0.6788333663344385,2019-11-18 13:35:50 101 | 3,3,1,1,5,15,200,100,False,False,True,200,0.001,0.0,0.8980000084638597,0.8005000436306001,2019-11-18 13:35:50 102 | 3,3,1,1,5,15,200,100,True,True,False,200,0.001,0.0,0.806500012166798,0.6045000344514849,2019-11-18 13:36:17 103 | 3,3,1,1,5,15,200,100,True,False,True,200,0.001,0.0,0.9025000071525572,0.727000040411949,2019-11-18 13:36:23 104 | 3,3,1,1,5,15,200,100,False,True,True,200,0.001,0.0,0.9233999785780909,0.8424000179767608,2019-11-18 13:36:28 105 | 3,3,1,1,5,15,200,100,True,True,True,200,0.001,0.0,0.8785999776422976,0.7844000184535982,2019-11-18 13:44:04 106 | 3,3,5,5,5,15,200,100,False,False,False,200,0.001,0.0,0.9806666806340218,0.9426666700839997,2019-11-18 13:45:26 107 | 2,2,1,1,5,15,200,100,True,True,False,200,0.001,0.0,0.8530000390112398,0.7184444588422775,2019-11-18 13:06:22 108 | 3,3,5,5,5,15,200,100,True,False,False,200,0.001,0.0,0.9723333519697193,0.9235555601119992,2019-11-18 13:45:26 109 | 3,3,5,5,5,15,200,100,False,True,False,200,0.001,0.0,0.9380000025033958,0.7865000450611114,2019-11-18 13:45:32 110 | 3,3,5,5,5,15,200,100,False,False,True,200,0.001,0.0,0.953750002682209,0.8986667144298551,2019-11-18 13:45:38 111 | 3,3,5,5,5,15,200,100,True,True,False,200,0.001,0.0,0.945749999880791,0.7861667084693909,2019-11-18 13:45:38 112 | 3,3,5,5,5,15,200,100,True,False,True,200,0.001,0.0,0.9522500017285348,0.85316671192646,2019-11-18 13:46:05 113 | 3,3,5,5,5,15,200,100,False,True,True,200,0.001,0.0,0.9675999853014949,0.9301333582401278,2019-11-18 13:46:26 114 | 3,3,5,5,5,15,200,100,True,True,True,200,0.001,0.0,0.9465999817848207,0.8909333544969558,2019-11-18 13:46:29 115 | 3,3,10,10,5,15,200,100,False,False,False,200,0.001,0.0,0.9796666818857194,0.9475555586814881,2019-11-18 13:47:16 116 | 3,3,10,10,5,15,200,100,True,False,False,200,0.001,0.0,0.9773333504796032,0.9342222261428832,2019-11-18 13:47:23 117 | 3,3,10,10,5,15,200,100,False,True,False,200,0.001,0.0,0.9507500007748608,0.8105000406503678,2019-11-18 13:47:57 118 | 2,2,1,1,5,15,200,100,True,False,True,200,0.001,0.0,0.8870000427961351,0.7835555678606034,2019-11-18 13:06:22 119 | 3,3,10,10,5,15,200,100,False,False,True,200,0.001,0.0,0.9612500017881391,0.9208333754539489,2019-11-18 13:49:34 120 | 3,3,10,10,5,15,200,100,True,True,False,200,0.001,0.0,0.9382500007748604,0.8018333745002744,2019-11-18 13:54:32 121 | 3,3,10,10,5,15,200,100,True,False,True,200,0.001,0.0,0.9492500039935116,0.8933333766460417,2019-11-18 13:54:45 122 | 3,3,10,10,5,15,200,100,False,True,True,200,0.001,0.0,0.9665999859571457,0.935866692662239,2019-11-18 13:54:49 123 | 3,3,10,10,5,15,200,100,True,True,True,200,0.001,0.0,0.9621999862790105,0.8901333570480344,2019-11-18 13:55:23 124 | 3,3,15,15,5,15,200,100,False,False,False,200,0.001,0.0,0.9843333455920221,0.9626666688919068,2019-11-18 13:55:26 125 | 3,3,15,15,5,15,200,100,True,False,False,200,0.001,0.0,0.9813333475589756,0.935555559396744,2019-11-18 13:55:35 126 | 3,3,15,15,5,15,200,100,False,True,False,200,0.001,0.0,0.9545000016689306,0.8276667094230653,2019-11-18 13:55:45 127 | 3,3,15,15,5,15,200,100,False,False,True,200,0.001,0.0,0.9697499996423721,0.923166711330414,2019-11-18 13:56:07 128 | 3,3,15,15,5,15,200,100,True,True,False,200,0.001,0.0,0.9525000008940699,0.7793333762884141,2019-11-18 13:56:14 129 | 2,2,1,1,5,15,200,100,False,True,True,200,0.001,0.0,0.9530000010132789,0.8890000444650649,2019-11-18 13:06:22 130 | 3,3,15,15,5,15,200,100,True,False,True,200,0.001,0.0,0.9475000044703485,0.8818333798646928,2019-11-18 13:56:28 131 | 3,3,15,15,5,15,200,100,False,True,True,200,0.001,0.0,0.9787999907135966,0.9365333610773083,2019-11-18 14:00:46 132 | 3,3,15,15,5,15,200,100,True,True,True,200,0.001,0.0,0.9501999813318249,0.9145333576202392,2019-11-18 14:05:25 133 | 3,3,20,20,5,15,200,100,False,False,False,200,0.001,0.0,0.9846666789054872,0.96244444668293,2019-11-18 14:05:36 134 | 3,3,20,20,5,15,200,100,True,False,False,200,0.001,0.0,0.962333358824253,0.9404444479942324,2019-11-18 14:07:09 135 | 3,3,20,20,5,15,200,100,False,True,False,200,0.001,0.0,0.9559999996423725,0.8086667090654371,2019-11-18 14:07:19 136 | 3,3,20,20,5,15,200,100,False,False,True,200,0.001,0.0,0.9677500003576278,0.9263333773612975,2019-11-18 14:08:58 137 | 3,3,20,20,5,15,200,100,True,True,False,200,0.001,0.0,0.9462500005960462,0.786166708469391,2019-11-18 14:09:06 138 | 3,3,20,20,5,15,200,100,True,False,True,200,0.001,0.0,0.9420000034570696,0.8971667093038559,2019-11-18 14:09:18 139 | 3,3,20,20,5,15,200,100,False,True,True,200,0.001,0.0,0.9693999862670899,0.9360000288486484,2019-11-18 14:10:43 140 | 2,2,1,1,5,15,200,100,True,True,True,200,0.001,0.0,0.9510000032186506,0.8451667094230655,2019-11-18 13:06:25 141 | 3,3,20,20,5,15,200,100,True,True,True,200,0.001,0.0,0.9623999840021131,0.9136000204086304,2019-11-18 14:13:59 142 | 4,4,1,1,5,15,200,100,False,False,False,200,0.001,0.0,0.9670000004768372,0.8005000460147857,2019-11-18 14:14:20 143 | 4,4,1,1,5,15,200,100,True,False,False,200,0.001,0.0,0.94900000423193,0.7750000435113907,2019-11-18 14:14:23 144 | 4,4,1,1,5,15,200,100,False,True,False,200,0.001,0.0,0.9143999817967413,0.6533333486318587,2019-11-18 14:16:10 145 | 4,4,1,1,5,15,200,100,False,False,True,200,0.001,0.0,0.9151999807357784,0.7392000177502632,2019-11-18 14:17:13 146 | 4,4,1,1,5,15,200,100,True,True,False,200,0.001,0.0,0.8411999779939653,0.5948000141978262,2019-11-18 14:17:16 147 | 4,4,1,1,5,15,200,100,True,False,True,200,0.001,0.0,0.9289999800920489,0.725866682231426,2019-11-18 14:23:28 148 | 4,4,1,1,5,15,200,100,False,True,True,200,0.001,0.0,0.9571666935086248,0.7875555795431134,2019-11-18 14:24:45 149 | 4,4,1,1,5,15,200,100,True,True,True,200,0.001,0.0,0.931333369612694,0.7612222439050672,2019-11-18 14:25:40 150 | 4,4,5,5,5,15,200,100,False,False,False,200,0.001,0.0,0.9884999987483026,0.9168333792686462,2019-11-18 14:27:08 151 | 2,2,5,5,5,15,200,100,False,False,False,200,0.001,0.0,0.9850000050663954,0.9696666902303691,2019-11-18 13:06:25 152 | 4,4,5,5,5,15,200,100,True,False,False,200,0.001,0.0,0.9752500003576278,0.8895000445842741,2019-11-18 14:27:20 153 | 4,4,5,5,5,15,200,100,False,True,False,200,0.001,0.0,0.9381999865174292,0.7708000177145002,2019-11-18 14:27:29 154 | 4,4,5,5,5,15,200,100,False,False,True,200,0.001,0.0,0.9555999827384947,0.8824000221490859,2019-11-18 14:27:32 155 | 4,4,5,5,5,15,200,100,True,True,False,200,0.001,0.0,0.9103999802470205,0.7224000173807146,2019-11-18 14:28:08 156 | 4,4,5,5,5,15,200,100,True,False,True,200,0.001,0.0,0.9653999847173693,0.8712000209093097,2019-11-18 14:29:57 157 | 4,4,5,5,5,15,200,100,False,True,True,200,0.001,0.0,0.9811666840314861,0.9023333561420444,2019-11-18 14:30:53 158 | 4,4,5,5,5,15,200,100,True,True,True,200,0.001,0.0,0.9666666913032529,0.8804444652795793,2019-11-18 14:33:47 159 | 4,4,10,10,5,15,200,100,False,False,False,200,0.001,0.0,0.9774999997019765,0.9283333730697635,2019-11-18 14:34:44 160 | 4,4,10,10,5,15,200,100,True,False,False,200,0.001,0.0,0.9565000012516969,0.9098333793878557,2019-11-18 14:36:06 161 | 4,4,10,10,5,15,200,100,False,True,False,200,0.001,0.0,0.9359999880194664,0.75613334774971,2019-11-18 14:36:17 162 | -------------------------------------------------------------------------------- /results/old/results_cnn_trad_fpool3.csv: -------------------------------------------------------------------------------- 1 | train.way,test.way,train.shot,test.shot,train.query,test.query,train.episodes,test.episodes,background,silence,unknown,epochs,lr,wd,train.acc,val.acc,timestamp 2 | 2,2,1,1,5,15,200,100,False,False,False,200,0.001,0.0,0.8100000092387198,0.7316667112708092,2019-11-05 10:33:36 3 | 2,2,5,5,5,15,200,100,True,False,False,200,0.001,0.0,0.9415000098943708,0.8873333770036698,2019-11-05 10:40:03 4 | 4,4,10,10,5,15,200,100,False,False,True,200,0.001,0.0,0.9237999808788299,0.7954666841030119,2019-11-05 12:04:50 5 | 4,4,10,10,5,15,200,100,True,True,False,200,0.001,0.0,0.9065999805927276,0.6289333462715151,2019-11-05 12:06:25 6 | 4,4,10,10,5,15,200,100,True,False,True,200,0.001,0.0,0.8661999768018726,0.7514666831493375,2019-11-05 12:06:57 7 | 4,4,10,10,5,15,200,100,False,True,True,200,0.001,0.0,0.9540000310540201,0.8148889112472534,2019-11-05 12:09:23 8 | 4,4,10,10,5,15,200,100,True,True,True,200,0.001,0.0,0.9536666986346244,0.7948889160156251,2019-11-05 12:10:32 9 | 4,4,15,15,5,15,200,100,False,False,False,200,0.001,0.0,0.9262500074505804,0.8298333787918093,2019-11-05 12:12:06 10 | 4,4,15,15,5,15,200,100,True,False,False,200,0.001,0.0,0.8960000097751617,0.7711667066812516,2019-11-05 12:12:10 11 | 4,4,15,15,5,15,200,100,False,True,False,200,0.001,0.0,0.8561999750137328,0.663600016236305,2019-11-05 12:12:59 12 | 4,4,15,15,5,15,200,100,False,False,True,200,0.001,0.0,0.9593999844789504,0.8045333516597747,2019-11-05 12:13:31 13 | 4,4,15,15,5,15,200,100,True,True,False,200,0.001,0.0,0.9443999892473223,0.6578666776418688,2019-11-05 12:15:50 14 | 2,2,5,5,5,15,200,100,False,True,False,200,0.001,0.0,0.944666686058044,0.6877777925133705,2019-11-05 10:40:16 15 | 4,4,15,15,5,15,200,100,True,False,True,200,0.001,0.0,0.928199977874756,0.7553333503007889,2019-11-05 12:16:25 16 | 4,4,15,15,5,15,200,100,False,True,True,200,0.001,0.0,0.9350000402331351,0.8330000263452528,2019-11-05 12:19:40 17 | 4,4,15,15,5,15,200,100,True,True,True,200,0.001,0.0,0.9658333587646482,0.8045555818080905,2019-11-05 12:20:56 18 | 4,4,20,20,5,15,200,100,False,False,False,200,0.001,0.0,0.9745000007748605,0.8596667122840878,2019-11-05 12:21:09 19 | 4,4,20,20,5,15,200,100,True,False,False,200,0.001,0.0,0.9315000057220454,0.8191667133569717,2019-11-05 12:21:30 20 | 4,4,20,20,5,15,200,100,False,True,False,200,0.001,0.0,0.9355999839305879,0.6840000146627424,2019-11-05 12:22:44 21 | 4,4,20,20,5,15,200,100,False,False,True,200,0.001,0.0,0.9223999789357187,0.8148000198602678,2019-11-05 12:23:18 22 | 4,4,20,20,5,15,200,100,True,True,False,200,0.001,0.0,0.9265999826788901,0.6477333456277846,2019-11-05 12:26:01 23 | 4,4,20,20,5,15,200,100,True,False,True,200,0.001,0.0,0.834199975430965,0.7434666854143142,2019-11-05 12:27:06 24 | 4,4,20,20,5,15,200,100,False,True,True,200,0.001,0.0,0.9551666951179506,0.8364444684982301,2019-11-05 12:30:46 25 | 2,2,5,5,5,15,200,100,False,False,True,200,0.001,0.0,0.9470000317692757,0.7993333452939987,2019-11-05 10:40:21 26 | 4,4,20,20,5,15,200,100,True,True,True,200,0.001,0.0,0.9690000233054161,0.8135555815696714,2019-11-05 12:31:11 27 | 5,5,1,1,5,15,200,100,False,False,False,200,0.001,0.0,0.8707999762892724,0.5569333450496199,2019-11-05 12:31:18 28 | 5,5,1,1,5,15,200,100,True,False,False,200,0.001,0.0,0.715599980726838,0.433733344823122,2019-11-05 12:32:16 29 | 5,5,1,1,5,15,200,100,False,True,False,200,0.001,0.0,0.7343333736062053,0.39088889628648765,2019-11-05 12:33:56 30 | 5,5,1,1,5,15,200,100,False,False,True,200,0.001,0.0,0.7750000429153442,0.5175555649399759,2019-11-05 12:34:24 31 | 5,5,1,1,5,15,200,100,True,True,False,200,0.001,0.0,0.6011666996777054,0.3484444519877433,2019-11-05 12:37:28 32 | 5,5,1,1,5,15,200,100,True,False,True,200,0.001,0.0,0.7180000391602512,0.4273333404958248,2019-11-05 12:38:40 33 | 5,5,1,1,5,15,200,100,False,True,True,200,0.001,0.0,0.7714285850524899,0.5519047859311104,2019-11-05 12:38:56 34 | 5,5,1,1,5,15,200,100,True,True,True,200,0.001,0.0,0.6508571614325048,0.47847621142864216,2019-11-05 12:39:21 35 | 5,5,5,5,5,15,200,100,False,False,False,200,0.001,0.0,0.895199978351593,0.7374666804075242,2019-11-05 12:40:47 36 | 2,2,5,5,5,15,200,100,True,True,False,200,0.001,0.0,0.9273333549499511,0.7057777929306032,2019-11-05 10:40:27 37 | 5,5,5,5,5,15,200,100,True,False,False,200,0.001,0.0,0.8685999777913097,0.5974666795134542,2019-11-05 12:40:49 38 | 5,5,5,5,5,15,200,100,False,True,False,200,0.001,0.0,0.8705000445246694,0.5745555660128592,2019-11-05 12:43:52 39 | 5,5,5,5,5,15,200,100,False,False,True,200,0.001,0.0,0.8590000492334366,0.7050000166893007,2019-11-05 12:44:12 40 | 5,5,5,5,5,15,200,100,True,True,False,200,0.001,0.0,0.863166713118553,0.577222234904766,2019-11-05 12:44:17 41 | 5,5,5,5,5,15,200,100,True,False,True,200,0.001,0.0,0.8185000446438788,0.6471111166477206,2019-11-05 12:45:35 42 | 5,5,5,5,5,15,200,100,False,True,True,200,0.001,0.0,0.9138571479916576,0.7912381076812743,2019-11-05 12:46:28 43 | 5,5,5,5,5,15,200,100,True,True,True,200,0.001,0.0,0.8425714379549029,0.7024762082099915,2019-11-05 12:47:05 44 | 5,5,10,10,5,15,200,100,False,False,False,200,0.001,0.0,0.9641999852657318,0.7802666831016544,2019-11-05 12:47:16 45 | 5,5,10,10,5,15,200,100,True,False,False,200,0.001,0.0,0.9343999812006955,0.7456000155210494,2019-11-05 12:47:19 46 | 5,5,10,10,5,15,200,100,False,True,False,200,0.001,0.0,0.8851667100191116,0.6430000084638594,2019-11-05 12:52:23 47 | 2,2,5,5,5,15,200,100,True,False,True,200,0.001,0.0,0.9173333728313445,0.7651111248135568,2019-11-05 10:40:36 48 | 5,5,10,10,5,15,200,100,False,False,True,200,0.001,0.0,0.8561667162179948,0.7637778037786482,2019-11-05 12:52:38 49 | 5,5,10,10,5,15,200,100,True,True,False,200,0.001,0.0,0.8675000461936,0.571111120581627,2019-11-05 12:53:54 50 | 5,5,10,10,5,15,200,100,True,False,True,200,0.001,0.0,0.9055000436306,0.6962222373485563,2019-11-05 12:54:10 51 | 5,5,10,10,5,15,200,100,False,True,True,200,0.001,0.0,0.9315714326500891,0.8213333439826963,2019-11-05 12:55:36 52 | 5,5,10,10,5,15,200,100,True,True,True,200,0.001,0.0,0.9081428626179697,0.751523824334145,2019-11-05 12:55:54 53 | 5,5,15,15,5,15,200,100,False,False,False,200,0.001,0.0,0.9325999787449836,0.7788000160455701,2019-11-05 12:56:02 54 | 5,5,15,15,5,15,200,100,True,False,False,200,0.001,0.0,0.9079999765753752,0.7572000181674956,2019-11-05 12:56:46 55 | 5,5,15,15,5,15,200,100,False,True,False,200,0.001,0.0,0.8958333757519724,0.6278888994455337,2019-11-05 13:02:13 56 | 5,5,15,15,5,15,200,100,False,False,True,200,0.001,0.0,0.936166707575321,0.7736666905879973,2019-11-05 13:02:19 57 | 5,5,15,15,5,15,200,100,True,True,False,200,0.001,0.0,0.849166713356972,0.5985555669665334,2019-11-05 13:03:58 58 | 2,2,5,5,5,15,200,100,False,True,True,200,0.001,0.0,0.8930000111460686,0.8096667069196704,2019-11-05 10:40:46 59 | 5,5,15,15,5,15,200,100,True,False,True,200,0.001,0.0,0.848333380818367,0.7228889065980914,2019-11-05 13:04:55 60 | 5,5,15,15,5,15,200,100,False,True,True,200,0.001,0.0,0.9464285746216776,0.8259047728776931,2019-11-05 13:05:47 61 | 5,5,15,15,5,15,200,100,True,True,True,200,0.001,0.0,0.8808571499586104,0.7546666812896727,2019-11-05 13:06:31 62 | 5,5,20,20,5,15,200,100,False,False,False,200,0.001,0.0,0.9355999794602395,0.8116000211238861,2019-11-05 13:06:37 63 | 5,5,20,20,5,15,200,100,True,False,False,200,0.001,0.0,0.9395999786257745,0.7577333486080169,2019-11-05 13:07:18 64 | 5,5,20,20,5,15,200,100,False,True,False,200,0.001,0.0,0.8998333731293681,0.6632222306728365,2019-11-05 13:13:30 65 | 5,5,20,20,5,15,200,100,False,False,True,200,0.001,0.0,0.9308333733677864,0.7620000231266022,2019-11-05 13:13:35 66 | 5,5,20,20,5,15,200,100,True,True,False,200,0.001,0.0,0.9078333699703215,0.6532222342491149,2019-11-05 13:15:22 67 | 5,5,20,20,5,15,200,100,True,False,True,200,0.001,0.0,0.9188333782553669,0.7173333525657655,2019-11-05 13:16:15 68 | 5,5,20,20,5,15,200,100,False,True,True,200,0.001,0.0,0.9160000050067901,0.8343809628486633,2019-11-05 13:17:40 69 | 2,2,5,5,5,15,200,100,True,True,True,200,0.001,0.0,0.9225000080466271,0.8050000452995304,2019-11-05 10:41:35 70 | 5,5,20,20,5,15,200,100,True,True,True,200,0.001,0.0,0.9475714316964151,0.7852381080389026,2019-11-05 13:18:22 71 | 2,2,10,10,5,15,200,100,False,False,False,200,0.001,0.0,0.9540000072121623,0.9203333669900893,2019-11-05 10:45:23 72 | 2,2,10,10,5,15,200,100,True,False,False,200,0.001,0.0,0.943500008285046,0.9016667097806931,2019-11-05 10:45:28 73 | 2,2,10,10,5,15,200,100,False,True,False,200,0.001,0.0,0.9253333497047425,0.7488889017701148,2019-11-05 10:46:44 74 | 2,2,1,1,5,15,200,100,True,False,False,200,0.001,0.0,0.8180000106617812,0.6533333707600832,2019-11-05 10:33:36 75 | 2,2,10,10,5,15,200,100,False,False,True,200,0.001,0.0,0.9496666994690894,0.8404444539546964,2019-11-05 10:46:54 76 | 2,2,10,10,5,15,200,100,True,True,False,200,0.001,0.0,0.9506666836142538,0.6695555734634399,2019-11-05 10:46:56 77 | 2,2,10,10,5,15,200,100,True,False,True,200,0.001,0.0,0.9530000257492071,0.7862222343683241,2019-11-05 10:47:19 78 | 2,2,10,10,5,15,200,100,False,True,True,200,0.001,0.0,0.8997500112652776,0.8211667108535768,2019-11-05 10:48:12 79 | 2,2,10,10,5,15,200,100,True,True,True,200,0.001,0.0,0.931250006556511,0.810833379626274,2019-11-05 10:49:46 80 | 2,2,15,15,5,15,200,100,False,False,False,200,0.001,0.0,0.9720000064373023,0.9286666989326476,2019-11-05 10:51:06 81 | 2,2,15,15,5,15,200,100,True,False,False,200,0.001,0.0,0.9535000090301033,0.9200000393390654,2019-11-05 10:51:06 82 | 2,2,15,15,5,15,200,100,False,True,False,200,0.001,0.0,0.9643333464860914,0.739777790904045,2019-11-05 10:53:30 83 | 2,2,15,15,5,15,200,100,False,False,True,200,0.001,0.0,0.9296667090058321,0.8168888998031616,2019-11-05 10:53:36 84 | 2,2,15,15,5,15,200,100,True,True,False,200,0.001,0.0,0.9410000237822533,0.7104444599151607,2019-11-05 10:53:41 85 | 2,2,1,1,5,15,200,100,False,True,False,200,0.001,0.0,0.7153333699703213,0.6031111273169515,2019-11-05 10:33:36 86 | 2,2,15,15,5,15,200,100,True,False,True,200,0.001,0.0,0.8643333822488786,0.7495555698871613,2019-11-05 10:54:36 87 | 2,2,15,15,5,15,200,100,False,True,True,200,0.001,0.0,0.9710000002384183,0.881500044465065,2019-11-05 10:56:33 88 | 2,2,15,15,5,15,200,100,True,True,True,200,0.001,0.0,0.9477500027418131,0.8465000456571579,2019-11-05 10:57:35 89 | 2,2,20,20,5,15,200,100,False,False,False,200,0.001,0.0,0.9550000080466279,0.9486666953563689,2019-11-05 10:57:38 90 | 2,2,20,20,5,15,200,100,True,False,False,200,0.001,0.0,0.9430000080168247,0.8906667160987854,2019-11-05 10:59:13 91 | 2,2,20,20,5,15,200,100,False,True,False,200,0.001,0.0,0.9240000179409981,0.7175555701553823,2019-11-05 11:01:58 92 | 2,2,20,20,5,15,200,100,False,False,True,200,0.001,0.0,0.9266667047142985,0.8475555646419523,2019-11-05 11:02:12 93 | 2,2,20,20,5,15,200,100,True,True,False,200,0.001,0.0,0.9530000191926955,0.7180000150203705,2019-11-05 11:02:15 94 | 2,2,20,20,5,15,200,100,True,False,True,200,0.001,0.0,0.9356667062640194,0.8155555665493011,2019-11-05 11:04:11 95 | 2,2,20,20,5,15,200,100,False,True,True,200,0.001,0.0,0.9417500057816499,0.8835000431537626,2019-11-05 11:05:01 96 | 2,2,1,1,5,15,200,100,False,False,True,200,0.001,0.0,0.7203333720564844,0.6324444612860678,2019-11-05 10:33:36 97 | 2,2,20,20,5,15,200,100,True,True,True,200,0.001,0.0,0.9300000083446505,0.8363333797454834,2019-11-05 11:05:43 98 | 3,3,1,1,5,15,200,100,False,False,False,200,0.001,0.0,0.8723333710432053,0.622444459795952,2019-11-05 11:06:19 99 | 3,3,1,1,5,15,200,100,True,False,False,200,0.001,0.0,0.810000044703483,0.6420000171661375,2019-11-05 11:07:06 100 | 3,3,1,1,5,15,200,100,False,True,False,200,0.001,0.0,0.8102500134706503,0.5088333603739739,2019-11-05 11:10:18 101 | 3,3,1,1,5,15,200,100,False,False,True,200,0.001,0.0,0.6912500119209292,0.5580000293254854,2019-11-05 11:10:31 102 | 3,3,1,1,5,15,200,100,True,True,False,200,0.001,0.0,0.5135000101290644,0.37733335524797434,2019-11-05 11:10:34 103 | 3,3,1,1,5,15,200,100,True,False,True,200,0.001,0.0,0.6997500118613241,0.5058333611488344,2019-11-05 11:11:59 104 | 3,3,1,1,5,15,200,100,False,True,True,200,0.001,0.0,0.8019999787211417,0.670933346748352,2019-11-05 11:12:40 105 | 3,3,1,1,5,15,200,100,True,True,True,200,0.001,0.0,0.7901999780535697,0.6245333465933799,2019-11-05 11:13:55 106 | 3,3,5,5,5,15,200,100,False,False,False,200,0.001,0.0,0.9623333570361138,0.8577777862548825,2019-11-05 11:14:45 107 | 2,2,1,1,5,15,200,100,True,True,False,200,0.001,0.0,0.6163333620131017,0.49600001588463766,2019-11-05 10:33:36 108 | 3,3,5,5,5,15,200,100,True,False,False,200,0.001,0.0,0.9536666923761367,0.836666676402092,2019-11-05 11:15:31 109 | 3,3,5,5,5,15,200,100,False,True,False,200,0.001,0.0,0.9302500021457671,0.6791667050123218,2019-11-05 11:16:23 110 | 3,3,5,5,5,15,200,100,False,False,True,200,0.001,0.0,0.8842500105500221,0.7531667059659956,2019-11-05 11:16:33 111 | 3,3,5,5,5,15,200,100,True,True,False,200,0.001,0.0,0.9180000036954882,0.6110000351071357,2019-11-05 11:16:37 112 | 3,3,5,5,5,15,200,100,True,False,True,200,0.001,0.0,0.8817500117421152,0.6998333752155304,2019-11-05 11:18:10 113 | 3,3,5,5,5,15,200,100,False,True,True,200,0.001,0.0,0.9111999753117564,0.8128000181913378,2019-11-05 11:19:51 114 | 3,3,5,5,5,15,200,100,True,True,True,200,0.001,0.0,0.9263999813795091,0.8060000193119052,2019-11-05 11:21:47 115 | 3,3,10,10,5,15,200,100,False,False,False,200,0.001,0.0,0.9180000379681581,0.8391111207008362,2019-11-05 11:21:57 116 | 3,3,10,10,5,15,200,100,True,False,False,200,0.001,0.0,0.969000020623207,0.8108888998627661,2019-11-05 11:22:28 117 | 3,3,10,10,5,15,200,100,False,True,False,200,0.001,0.0,0.9385000011324879,0.6906667086482045,2019-11-05 11:23:07 118 | 2,2,1,1,5,15,200,100,True,False,True,200,0.001,0.0,0.7633333727717397,0.6075555738806725,2019-11-05 10:33:36 119 | 3,3,10,10,5,15,200,100,False,False,True,200,0.001,0.0,0.9147500085830692,0.7765000426769254,2019-11-05 11:23:38 120 | 3,3,10,10,5,15,200,100,True,True,False,200,0.001,0.0,0.898750007450581,0.6523333698511126,2019-11-05 11:23:41 121 | 3,3,10,10,5,15,200,100,True,False,True,200,0.001,0.0,0.887750012874603,0.7751667064428329,2019-11-05 11:25:22 122 | 3,3,10,10,5,15,200,100,False,True,True,200,0.001,0.0,0.9597999846935269,0.8433333545923234,2019-11-05 11:26:41 123 | 3,3,10,10,5,15,200,100,True,True,True,200,0.001,0.0,0.9311999797821048,0.8080000215768813,2019-11-05 11:28:28 124 | 3,3,15,15,5,15,200,100,False,False,False,200,0.001,0.0,0.976333351135254,0.873333340883255,2019-11-05 11:28:49 125 | 3,3,15,15,5,15,200,100,True,False,False,200,0.001,0.0,0.9616666933894158,0.8675555634498596,2019-11-05 11:30:15 126 | 3,3,15,15,5,15,200,100,False,True,False,200,0.001,0.0,0.9017500063776973,0.6848333740234378,2019-11-05 11:30:50 127 | 3,3,15,15,5,15,200,100,False,False,True,200,0.001,0.0,0.9550000008940693,0.8221667104959485,2019-11-05 11:31:32 128 | 3,3,15,15,5,15,200,100,True,True,False,200,0.001,0.0,0.9290000048279756,0.653000037074089,2019-11-05 11:31:38 129 | 2,2,1,1,5,15,200,100,False,True,True,200,0.001,0.0,0.7357500104606155,0.7193333765864369,2019-11-05 10:33:36 130 | 3,3,15,15,5,15,200,100,True,False,True,200,0.001,0.0,0.9120000094175339,0.7630000424385072,2019-11-05 11:33:41 131 | 3,3,15,15,5,15,200,100,False,True,True,200,0.001,0.0,0.9543999826908107,0.8522666889429097,2019-11-05 11:35:38 132 | 3,3,15,15,5,15,200,100,True,True,True,200,0.001,0.0,0.9185999786853793,0.81213335454464,2019-11-05 11:37:47 133 | 3,3,20,20,5,15,200,100,False,False,False,200,0.001,0.0,0.9703333550691603,0.8615555638074874,2019-11-05 11:38:50 134 | 3,3,20,20,5,15,200,100,True,False,False,200,0.001,0.0,0.9343333706259727,0.8271111214160919,2019-11-05 11:39:21 135 | 3,3,20,20,5,15,200,100,False,True,False,200,0.001,0.0,0.8880000051856044,0.6963333705067634,2019-11-05 11:40:03 136 | 3,3,20,20,5,15,200,100,False,False,True,200,0.001,0.0,0.9200000083446498,0.8213333767652512,2019-11-05 11:40:16 137 | 3,3,20,20,5,15,200,100,True,True,False,200,0.001,0.0,0.9477500011026863,0.7138333737850188,2019-11-05 11:40:20 138 | 3,3,20,20,5,15,200,100,True,False,True,200,0.001,0.0,0.9132500088214874,0.7830000436305998,2019-11-05 11:42:39 139 | 3,3,20,20,5,15,200,100,False,True,True,200,0.001,0.0,0.9475999817252158,0.8661333549022675,2019-11-05 11:45:56 140 | 2,2,1,1,5,15,200,100,True,True,True,200,0.001,0.0,0.8605000081658367,0.6991667053103446,2019-11-05 10:33:36 141 | 3,3,20,20,5,15,200,100,True,True,True,200,0.001,0.0,0.9221999794244762,0.8284000223875047,2019-11-05 11:47:24 142 | 4,4,1,1,5,15,200,100,False,False,False,200,0.001,0.0,0.7272500123828652,0.5233333599567414,2019-11-05 11:48:10 143 | 4,4,1,1,5,15,200,100,True,False,False,200,0.001,0.0,0.622250010445714,0.40900002270936964,2019-11-05 11:48:19 144 | 4,4,1,1,5,15,200,100,False,True,False,200,0.001,0.0,0.673999980539084,0.4393333446979523,2019-11-05 11:49:53 145 | 4,4,1,1,5,15,200,100,False,False,True,200,0.001,0.0,0.7307999813556674,0.5754666778445245,2019-11-05 11:49:56 146 | 4,4,1,1,5,15,200,100,True,True,False,200,0.001,0.0,0.6413999849185351,0.41093334332108505,2019-11-05 11:49:59 147 | 4,4,1,1,5,15,200,100,True,False,True,200,0.001,0.0,0.6821999828517434,0.4846666783094407,2019-11-05 11:52:38 148 | 4,4,1,1,5,15,200,100,False,True,True,200,0.001,0.0,0.7965000453591343,0.5977777880430221,2019-11-05 11:54:08 149 | 4,4,1,1,5,15,200,100,True,True,True,200,0.001,0.0,0.7331667090952396,0.5792222324013713,2019-11-05 11:54:13 150 | 4,4,5,5,5,15,200,100,False,False,False,200,0.001,0.0,0.9172500088810926,0.7441667079925538,2019-11-05 11:57:09 151 | 2,2,5,5,5,15,200,100,False,False,False,200,0.001,0.0,0.947500008940697,0.8906667113304135,2019-11-05 10:39:57 152 | 4,4,5,5,5,15,200,100,True,False,False,200,0.001,0.0,0.8975000107288362,0.7260000425577162,2019-11-05 11:57:12 153 | 4,4,5,5,5,15,200,100,False,True,False,200,0.001,0.0,0.8647999769449233,0.6361333480477328,2019-11-05 11:57:29 154 | 4,4,5,5,5,15,200,100,False,False,True,200,0.001,0.0,0.8963999778032307,0.7360000151395798,2019-11-05 11:58:34 155 | 4,4,5,5,5,15,200,100,True,True,False,200,0.001,0.0,0.9235999846458434,0.6098666819930076,2019-11-05 11:58:41 156 | 4,4,5,5,5,15,200,100,True,False,True,200,0.001,0.0,0.8719999769330022,0.6918666827678679,2019-11-05 12:00:10 157 | 4,4,5,5,5,15,200,100,False,True,True,200,0.001,0.0,0.9413333678245548,0.7825555807352066,2019-11-05 12:00:46 158 | 4,4,5,5,5,15,200,100,True,True,True,200,0.001,0.0,0.935000040233135,0.7571111363172531,2019-11-05 12:00:48 159 | 4,4,10,10,5,15,200,100,False,False,False,200,0.001,0.0,0.8570000135898596,0.7923333770036702,2019-11-05 12:03:56 160 | 4,4,10,10,5,15,200,100,True,False,False,200,0.001,0.0,0.9115000104904175,0.786000043749809,2019-11-05 12:04:11 161 | 4,4,10,10,5,15,200,100,False,True,False,200,0.001,0.0,0.899599979221821,0.6684000152349472,2019-11-05 12:04:18 162 | -------------------------------------------------------------------------------- /results/old/results_cnn_trad_fpool3_simple.csv: -------------------------------------------------------------------------------- 1 | train.way,test.way,train.shot,test.shot,train.query,test.query,train.episodes,test.episodes,background,silence,unknown,epochs,lr,wd,train.acc,val.acc,timestamp 2 | 2,2,1,1,5,15,200,100,False,False,False,200,0.001,0.0,0.7825000106915833,0.6883333736658094,2019-11-04 13:58:37 3 | 2,2,5,5,5,15,200,100,True,False,False,200,0.001,0.0,0.8955000112950806,0.7990000480413441,2019-11-04 13:58:40 4 | 4,4,10,10,5,15,200,100,False,False,True,200,0.001,0.0,0.9133999776840214,0.7536000168323516,2019-11-04 14:51:13 5 | 4,4,10,10,5,15,200,100,True,True,False,200,0.001,0.0,0.8529999747872354,0.6041333481669423,2019-11-04 14:52:57 6 | 4,4,10,10,5,15,200,100,True,False,True,200,0.001,0.0,0.8603999745845798,0.6144000115990637,2019-11-04 14:53:47 7 | 4,4,10,10,5,15,200,100,False,True,True,200,0.001,0.0,0.9176667088270192,0.8081111377477647,2019-11-04 14:54:08 8 | 4,4,10,10,5,15,200,100,True,True,True,200,0.001,0.0,0.8788333806395535,0.6884444582462307,2019-11-04 14:54:18 9 | 4,4,15,15,5,15,200,100,False,False,False,200,0.001,0.0,0.9332500073313713,0.84633338034153,2019-11-04 14:54:26 10 | 4,4,15,15,5,15,200,100,True,False,False,200,0.001,0.0,0.8942500117421156,0.7385000401735308,2019-11-04 14:54:30 11 | 4,4,15,15,5,15,200,100,False,True,False,200,0.001,0.0,0.8805999767780305,0.6766666781902313,2019-11-04 14:55:37 12 | 4,4,15,15,5,15,200,100,False,False,True,200,0.001,0.0,0.9185999765992165,0.7893333536386491,2019-11-04 14:58:03 13 | 4,4,15,15,5,15,200,100,True,True,False,200,0.001,0.0,0.8679999780654908,0.5910666778683666,2019-11-04 14:58:57 14 | 2,2,5,5,5,15,200,100,False,True,False,200,0.001,0.0,0.897333369255066,0.7042222383618358,2019-11-04 13:58:40 15 | 4,4,15,15,5,15,200,100,True,False,True,200,0.001,0.0,0.8405999764800073,0.6602666783332823,2019-11-04 14:59:18 16 | 4,4,15,15,5,15,200,100,False,True,True,200,0.001,0.0,0.9201667127013212,0.8187778025865554,2019-11-04 14:59:39 17 | 4,4,15,15,5,15,200,100,True,True,True,200,0.001,0.0,0.9036667147278785,0.7212222445011139,2019-11-04 14:59:43 18 | 4,4,20,20,5,15,200,100,False,False,False,200,0.001,0.0,0.9397500047087669,0.8405000436305998,2019-11-04 15:01:10 19 | 4,4,20,20,5,15,200,100,True,False,False,200,0.001,0.0,0.9112500107288363,0.7545000439882278,2019-11-04 15:03:36 20 | 4,4,20,20,5,15,200,100,False,True,False,200,0.001,0.0,0.9013999778032301,0.6969333463907241,2019-11-04 15:03:38 21 | 4,4,20,20,5,15,200,100,False,False,True,200,0.001,0.0,0.9229999756813051,0.8068000191450118,2019-11-04 15:03:43 22 | 4,4,20,20,5,15,200,100,True,True,False,200,0.001,0.0,0.8873999780416488,0.6341333484649658,2019-11-04 15:04:13 23 | 4,4,20,20,5,15,200,100,True,False,True,200,0.001,0.0,0.8893999776244165,0.7021333467960358,2019-11-04 15:04:16 24 | 4,4,20,20,5,15,200,100,False,True,True,200,0.001,0.0,0.9348333740234372,0.8368889129161838,2019-11-04 15:04:46 25 | 2,2,5,5,5,15,200,100,False,False,True,200,0.001,0.0,0.8823333847522739,0.8293333435058595,2019-11-04 13:58:40 26 | 4,4,20,20,5,15,200,100,True,True,True,200,0.001,0.0,0.8976667121052742,0.7390000212192535,2019-11-04 15:06:15 27 | 5,5,1,1,5,15,200,100,False,False,False,200,0.001,0.0,0.6883999824523929,0.45026667669415477,2019-11-04 15:08:05 28 | 5,5,1,1,5,15,200,100,True,False,False,200,0.001,0.0,0.5645999862253662,0.34320000767707814,2019-11-04 15:08:45 29 | 5,5,1,1,5,15,200,100,False,True,False,200,0.001,0.0,0.591000033915043,0.38155556246638317,2019-11-04 15:09:55 30 | 5,5,1,1,5,15,200,100,False,False,True,200,0.001,0.0,0.5748333664238457,0.3742222295701505,2019-11-04 15:10:49 31 | 5,5,1,1,5,15,200,100,True,True,False,200,0.001,0.0,0.5333333633840083,0.32322223111987114,2019-11-04 15:10:51 32 | 5,5,1,1,5,15,200,100,True,False,True,200,0.001,0.0,0.47000002697110177,0.309555565044284,2019-11-04 15:12:01 33 | 5,5,1,1,5,15,200,100,False,True,True,200,0.001,0.0,0.6532857331633569,0.5191428819298746,2019-11-04 15:13:43 34 | 5,5,1,1,5,15,200,100,True,True,True,200,0.001,0.0,0.5204285871982575,0.3930476304888725,2019-11-04 15:14:45 35 | 5,5,5,5,5,15,200,100,False,False,False,200,0.001,0.0,0.8927999749779701,0.6974666798114777,2019-11-04 15:15:11 36 | 2,2,5,5,5,15,200,100,True,True,False,200,0.001,0.0,0.8793333795666695,0.6537777957320213,2019-11-04 13:58:40 37 | 5,5,5,5,5,15,200,100,True,False,False,200,0.001,0.0,0.8281999766826629,0.5776000162959101,2019-11-04 15:15:17 38 | 5,5,5,5,5,15,200,100,False,True,False,200,0.001,0.0,0.8238333782553672,0.5228888967633247,2019-11-04 15:15:26 39 | 5,5,5,5,5,15,200,100,False,False,True,200,0.001,0.0,0.8670000484585763,0.6858889001607895,2019-11-04 15:15:40 40 | 5,5,5,5,5,15,200,100,True,True,False,200,0.001,0.0,0.7875000447034837,0.48044445157051097,2019-11-04 15:16:08 41 | 5,5,5,5,5,15,200,100,True,False,True,200,0.001,0.0,0.8020000472664832,0.5231111222505567,2019-11-04 15:16:13 42 | 5,5,5,5,5,15,200,100,False,True,True,200,0.001,0.0,0.8747142931818962,0.7325714445114136,2019-11-04 15:16:54 43 | 5,5,5,5,5,15,200,100,True,True,True,200,0.001,0.0,0.8192857249081135,0.5850476437807083,2019-11-04 15:17:05 44 | 5,5,10,10,5,15,200,100,False,False,False,200,0.001,0.0,0.9151999777555468,0.7402666836977002,2019-11-04 15:18:03 45 | 5,5,10,10,5,15,200,100,True,False,False,200,0.001,0.0,0.8795999780297278,0.6717333477735516,2019-11-04 15:18:39 46 | 5,5,10,10,5,15,200,100,False,True,False,200,0.001,0.0,0.8731667116284371,0.6198888975381851,2019-11-04 15:20:41 47 | 2,2,5,5,5,15,200,100,True,False,True,200,0.001,0.0,0.8463333815336224,0.6871111291646955,2019-11-04 13:58:40 48 | 5,5,10,10,5,15,200,100,False,False,True,200,0.001,0.0,0.8913333809375761,0.7445555788278582,2019-11-04 15:21:25 49 | 5,5,10,10,5,15,200,100,True,True,False,200,0.001,0.0,0.841166713237763,0.5541111227869989,2019-11-04 15:21:28 50 | 5,5,10,10,5,15,200,100,True,False,True,200,0.001,0.0,0.8466667163372045,0.5961111217737198,2019-11-04 15:21:33 51 | 5,5,10,10,5,15,200,100,False,True,True,200,0.001,0.0,0.893142863512039,0.7845714414119721,2019-11-04 15:22:03 52 | 5,5,10,10,5,15,200,100,True,True,True,200,0.001,0.0,0.8654285794496536,0.6423809736967089,2019-11-04 15:23:22 53 | 5,5,15,15,5,15,200,100,False,False,False,200,0.001,0.0,0.937399979829788,0.8105333524942397,2019-11-04 15:24:31 54 | 5,5,15,15,5,15,200,100,True,False,False,200,0.001,0.0,0.8949999758601189,0.6852000153064723,2019-11-04 15:24:49 55 | 5,5,15,15,5,15,200,100,False,True,False,200,0.001,0.0,0.8720000448822972,0.6530000078678133,2019-11-04 15:24:57 56 | 5,5,15,15,5,15,200,100,False,False,True,200,0.001,0.0,0.910833377838135,0.7494444674253465,2019-11-04 15:26:18 57 | 5,5,15,15,5,15,200,100,True,True,False,200,0.001,0.0,0.8535000455379486,0.589888898134232,2019-11-04 15:27:31 58 | 2,2,5,5,5,15,200,100,False,True,True,200,0.001,0.0,0.9102500092983249,0.860333377122879,2019-11-04 14:04:36 59 | 5,5,15,15,5,15,200,100,True,False,True,200,0.001,0.0,0.8483333823084832,0.6425555616617202,2019-11-04 15:27:34 60 | 5,5,15,15,5,15,200,100,False,True,True,200,0.001,0.0,0.907142862677574,0.7703809660673143,2019-11-04 15:27:54 61 | 5,5,15,15,5,15,200,100,True,True,True,200,0.001,0.0,0.8848571497201919,0.6753333526849745,2019-11-04 15:31:17 62 | 5,5,20,20,5,15,200,100,False,False,False,200,0.001,0.0,0.9351999804377549,0.8257333528995512,2019-11-04 15:31:26 63 | 5,5,20,20,5,15,200,100,True,False,False,200,0.001,0.0,0.8991999751329419,0.7097333484888075,2019-11-04 15:32:09 64 | 5,5,20,20,5,15,200,100,False,True,False,200,0.001,0.0,0.898166707754135,0.6584444528818127,2019-11-04 15:32:17 65 | 5,5,20,20,5,15,200,100,False,False,True,200,0.001,0.0,0.9245000410079955,0.7968889176845552,2019-11-04 15:33:09 66 | 5,5,20,20,5,15,200,100,True,True,False,200,0.001,0.0,0.8720000457763674,0.5947777882218359,2019-11-04 15:34:25 67 | 5,5,20,20,5,15,200,100,True,False,True,200,0.001,0.0,0.8851667159795762,0.6760000091791151,2019-11-04 15:35:05 68 | 5,5,20,20,5,15,200,100,False,True,True,200,0.001,0.0,0.9271428614854815,0.8015238213539124,2019-11-04 15:35:08 69 | 2,2,5,5,5,15,200,100,True,True,True,200,0.001,0.0,0.902500008940697,0.7601667100191114,2019-11-04 14:04:44 70 | 5,5,20,20,5,15,200,100,True,True,True,200,0.001,0.0,0.8954285776615143,0.7067619222402575,2019-11-04 15:37:04 71 | 2,2,10,10,5,15,200,100,False,False,False,200,0.001,0.0,0.9340000119805338,0.894333375692368,2019-11-04 14:04:47 72 | 2,2,10,10,5,15,200,100,True,False,False,200,0.001,0.0,0.9185000117123127,0.8363333815336228,2019-11-04 14:04:50 73 | 2,2,10,10,5,15,200,100,False,True,False,200,0.001,0.0,0.9016667032241824,0.7091111263632774,2019-11-04 14:04:50 74 | 2,2,1,1,5,15,200,100,True,False,False,200,0.001,0.0,0.7300000093877317,0.5983333650231359,2019-11-04 13:58:37 75 | 2,2,10,10,5,15,200,100,False,False,True,200,0.001,0.0,0.9246667060256003,0.842222231626511,2019-11-04 14:05:44 76 | 2,2,10,10,5,15,200,100,True,True,False,200,0.001,0.0,0.8766667082905772,0.6917777949571607,2019-11-04 14:05:44 77 | 2,2,10,10,5,15,200,100,True,False,True,200,0.001,0.0,0.8530000516772273,0.7202222388982771,2019-11-04 14:05:44 78 | 2,2,10,10,5,15,200,100,False,True,True,200,0.001,0.0,0.9360000076889993,0.8833333754539492,2019-11-04 14:06:25 79 | 2,2,10,10,5,15,200,100,True,True,True,200,0.001,0.0,0.9200000083446505,0.7868333792686462,2019-11-04 14:06:25 80 | 2,2,15,15,5,15,200,100,False,False,False,200,0.001,0.0,0.9435000115633012,0.9073333758115767,2019-11-04 14:07:07 81 | 2,2,15,15,5,15,200,100,True,False,False,200,0.001,0.0,0.930500012487173,0.8540000510215758,2019-11-04 14:07:10 82 | 2,2,15,15,5,15,200,100,False,True,False,200,0.001,0.0,0.9350000289082528,0.7313333478569984,2019-11-04 14:07:10 83 | 2,2,15,15,5,15,200,100,False,False,True,200,0.001,0.0,0.9370000389218329,0.8475555646419521,2019-11-04 14:07:46 84 | 2,2,15,15,5,15,200,100,True,True,False,200,0.001,0.0,0.9283333653211598,0.693555572628975,2019-11-04 14:09:29 85 | 2,2,1,1,5,15,200,100,False,True,False,200,0.001,0.0,0.7923333737254143,0.5460000160336493,2019-11-04 13:58:37 86 | 2,2,15,15,5,15,200,100,True,False,True,200,0.001,0.0,0.8760000497102733,0.7515555703639984,2019-11-04 14:09:42 87 | 2,2,15,15,5,15,200,100,False,True,True,200,0.001,0.0,0.9397500064969061,0.8836667114496229,2019-11-04 14:10:42 88 | 2,2,15,15,5,15,200,100,True,True,True,200,0.001,0.0,0.9235000088810918,0.8046667128801347,2019-11-04 14:11:01 89 | 2,2,20,20,5,15,200,100,False,False,False,200,0.001,0.0,0.9470000118017196,0.9070000475645066,2019-11-04 14:11:05 90 | 2,2,20,20,5,15,200,100,True,False,False,200,0.001,0.0,0.9390000122785569,0.8583333832025524,2019-11-04 14:12:25 91 | 2,2,20,20,5,15,200,100,False,True,False,200,0.001,0.0,0.9203333637118339,0.755333346426487,2019-11-04 14:12:28 92 | 2,2,20,20,5,15,200,100,False,False,True,200,0.001,0.0,0.9466666990518565,0.8666666746139526,2019-11-04 14:12:53 93 | 2,2,20,20,5,15,200,100,True,True,False,200,0.001,0.0,0.9066667026281358,0.7206666818261145,2019-11-04 14:12:58 94 | 2,2,20,20,5,15,200,100,True,False,True,200,0.001,0.0,0.8943333846330638,0.7675555694103242,2019-11-04 14:14:35 95 | 2,2,20,20,5,15,200,100,False,True,True,200,0.001,0.0,0.9570000010728839,0.8931667107343675,2019-11-04 14:14:38 96 | 2,2,1,1,5,15,200,100,False,False,True,200,0.001,0.0,0.7406667080894113,0.632000018060207,2019-11-04 13:58:37 97 | 2,2,20,20,5,15,200,100,True,True,True,200,0.001,0.0,0.9185000097751618,0.7941667097806931,2019-11-04 14:15:56 98 | 3,3,1,1,5,15,200,100,False,False,False,200,0.001,0.0,0.7646667091548442,0.6268889065086841,2019-11-04 14:16:52 99 | 3,3,1,1,5,15,200,100,True,False,False,200,0.001,0.0,0.6286666993796828,0.470000012665987,2019-11-04 14:16:52 100 | 3,3,1,1,5,15,200,100,False,True,False,200,0.001,0.0,0.7320000115036964,0.4601666927337647,2019-11-04 14:18:14 101 | 3,3,1,1,5,15,200,100,False,False,True,200,0.001,0.0,0.6515000121295451,0.5375000277161597,2019-11-04 14:19:06 102 | 3,3,1,1,5,15,200,100,True,True,False,200,0.001,0.0,0.5902500125020739,0.38533335655927653,2019-11-04 14:19:09 103 | 3,3,1,1,5,15,200,100,True,False,True,200,0.001,0.0,0.5647500101476907,0.41850002244114876,2019-11-04 14:19:30 104 | 3,3,1,1,5,15,200,100,False,True,True,200,0.001,0.0,0.7091999800503253,0.6210666787624359,2019-11-04 14:19:45 105 | 3,3,1,1,5,15,200,100,True,True,True,200,0.001,0.0,0.6065999858826399,0.5100000119209293,2019-11-04 14:21:35 106 | 3,3,5,5,5,15,200,100,False,False,False,200,0.001,0.0,0.9180000442266462,0.8102222335338596,2019-11-04 14:21:35 107 | 2,2,1,1,5,15,200,100,True,True,False,200,0.001,0.0,0.5693333631008864,0.4202222333848477,2019-11-04 13:58:37 108 | 3,3,5,5,5,15,200,100,True,False,False,200,0.001,0.0,0.8583333814144134,0.7102222394943236,2019-11-04 14:22:00 109 | 3,3,5,5,5,15,200,100,False,True,False,200,0.001,0.0,0.8597500093281273,0.6493333706259731,2019-11-04 14:22:00 110 | 3,3,5,5,5,15,200,100,False,False,True,200,0.001,0.0,0.8815000158548353,0.77750004529953,2019-11-04 14:22:57 111 | 3,3,5,5,5,15,200,100,True,True,False,200,0.001,0.0,0.8260000124573706,0.5623333618044853,2019-11-04 14:23:40 112 | 3,3,5,5,5,15,200,100,True,False,True,200,0.001,0.0,0.8305000151693819,0.6111666959524154,2019-11-04 14:23:44 113 | 3,3,5,5,5,15,200,100,False,True,True,200,0.001,0.0,0.9063999760150913,0.7917333525419236,2019-11-04 14:24:16 114 | 3,3,5,5,5,15,200,100,True,True,True,200,0.001,0.0,0.8273999756574629,0.6632000136375429,2019-11-04 14:24:50 115 | 3,3,10,10,5,15,200,100,False,False,False,200,0.001,0.0,0.9286667042970657,0.8133333444595334,2019-11-04 14:24:59 116 | 3,3,10,10,5,15,200,100,True,False,False,200,0.001,0.0,0.8763333791494369,0.7566666811704634,2019-11-04 14:25:10 117 | 3,3,10,10,5,15,200,100,False,True,False,200,0.001,0.0,0.890500011146069,0.6863333681225778,2019-11-04 14:26:02 118 | 2,2,1,1,5,15,200,100,True,False,True,200,0.001,0.0,0.6126666985452177,0.5062222382426262,2019-11-04 13:58:37 119 | 3,3,10,10,5,15,200,100,False,False,True,200,0.001,0.0,0.9145000079274179,0.7925000464916225,2019-11-04 14:26:50 120 | 3,3,10,10,5,15,200,100,True,True,False,200,0.001,0.0,0.8617500132322309,0.6370000323653222,2019-11-04 14:27:50 121 | 3,3,10,10,5,15,200,100,True,False,True,200,0.001,0.0,0.8627500146627428,0.6980000376701356,2019-11-04 14:27:58 122 | 3,3,10,10,5,15,200,100,False,True,True,200,0.001,0.0,0.9263999786972998,0.8430666851997376,2019-11-04 14:28:28 123 | 3,3,10,10,5,15,200,100,True,True,True,200,0.001,0.0,0.8855999752879146,0.7313333511352541,2019-11-04 14:28:41 124 | 3,3,15,15,5,15,200,100,False,False,False,200,0.001,0.0,0.932333372235298,0.868000007867813,2019-11-04 14:30:11 125 | 3,3,15,15,5,15,200,100,True,False,False,200,0.001,0.0,0.9010000437498088,0.7735555690526963,2019-11-04 14:30:14 126 | 3,3,15,15,5,15,200,100,False,True,False,200,0.001,0.0,0.8985000061988833,0.6933333724737168,2019-11-04 14:30:20 127 | 3,3,15,15,5,15,200,100,False,False,True,200,0.001,0.0,0.923750009536743,0.8281667143106459,2019-11-04 14:30:30 128 | 3,3,15,15,5,15,200,100,True,True,False,200,0.001,0.0,0.8937500101327895,0.6501667049527169,2019-11-04 14:30:52 129 | 2,2,1,1,5,15,200,100,False,True,True,200,0.001,0.0,0.7820000122487547,0.709666705429554,2019-11-04 13:58:40 130 | 3,3,15,15,5,15,200,100,True,False,True,200,0.001,0.0,0.8770000138878825,0.6936667037010193,2019-11-04 14:31:35 131 | 3,3,15,15,5,15,200,100,False,True,True,200,0.001,0.0,0.928999977111816,0.8530666828155518,2019-11-04 14:31:41 132 | 3,3,15,15,5,15,200,100,True,True,True,200,0.001,0.0,0.90199997484684,0.7614666837453845,2019-11-04 14:34:02 133 | 3,3,20,20,5,15,200,100,False,False,False,200,0.001,0.0,0.9503333631157876,0.84911112010479,2019-11-04 14:35:29 134 | 3,3,20,20,5,15,200,100,True,False,False,200,0.001,0.0,0.9196667090058328,0.8008889007568358,2019-11-04 14:36:00 135 | 3,3,20,20,5,15,200,100,False,True,False,200,0.001,0.0,0.8972500075399875,0.7276667082309723,2019-11-04 14:36:48 136 | 3,3,20,20,5,15,200,100,False,False,True,200,0.001,0.0,0.9290000078082085,0.8461667102575303,2019-11-04 14:38:02 137 | 3,3,20,20,5,15,200,100,True,True,False,200,0.001,0.0,0.8770000085234645,0.6683333688974378,2019-11-04 14:38:24 138 | 3,3,20,20,5,15,200,100,True,False,True,200,0.001,0.0,0.9035000100731846,0.7131667065620418,2019-11-04 14:38:27 139 | 3,3,20,20,5,15,200,100,False,True,True,200,0.001,0.0,0.9433999791741371,0.8704000210762022,2019-11-04 14:39:09 140 | 2,2,1,1,5,15,200,100,True,True,True,200,0.001,0.0,0.6600000125169749,0.5986667010188105,2019-11-04 13:58:40 141 | 3,3,20,20,5,15,200,100,True,True,True,200,0.001,0.0,0.9227999779582018,0.7692000186443326,2019-11-04 14:39:12 142 | 4,4,1,1,5,15,200,100,False,False,False,200,0.001,0.0,0.6667500144243238,0.480500026345253,2019-11-04 14:39:20 143 | 4,4,1,1,5,15,200,100,True,False,False,200,0.001,0.0,0.5612500116229059,0.4035000221431255,2019-11-04 14:40:03 144 | 4,4,1,1,5,15,200,100,False,True,False,200,0.001,0.0,0.630599984079599,0.39080000817775734,2019-11-04 14:40:49 145 | 4,4,1,1,5,15,200,100,False,False,True,200,0.001,0.0,0.6059999834746119,0.45080001115798957,2019-11-04 14:42:24 146 | 4,4,1,1,5,15,200,100,True,True,False,200,0.001,0.0,0.5051999877393243,0.34640000730752946,2019-11-04 14:43:37 147 | 4,4,1,1,5,15,200,100,True,False,True,200,0.001,0.0,0.513799987360835,0.3608000084757803,2019-11-04 14:44:36 148 | 4,4,1,1,5,15,200,100,False,True,True,200,0.001,0.0,0.6708333706855776,0.5450000083446505,2019-11-04 14:45:11 149 | 4,4,1,1,5,15,200,100,True,True,True,200,0.001,0.0,0.5466666948795317,0.43833333939313895,2019-11-04 14:45:26 150 | 4,4,5,5,5,15,200,100,False,False,False,200,0.001,0.0,0.8827500140666957,0.7261667072772976,2019-11-04 14:46:38 151 | 2,2,5,5,5,15,200,100,False,False,False,200,0.001,0.0,0.9175000111758707,0.8743333768844601,2019-11-04 13:58:40 152 | 4,4,5,5,5,15,200,100,True,False,False,200,0.001,0.0,0.8522500127553941,0.6518333697319033,2019-11-04 14:47:34 153 | 4,4,5,5,5,15,200,100,False,True,False,200,0.001,0.0,0.8539999759197235,0.5830666786432266,2019-11-04 14:47:40 154 | 4,4,5,5,5,15,200,100,False,False,True,200,0.001,0.0,0.8747999766468998,0.7317333507537841,2019-11-04 14:48:05 155 | 4,4,5,5,5,15,200,100,True,True,False,200,0.001,0.0,0.8073999780416492,0.520666679143906,2019-11-04 14:48:10 156 | 4,4,5,5,5,15,200,100,True,False,True,200,0.001,0.0,0.8217999750375745,0.5608000135421752,2019-11-04 14:48:53 157 | 4,4,5,5,5,15,200,100,False,True,True,200,0.001,0.0,0.8873333823680877,0.7620000237226486,2019-11-04 14:50:10 158 | 4,4,5,5,5,15,200,100,True,True,True,200,0.001,0.0,0.8415000486373911,0.6304444539546967,2019-11-04 14:50:20 159 | 4,4,10,10,5,15,200,100,False,False,False,200,0.001,0.0,0.9195000088214874,0.7745000469684603,2019-11-04 14:50:29 160 | 4,4,10,10,5,15,200,100,True,False,False,200,0.001,0.0,0.8775000110268596,0.7093333739042282,2019-11-04 14:50:50 161 | 4,4,10,10,5,15,200,100,False,True,False,200,0.001,0.0,0.8823999783396719,0.6412000107765194,2019-11-04 14:51:10 162 | -------------------------------------------------------------------------------- /results/performance-Core+Background.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ArchitParnami/Few-Shot-KWS/04823543110be9d186b3a5ecbc695548a6cec097/results/performance-Core+Background.png -------------------------------------------------------------------------------- /results/performance-Core+Unknown+Background+Silence.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ArchitParnami/Few-Shot-KWS/04823543110be9d186b3a5ecbc695548a6cec097/results/performance-Core+Unknown+Background+Silence.png -------------------------------------------------------------------------------- /results/performance-Core+Unknown.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ArchitParnami/Few-Shot-KWS/04823543110be9d186b3a5ecbc695548a6cec097/results/performance-Core+Unknown.png -------------------------------------------------------------------------------- /results/performance-Core.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ArchitParnami/Few-Shot-KWS/04823543110be9d186b3a5ecbc695548a6cec097/results/performance-Core.png -------------------------------------------------------------------------------- /results/results.tgn: -------------------------------------------------------------------------------- 1 | {"rows_views":[[{"style":{"borders":"ltrb","font_style":{"bold":true},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"ltrb","font_style":{"bold":true},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"ltrb","font_style":{"bold":true},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"ltrb","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"ltrb","font_style":{"bold":true},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"ltrb","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}}],[{"style":{"borders":"ltrb","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"ltrb","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"ltrb","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"ltrb","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"ltrb","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"ltrb","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}}],[{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}}],[{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}}],[{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}}],[{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{"bold":true},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{"bold":true},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{"bold":true},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{"bold":true},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}}],[{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}}],[{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}}],[{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{"bold":true},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}}],[{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{"bold":true},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{"bold":true},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{"bold":true},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}}],[{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}}],[{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}}],[{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}}],[{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{"bold":true},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{"bold":true},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{"bold":true},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{"bold":true},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}}],[{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}}],[{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}}],[{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}}],[{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{"bold":true},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{"bold":true},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{"bold":true},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}},{"style":{"borders":"lr","font_style":{"bold":true},"text_color":"","bg_color":"","halign":"center","valign":"top","padding":{"top":10,"right":5,"bottom":10,"left":5},"border_color":""}}]],"model":{"rows":[[{"value":"Case","cspan":1,"rspan":2,"markup":[1,4]},{"value":"Embedding\nNetwork\n","cspan":1,"rspan":2,"markup":[1,18]},{"value":"2-way Acc.","cspan":2,"rspan":1,"markup":[1,10]},{"value":"","cspan":-1,"rspan":1,"markup":[]},{"value":"4-way Acc.","cspan":2,"rspan":1,"markup":[1,10]},{"value":"","cspan":-1,"rspan":1,"markup":[]}],[{"value":"","cspan":1,"rspan":-1,"markup":[]},{"value":"","cspan":1,"rspan":-1,"markup":[]},{"value":"1-shot","cspan":1,"rspan":1,"markup":[1,6]},{"value":"5-shot","cspan":1,"rspan":1,"markup":[1,6]},{"value":"1-shot","cspan":1,"rspan":1,"markup":[1,6]},{"value":"5-shot","cspan":1,"rspan":1,"markup":[1,6]}],[{"value":"core","cspan":1,"rspan":4,"markup":[1,4]},{"value":"cnn_trad_fpool3","cspan":1,"rspan":1,"markup":[0,15]},{"value":"69.23 $\\pm$ 0.03","cspan":1,"rspan":1,"markup":[0,16]},{"value":"87.07 $\\pm$ 0.02","cspan":1,"rspan":1,"markup":[0,16]},{"value":"48.83 $\\pm$ 0.02","cspan":1,"rspan":1,"markup":[0,16]},{"value":"75.93 $\\pm$ 0.01","cspan":1,"rspan":1,"markup":[0,16]}],[{"value":"core","cspan":1,"rspan":-1,"markup":[0,4]},{"value":"C64","cspan":1,"rspan":1,"markup":[0,3]},{"value":"77.20 $\\pm$ 0.03","cspan":1,"rspan":1,"markup":[1,16]},{"value":"89.97 $\\pm$ 0.02","cspan":1,"rspan":1,"markup":[0,16]},{"value":"62.63 $\\pm$ 0.02","cspan":1,"rspan":1,"markup":[0,16]},{"value":"80.48 $\\pm$ 0.01","cspan":1,"rspan":1,"markup":[0,16]}],[{"value":"core","cspan":1,"rspan":-2,"markup":[0,4]},{"value":"TCResNet8","cspan":1,"rspan":1,"markup":[0,9]},{"value":"82.70 $\\pm$ 0.03","cspan":1,"rspan":1,"markup":[1,16]},{"value":"89.00 $\\pm$ 0.02","cspan":1,"rspan":1,"markup":[1,16]},{"value":"69.47 $\\pm$ 0.02","cspan":1,"rspan":1,"markup":[0,16]},{"value":"81.20 $\\pm$ 0.01","cspan":1,"rspan":1,"markup":[1,16]}],[{"value":"core","cspan":1,"rspan":-3,"markup":[0,4]},{"value":"TDResNet8 (ours)","cspan":1,"rspan":1,"markup":[1,16]},{"value":"85.43 $\\pm$ 0.03","cspan":1,"rspan":1,"markup":[0,16]},{"value":"94.10 $\\pm$ 0.01","cspan":1,"rspan":1,"markup":[1,16]},{"value":"75.22 $\\pm$ 0.02","cspan":1,"rspan":1,"markup":[0,16]},{"value":"83.48 $\\pm$ 0.02","cspan":1,"rspan":1,"markup":[0,16]}],[{"value":"core \n+ \nbackground","cspan":1,"rspan":4,"markup":[1,19]},{"value":"cnn_trad_fpool3","cspan":1,"rspan":1,"markup":[0,15]},{"value":"69.53 $\\pm$ 0.04","cspan":1,"rspan":1,"markup":[0,16]},{"value":"86.8 $\\pm$ 0.02","cspan":1,"rspan":1,"markup":[0,15]},{"value":"43.3 $\\pm$ 0.02","cspan":1,"rspan":1,"markup":[0,15]},{"value":"67.42 $\\pm$ 0.01","cspan":1,"rspan":1,"markup":[0,16]}],[{"value":"core + background","cspan":1,"rspan":-1,"markup":[0,17]},{"value":"C64","cspan":1,"rspan":1,"markup":[0,3]},{"value":"78.30 $\\pm$ 0.03","cspan":1,"rspan":1,"markup":[1,16]},{"value":"90.03 $\\pm$ 0.02","cspan":1,"rspan":1,"markup":[0,16]},{"value":"58.83 $\\pm$ 0.02","cspan":1,"rspan":1,"markup":[0,16]},{"value":"80.52 $\\pm$ 0.01","cspan":1,"rspan":1,"markup":[0,16]}],[{"value":"core + background","cspan":1,"rspan":-2,"markup":[0,17]},{"value":"TCResNet8","cspan":1,"rspan":1,"markup":[1,9]},{"value":"77.40 $\\pm$ 0.03","cspan":1,"rspan":1,"markup":[1,16]},{"value":"91.40 $\\pm$ 0.02","cspan":1,"rspan":1,"markup":[1,16]},{"value":"64.23 $\\pm$ 0.02","cspan":1,"rspan":1,"markup":[0,16]},{"value":"79.25 $\\pm$ 0.01","cspan":1,"rspan":1,"markup":[0,16]}],[{"value":"core + background","cspan":1,"rspan":-3,"markup":[0,17]},{"value":"TDResNet8 (ours)","cspan":1,"rspan":1,"markup":[1,16]},{"value":"82.23 $\\pm$ 0.03","cspan":1,"rspan":1,"markup":[1,16]},{"value":"91.00 $\\pm$ 0.02","cspan":1,"rspan":1,"markup":[1,16]},{"value":"71.58 $\\pm$ 0.02","cspan":1,"rspan":1,"markup":[0,16]},{"value":"85.65 $\\pm$ 0.01","cspan":1,"rspan":1,"markup":[0,16]}],[{"value":"core \n+ \nunknown","cspan":1,"rspan":4,"markup":[1,16]},{"value":"cnn_trad_fpool3","cspan":1,"rspan":1,"markup":[0,15]},{"value":"58.33 $\\pm$ 0.03","cspan":1,"rspan":1,"markup":[0,16]},{"value":"78.36 $\\pm$ 0.02","cspan":1,"rspan":1,"markup":[0,16]},{"value":"50.15 $\\pm$ 0.02","cspan":1,"rspan":1,"markup":[0,16]},{"value":"69.25 $\\pm$ 0.01","cspan":1,"rspan":1,"markup":[0,16]}],[{"value":"core + unknown","cspan":1,"rspan":-1,"markup":[0,14]},{"value":"C64","cspan":1,"rspan":1,"markup":[0,3]},{"value":"63.42 $\\pm$ 0.03","cspan":1,"rspan":1,"markup":[0,16]},{"value":"78.47 $\\pm$ 0.02","cspan":1,"rspan":1,"markup":[0,16]},{"value":"53.69 $\\pm$ 0.02","cspan":1,"rspan":1,"markup":[0,16]},{"value":"76.43 $\\pm$ 0.01","cspan":1,"rspan":1,"markup":[0,16]}],[{"value":"core + unknown","cspan":1,"rspan":-2,"markup":[0,14]},{"value":"TCResNet8","cspan":1,"rspan":1,"markup":[1,9]},{"value":"68.84 $\\pm$ 0.03","cspan":1,"rspan":1,"markup":[0,16]},{"value":"80.49 $\\pm$ 0.02","cspan":1,"rspan":1,"markup":[0,16]},{"value":"59.08 $\\pm$ 0.02","cspan":1,"rspan":1,"markup":[0,16]},{"value":"78.07 $\\pm$ 0.01","cspan":1,"rspan":1,"markup":[0,16]}],[{"value":"core + unknown","cspan":1,"rspan":-3,"markup":[0,14]},{"value":"TDResNet8 (ours)","cspan":1,"rspan":1,"markup":[1,16]},{"value":"77.24 $\\pm$ 0.02","cspan":1,"rspan":1,"markup":[0,16]},{"value":"87.22 $\\pm$ 0.01","cspan":1,"rspan":1,"markup":[0,16]},{"value":"70.45 $\\pm$ 0.02","cspan":1,"rspan":1,"markup":[0,16]},{"value":"81.88 $\\pm$ 0.01","cspan":1,"rspan":1,"markup":[0,16]}],[{"value":"core +  \nunknown +\nbackground +\nsilence","cspan":1,"rspan":4,"markup":[1,39]},{"value":"cnn_trad_fpool3","cspan":1,"rspan":1,"markup":[0,15]},{"value":"67.43 $\\pm$ 0.02","cspan":1,"rspan":1,"markup":[0,16]},{"value":"82.32 $\\pm$ 0.01","cspan":1,"rspan":1,"markup":[0,16]},{"value":"53.51 $\\pm$ 0.02","cspan":1,"rspan":1,"markup":[0,16]},{"value":"74.54 $\\pm$ 0.01","cspan":1,"rspan":1,"markup":[0,16]}],[{"value":"core + background + silence + unknown","cspan":1,"rspan":-1,"markup":[0,37]},{"value":"C64","cspan":1,"rspan":1,"markup":[0,3]},{"value":"65.83 $\\pm$ 0.02","cspan":1,"rspan":1,"markup":[0,16]},{"value":"81.15 $\\pm$ 0.01","cspan":1,"rspan":1,"markup":[0,16]},{"value":"56.38 $\\pm$ 0.01","cspan":1,"rspan":1,"markup":[0,16]},{"value":"73.20 $\\pm$ 0.01","cspan":1,"rspan":1,"markup":[1,16]}],[{"value":"core + background + silence + unknown","cspan":1,"rspan":-2,"markup":[0,37]},{"value":"TCResNet8","cspan":1,"rspan":1,"markup":[0,9]},{"value":"78.63 $\\pm$ 0.02","cspan":1,"rspan":1,"markup":[0,16]},{"value":"85.98 $\\pm$ 0.01","cspan":1,"rspan":1,"markup":[0,16]},{"value":"63.37 $\\pm$ 0.02","cspan":1,"rspan":1,"markup":[0,16]},{"value":"80.39 $\\pm$ 0.01","cspan":1,"rspan":1,"markup":[0,16]}],[{"value":"core + background + silence + unknown","cspan":1,"rspan":-3,"markup":[0,37]},{"value":"TDResNet8 (ours)","cspan":1,"rspan":1,"markup":[1,16]},{"value":"82.77 $\\pm$ 0.02","cspan":1,"rspan":1,"markup":[0,16]},{"value":"89.45 $\\pm$ 0.01","cspan":1,"rspan":1,"markup":[0,16]},{"value":"69.34 $\\pm$ 0.01","cspan":1,"rspan":1,"markup":[0,16]},{"value":"82.50 $\\pm$ 0.01","cspan":1,"rspan":1,"markup":[1,16]}]]},"theme":null,"fixed_layout":false,"markup":{"instances":[{},{"style":{}}]},"options":{"table_caption":"","table_label":"tab:my-table"}} -------------------------------------------------------------------------------- /results/table.csv: -------------------------------------------------------------------------------- 1 | core,cnn_trad_fpool3,69.23 $\pm$ 0.03,87.07 $\pm$ 0.02,48.83 $\pm$ 0.02,75.93 $\pm$ 0.01 2 | core,C64,77.2 $\pm$ 0.03,89.97 $\pm$ 0.02,62.63 $\pm$ 0.02,80.48 $\pm$ 0.01 3 | core,TCResNet8,82.7 $\pm$ 0.03,89.0 $\pm$ 0.02,69.47 $\pm$ 0.02,81.2 $\pm$ 0.01 4 | core,TCResNetDilated,85.43 $\pm$ 0.03,94.1 $\pm$ 0.01,75.22 $\pm$ 0.02,83.48 $\pm$ 0.02 5 | core + background,cnn_trad_fpool3,69.53 $\pm$ 0.04,86.8 $\pm$ 0.02,43.3 $\pm$ 0.02,67.42 $\pm$ 0.01 6 | core + background,C64,78.3 $\pm$ 0.03,90.03 $\pm$ 0.02,58.83 $\pm$ 0.02,80.52 $\pm$ 0.01 7 | core + background,TCResNet8,77.4 $\pm$ 0.03,91.4 $\pm$ 0.02,64.23 $\pm$ 0.02,79.25 $\pm$ 0.01 8 | core + background,TCResNetDilated,82.23 $\pm$ 0.03,91.0 $\pm$ 0.02,71.58 $\pm$ 0.02,85.65 $\pm$ 0.01 9 | core + unknown,cnn_trad_fpool3,58.33 $\pm$ 0.03,78.36 $\pm$ 0.02,50.15 $\pm$ 0.02,69.25 $\pm$ 0.01 10 | core + unknown,C64,63.42 $\pm$ 0.03,78.47 $\pm$ 0.02,53.69 $\pm$ 0.02,76.43 $\pm$ 0.01 11 | core + unknown,TCResNet8,68.84 $\pm$ 0.03,80.49 $\pm$ 0.02,59.08 $\pm$ 0.02,78.07 $\pm$ 0.01 12 | core + unknown,TCResNetDilated,77.24 $\pm$ 0.02,87.22 $\pm$ 0.01,70.45 $\pm$ 0.02,81.88 $\pm$ 0.01 13 | core + background + silence + unknown,cnn_trad_fpool3,67.43 $\pm$ 0.02,82.32 $\pm$ 0.01,53.51 $\pm$ 0.02,74.54 $\pm$ 0.01 14 | core + background + silence + unknown,C64,65.83 $\pm$ 0.02,81.15 $\pm$ 0.01,56.38 $\pm$ 0.01,73.2 $\pm$ 0.01 15 | core + background + silence + unknown,TCResNet8,78.63 $\pm$ 0.02,85.98 $\pm$ 0.01,63.37 $\pm$ 0.02,80.39 $\pm$ 0.01 16 | core + background + silence + unknown,TCResNetDilated,82.77 $\pm$ 0.02,89.45 $\pm$ 0.01,69.34 $\pm$ 0.01,82.5 $\pm$ 0.01 17 | -------------------------------------------------------------------------------- /scripts/debug_args.py: -------------------------------------------------------------------------------- 1 | #!/usr/bin/env python3 2 | # -*- coding: utf-8 -*- 3 | """ 4 | Created on Tue Nov 5 09:23:09 2019 5 | 6 | @author: aparnami 7 | """ 8 | 9 | def get_args(arg_str): 10 | args = arg_str.split('\t\t') 11 | debug_args = [] 12 | for arg in args[1:]: 13 | argsVal = arg.split() 14 | for item in argsVal: 15 | result = '"{}",'.format(item) 16 | debug_args.append(result) 17 | 18 | print('\n'.join(debug_args)) 19 | -------------------------------------------------------------------------------- /scripts/parse_results.py: -------------------------------------------------------------------------------- 1 | import ast 2 | import sys 3 | import os 4 | import json 5 | import pandas as pd 6 | 7 | root_dir = '' 8 | 9 | 10 | def read_opt(opt_file): 11 | with open(opt_file, 'r') as rf: 12 | line = rf.readline() 13 | opt = json.loads(line) 14 | return opt 15 | 16 | def read_trace(trace_file, val_every=1): 17 | train_acc = []; train_loss = [] 18 | val_acc = []; val_loss = [] 19 | with open(trace_file, 'r') as rf: 20 | for line in rf: 21 | result = ast.literal_eval(line) 22 | train_epoch = result['epoch'] 23 | train_acc.append(result['train']['acc']) 24 | train_loss.append(result['train']['loss']) 25 | if train_epoch % val_every == 0: 26 | val_acc.append(result['val']['acc']) 27 | val_loss.append(result['val']['loss']) 28 | return {'train' : {'acc' : train_acc, 'loss' : train_loss}, 29 | 'val' : {'acc' : val_acc, 'loss' : val_loss, 'every' : val_every}} 30 | 31 | def get_best_results(trace): 32 | val_epochs = range(trace['val']['every'], 33 | len(trace['train']['acc'])+1, trace['val']['every']) 34 | min_loss_val = min(trace['val']['loss']) 35 | best_index = trace['val']['loss'].index(min_loss_val) 36 | best_epoch = val_epochs[best_index] 37 | min_loss_train = trace['train']['loss'][best_epoch-1] 38 | max_acc_train = trace['train']['acc'][best_epoch-1] 39 | max_acc_val = trace['val']['acc'][best_index] 40 | return {'train' : {'acc' : max_acc_train, 'loss' : min_loss_train}, 41 | 'val' : {'acc' : max_acc_val, 'loss' :min_loss_val}, 42 | 'epoch' : best_epoch} 43 | 44 | def read_eval(eval_file): 45 | if os.path.exists(eval_file): 46 | with open(eval_file, 'r') as rf: 47 | line = rf.readline() 48 | test = ast.literal_eval(line) 49 | return test['test'] 50 | else: 51 | return None 52 | 53 | def join_results(timestamp, opt, result, test): 54 | row = [opt['data.way'], opt['data.test_way'], opt['data.shot'], 55 | opt['data.test_shot'], opt['data.query'], opt['data.test_query'], 56 | opt['data.train_episodes'], opt['data.test_episodes'], 57 | opt['speech.include_background'], opt['speech.include_silence'], 58 | opt['speech.include_unknown'], opt['train.epochs'], opt['train.learning_rate'], 59 | opt['train.weight_decay'], result['train']['acc'], result['val']['acc'], 60 | timestamp] 61 | if test is not None: 62 | row += [test['loss']['mean'], test['loss']['confidence'], 63 | test['acc']['mean'], test['acc']['confidence']] 64 | return row 65 | 66 | def read_results(root_dir): 67 | dirs = sorted(os.listdir(root_dir)) 68 | columns = ['train.way', 'test.way', 'train.shot', 'test.shot', 69 | 'train.query', 'test.query', 'train.episodes', 'test.episodes', 70 | 'background', 'silence', 'unknown', 'epochs', 'lr', 'wd', 71 | 'train.acc', 'val.acc','timestamp', 72 | 'test.loss.mean', 'test.loss.confidence', 73 | 'test.acc.mean', 'test.acc.confidence'] 74 | df = pd.DataFrame(columns=columns) 75 | 76 | for i, result_dir in enumerate(dirs): 77 | result_path = os.path.join(root_dir, result_dir) 78 | timestamp = os.listdir(result_path)[0] 79 | opt =read_opt(os.path.join(result_path, timestamp, 'opt.json')) 80 | trace = read_trace(os.path.join(result_path, timestamp, 'trace.txt')) 81 | result = get_best_results(trace) 82 | test = read_eval(os.path.join(result_path, timestamp, 'eval.txt')) 83 | row = join_results(timestamp, opt, result, test) 84 | df.loc[i] = row 85 | 86 | return df 87 | 88 | 89 | if __name__ == '__main__': 90 | result_root = sys.argv[1] 91 | df = read_results(result_root) 92 | if len(sys.argv) > 2: 93 | output_file = sys.argv[2] 94 | df.to_csv(output_file + '.csv', index=None, header=True) 95 | else: 96 | print(df) 97 | 98 | -------------------------------------------------------------------------------- /scripts/plot.py: -------------------------------------------------------------------------------- 1 | import ast 2 | import sys 3 | import matplotlib.pyplot as plt 4 | import os 5 | import json 6 | 7 | def plot_results(result_dir): 8 | train_acc = []; train_loss = [] 9 | val_acc = []; val_loss = [] 10 | result_file = os.path.join(result_dir, 'trace.txt') 11 | opt_file = os.path.join(result_dir, 'opt.json') 12 | opt = {} 13 | #get the options 14 | with open(opt_file, 'r') as rf: 15 | line = rf.readline() 16 | opt = json.loads(line) 17 | 18 | #get the results 19 | val_every = 1 20 | with open(result_file, 'r') as rf: 21 | for line in rf: 22 | result = ast.literal_eval(line) 23 | train_epoch = result['epoch'] 24 | train_acc.append(result['train']['acc']) 25 | train_loss.append(result['train']['loss']) 26 | if train_epoch % val_every == 0: 27 | val_acc.append(result['val']['acc']) 28 | val_loss.append(result['val']['loss']) 29 | 30 | train_epochs = range(1, len(train_acc)+1) 31 | val_epochs = range(val_every, len(train_acc)+1, val_every) 32 | plt.figure() 33 | plt.subplot(2,1,1) 34 | plt.plot(train_epochs, train_loss, label='train') 35 | plt.plot(val_epochs, val_loss, label='val') 36 | plt.ylabel('Loss') 37 | plt.grid(True) 38 | plt.legend() 39 | 40 | data_opt = 'Way,Shot,Query = (Train:[{},{},{}], Val:[{} {} {}])'.format( 41 | opt['data.way'], opt['data.shot'], opt['data.query'], 42 | opt['data.test_way'], opt['data.test_shot'], opt['data.test_query']) 43 | 44 | best_loss = min(val_loss) 45 | best_index = val_loss.index(best_loss) 46 | best_epoch = val_epochs[best_index] 47 | min_loss = 'Best Loss = (Train : {:.4}, Val: {:.4})'.format(train_loss[best_epoch-1], best_loss) 48 | max_acc = 'Best Accuracy = (Train : {:.4}, Val: {:.4})'.format(train_acc[best_epoch-1], val_acc[best_index]) 49 | title = '{}\n{}\n{}'.format(data_opt, min_loss, max_acc) 50 | plt.title(title) 51 | plt.plot([best_epoch, best_epoch],[train_loss[best_epoch-1],best_loss], '--bo') 52 | 53 | plt.subplot(2,1,2) 54 | plt.plot(train_epochs, train_acc, label='train') 55 | plt.plot(val_epochs, val_acc, label='val') 56 | plt.xlabel('Number of Epochs') 57 | plt.ylabel('Accuracy') 58 | plt.grid(True) 59 | plt.legend() 60 | 61 | save_file = os.path.join(result_dir , 'plot.png') 62 | plt.savefig(save_file) 63 | 64 | 65 | if __name__ == '__main__': 66 | result_dir = sys.argv[1] 67 | experiments = os.listdir(result_dir) 68 | for exp in experiments: 69 | exp_path = os.path.join(result_dir, exp) 70 | if os.path.isdir(exp_path): 71 | plot_results(exp_path) 72 | 73 | 74 | 75 | -------------------------------------------------------------------------------- /scripts/predict/few_shot/eval.py: -------------------------------------------------------------------------------- 1 | import os 2 | import json 3 | import math 4 | from tqdm import tqdm 5 | 6 | import torch 7 | import torchnet as tnt 8 | 9 | from protonets.utils import filter_opt, merge_dict 10 | import protonets.utils.data as data_utils 11 | import protonets.utils.model as model_utils 12 | 13 | def main(opt): 14 | # load model 15 | model = torch.load(opt['model.model_path']) 16 | model.eval() 17 | 18 | # load opts 19 | model_opt_file = os.path.join(os.path.dirname(opt['model.model_path']), 'opt.json') 20 | with open(model_opt_file, 'r') as f: 21 | model_opt = json.load(f) 22 | 23 | # Postprocess arguments 24 | model_opt['model.x_dim'] = map(int, model_opt['model.x_dim'].split(',')) 25 | model_opt['log.fields'] = model_opt['log.fields'].split(',') 26 | 27 | # construct data 28 | data_opt = { 'data.' + k: v for k,v in filter_opt(model_opt, 'data').items() } 29 | 30 | 31 | # episode_fields = { 32 | # 'data.test_way': 'data.way', 33 | # 'data.test_shot': 'data.shot', 34 | # 'data.test_query': 'data.query', 35 | # 'data.test_episodes': 'data.train_episodes' 36 | # } 37 | 38 | # for k,v in episode_fields.items(): 39 | # if opt[k] != 0: 40 | # data_opt[k] = opt[k] 41 | # elif model_opt[k] != 0: 42 | # data_opt[k] = model_opt[k] 43 | # else: 44 | # data_opt[k] = model_opt[v] 45 | 46 | print("Evaluating {:d}-way, {:d}-shot with {:d} query examples/class over {:d} episodes".format( 47 | data_opt['data.test_way'], data_opt['data.test_shot'], 48 | data_opt['data.test_query'], data_opt['data.test_episodes'])) 49 | 50 | torch.manual_seed(1234) 51 | if data_opt['data.cuda']: 52 | torch.cuda.manual_seed(1234) 53 | 54 | data = data_utils.load(model_opt, ['test']) 55 | 56 | if data_opt['data.cuda']: 57 | model.cuda() 58 | 59 | meters = { field: tnt.meter.AverageValueMeter() for field in model_opt['log.fields'] } 60 | 61 | model_utils.evaluate(model, data['test'], meters, desc="test") 62 | 63 | output = {"test" : {}} 64 | for field,meter in meters.items(): 65 | mean, std = meter.value() 66 | output["test"][field] = {} 67 | output["test"][field]["mean"] = mean 68 | output["test"][field]["confidence"] = 1.96 * std / math.sqrt(data_opt['data.test_episodes']) 69 | #print("test {:s}: {:0.6f} +/- {:0.6f}".format(field, mean, 1.96 * std / math.sqrt(data_opt['data.test_episodes']))) 70 | 71 | output_file = os.path.join(os.path.dirname(opt['model.model_path']), 'eval.txt') 72 | with open(output_file, 'w') as fp: 73 | json.dump(output, fp) -------------------------------------------------------------------------------- /scripts/predict/few_shot/eval_results.py: -------------------------------------------------------------------------------- 1 | import os 2 | import sys 3 | from eval import main 4 | import argparse 5 | 6 | def eval_model(model_path): 7 | args = {'model.model_path' : model_path} 8 | main(args) 9 | 10 | def evaluate(root_dir): 11 | dirs = sorted(os.listdir(root_dir)) 12 | for i, result_dir in enumerate(dirs): 13 | result_path = os.path.join(root_dir, result_dir) 14 | timestamp = os.listdir(result_path)[0] 15 | model_path = os.path.join(result_path, timestamp, 'best_model.pt') 16 | eval_model(model_path) 17 | 18 | if __name__ == '__main__': 19 | result_root = sys.argv[1] 20 | evaluate(result_root) 21 | 22 | 23 | -------------------------------------------------------------------------------- /scripts/predict/few_shot/qsub_eval.sh: -------------------------------------------------------------------------------- 1 | #!/bin/sh 2 | # 3 | # ======= PBS OPTIONS ======= (user input required) 4 | # 5 | ### Specify queue to run 6 | #PBS -q titan 7 | ### Set the job name 8 | #PBS -N FKSeval 9 | ### Specify the # of cpus for your job. 10 | #PBS -l nodes=1:ppn=1:gpus=1,mem=15GB 11 | #PBS -l walltime=24:00:00 12 | ### pass the full environment 13 | #PBS -V 14 | # 15 | # ===== END PBS OPTIONS ===== 16 | 17 | ### run job 18 | 19 | source activate pyenv 20 | 21 | cd $PBS_O_WORKDIR 22 | 23 | python eval_results.py "$result_dir" 24 | 25 | conda deactivate 26 | -------------------------------------------------------------------------------- /scripts/predict/few_shot/run_eval.py: -------------------------------------------------------------------------------- 1 | import argparse 2 | 3 | from eval import main 4 | 5 | parser = argparse.ArgumentParser(description='Evaluate few-shot prototypical networks') 6 | 7 | default_model_path = 'results/best_model.pt' 8 | parser.add_argument('--model.model_path', type=str, default=default_model_path, metavar='MODELPATH', 9 | help="location of pretrained model to evaluate (default: {:s})".format(default_model_path)) 10 | 11 | parser.add_argument('--data.test_way', type=int, default=0, metavar='TESTWAY', 12 | help="number of classes per episode in test. 0 means same as model's data.test_way (default: 0)") 13 | parser.add_argument('--data.test_shot', type=int, default=0, metavar='TESTSHOT', 14 | help="number of support examples per class in test. 0 means same as model's data.shot (default: 0)") 15 | parser.add_argument('--data.test_query', type=int, default=0, metavar='TESTQUERY', 16 | help="number of query examples per class in test. 0 means same as model's data.query (default: 0)") 17 | parser.add_argument('--data.test_episodes', type=int, default=0, metavar='NTEST', 18 | help="number of test episodes per epoch (default: 0)") 19 | 20 | args = vars(parser.parse_args()) 21 | 22 | main(args) 23 | -------------------------------------------------------------------------------- /scripts/predict/few_shot/submit.sh: -------------------------------------------------------------------------------- 1 | qsub -v result_dir=$1 qsub_eval.sh -------------------------------------------------------------------------------- /scripts/train/few_shot/fewshotspeech/loop_qsub.sh: -------------------------------------------------------------------------------- 1 | id=1 2 | for way in 2 3 4 5 3 | do 4 | for shot in 1 5 10 15 20 5 | do 6 | for i in 0 1 2 3 4 5 6 7 7 | do 8 | qsub -v way=$way,shot=$shot,flag=$i,id=$id qsub_train.sh 9 | id=$((id+1)) 10 | done 11 | done 12 | 13 | 14 | done 15 | -------------------------------------------------------------------------------- /scripts/train/few_shot/fewshotspeech/qsub_train.sh: -------------------------------------------------------------------------------- 1 | #!/bin/sh 2 | # 3 | # ======= PBS OPTIONS ======= (user input required) 4 | # 5 | ### Specify queue to run 6 | #PBS -q titan 7 | ### Set the job name 8 | #PBS -N FKS 9 | ### Specify the # of cpus for your job. 10 | #PBS -l nodes=1:ppn=1:gpus=1,mem=15GB 11 | #PBS -l walltime=24:00:00 12 | ### pass the full environment 13 | #PBS -V 14 | # 15 | # ===== END PBS OPTIONS ===== 16 | 17 | ### run job 18 | 19 | source activate pyenv 20 | 21 | cd $PBS_O_WORKDIR 22 | ./train.sh "$way" "$shot" "$flag" "$id" 23 | 24 | conda deactivate 25 | -------------------------------------------------------------------------------- /scripts/train/few_shot/fewshotspeech/train.sh: -------------------------------------------------------------------------------- 1 | way=$1 2 | shot=$2 3 | i=$3 4 | id=$4 5 | 6 | args="../run_train.py \ 7 | --data.dataset=googlespeech --data.way=$way --data.shot=$shot \ 8 | --data.query=5 --data.test_way=$way --data.test_shot=$shot --data.test_query=15 \ 9 | --data.train_episodes=200 --data.test_episodes=100 \ 10 | --model.model_name=protonet_conv --model.x_dim=1,51,40 --model.hid_dim=64 --model.z_dim=64 \ 11 | --train.epochs=200 --train.optim_method=Adam --train.learning_rate=0.001 --train.decay_every=20 \ 12 | --train.weight_decay=0.0 --train.patience=200 \ 13 | --log.exp_dir=./$id \ 14 | --data.cuda \ 15 | --model.encoding=TCResNet8Dilated" 16 | 17 | if [ "$i" -eq "1" ]; then 18 | args+=" --speech.include_background" 19 | 20 | elif [ "$i" -eq "2" ]; then 21 | args+=" --speech.include_silence" 22 | 23 | elif [ "$i" -eq "3" ]; then 24 | args+=" --speech.include_unknown" 25 | 26 | elif [ "$i" -eq "4" ]; then 27 | args+=" --speech.include_background" 28 | args+=" --speech.include_silence" 29 | 30 | elif [ "$i" -eq "5" ]; then 31 | args+=" --speech.include_background" 32 | args+=" --speech.include_unknown" 33 | 34 | elif [ "$i" -eq "6" ]; then 35 | args+=" --speech.include_unknown" 36 | args+=" --speech.include_silence" 37 | 38 | elif [ "$i" -eq "7" ]; then 39 | args+=" --speech.include_background" 40 | args+=" --speech.include_unknown" 41 | args+=" --speech.include_silence" 42 | fi 43 | 44 | python $args 45 | -------------------------------------------------------------------------------- /scripts/train/few_shot/fewshotspeech/train_fewshotspeech.sh: -------------------------------------------------------------------------------- 1 | for way in 2 2 | do 3 | for shot in 1 5 10 4 | do 5 | 6 | for i in 0 1 2 3 4 5 6 7 7 | do 8 | args="../run_train.py \ 9 | --data.dataset=googlespeech --data.way=$way --data.shot=$shot \ 10 | --data.query=5 --data.test_way=$way --data.test_shot=$shot --data.test_query=15 \ 11 | --data.train_episodes=100 --data.test_episodes=100 \ 12 | --model.model_name=protonet_conv --model.x_dim=1,51,40 --model.hid_dim=64 --model.z_dim=64 \ 13 | --train.epochs=200 --train.optim_method=Adam --train.learning_rate=0.001 --train.decay_every=20 \ 14 | --train.weight_decay=0.0 --train.patience=200 \ 15 | --log.exp_dir=. \ 16 | --data.cuda" 17 | 18 | if [ "$i" -eq "1" ]; then 19 | args+=" --speech.include_background" 20 | 21 | elif [ "$i" -eq "2" ]; then 22 | args+=" --speech.include_silence" 23 | 24 | elif [ "$i" -eq "3" ]; then 25 | args+=" --speech.include_unknown" 26 | 27 | elif [ "$i" -eq "4" ]; then 28 | args+=" --speech.include_background" 29 | args+=" --speech.include_silence" 30 | 31 | elif [ "$i" -eq "5" ]; then 32 | args+=" --speech.include_background" 33 | args+=" --speech.include_unknown" 34 | 35 | elif [ "$i" -eq "6" ]; then 36 | args+=" --speech.include_unknown" 37 | args+=" --speech.include_silence" 38 | 39 | elif [ "$i" -eq "7" ]; then 40 | args+=" --speech.include_background" 41 | args+=" --speech.include_unknown" 42 | args+=" --speech.include_silence" 43 | fi 44 | 45 | python $args 46 | 47 | done 48 | 49 | 50 | done 51 | 52 | 53 | done 54 | 55 | 56 | -------------------------------------------------------------------------------- /scripts/train/few_shot/run_train.py: -------------------------------------------------------------------------------- 1 | import argparse 2 | 3 | from train import main 4 | 5 | parser = argparse.ArgumentParser(description='Train prototypical networks') 6 | 7 | # data args 8 | default_dataset = 'omniglot' 9 | parser.add_argument('--data.dataset', type=str, default=default_dataset, metavar='DS', 10 | help="data set name (default: {:s})".format(default_dataset)) 11 | default_split = 'vinyals' 12 | parser.add_argument('--data.split', type=str, default=default_split, metavar='SP', 13 | help="split name (default: {:s})".format(default_split)) 14 | parser.add_argument('--data.way', type=int, default=60, metavar='WAY', 15 | help="number of classes per episode (default: 60)") 16 | parser.add_argument('--data.shot', type=int, default=5, metavar='SHOT', 17 | help="number of support examples per class (default: 5)") 18 | parser.add_argument('--data.query', type=int, default=5, metavar='QUERY', 19 | help="number of query examples per class (default: 5)") 20 | parser.add_argument('--data.test_way', type=int, default=5, metavar='TESTWAY', 21 | help="number of classes per episode in test. 0 means same as data.way (default: 5)") 22 | parser.add_argument('--data.test_shot', type=int, default=0, metavar='TESTSHOT', 23 | help="number of support examples per class in test. 0 means same as data.shot (default: 0)") 24 | parser.add_argument('--data.test_query', type=int, default=15, metavar='TESTQUERY', 25 | help="number of query examples per class in test. 0 means same as data.query (default: 15)") 26 | parser.add_argument('--data.train_episodes', type=int, default=100, metavar='NTRAIN', 27 | help="number of train episodes per epoch (default: 100)") 28 | parser.add_argument('--data.test_episodes', type=int, default=100, metavar='NTEST', 29 | help="number of test episodes per epoch (default: 100)") 30 | parser.add_argument('--data.trainval', action='store_true', help="run in train+validation mode (default: False)") 31 | parser.add_argument('--data.sequential', action='store_true', help="use sequential sampler instead of episodic (default: False)") 32 | parser.add_argument('--data.cuda', action='store_true', help="run in CUDA mode (default: False)") 33 | 34 | # model args 35 | default_model_name = 'protonet_conv' 36 | default_encoding = 'C64' 37 | parser.add_argument('--model.model_name', type=str, default=default_model_name, metavar='MODELNAME', 38 | help="model name (default: {:s})".format(default_model_name)) 39 | parser.add_argument('--model.x_dim', type=str, default='1,28,28', metavar='XDIM', 40 | help="dimensionality of input images (default: '1,28,28')") 41 | parser.add_argument('--model.hid_dim', type=int, default=64, metavar='HIDDIM', 42 | help="dimensionality of hidden layers (default: 64)") 43 | parser.add_argument('--model.z_dim', type=int, default=64, metavar='ZDIM', 44 | help="dimensionality of input images (default: 64)") 45 | parser.add_argument('--model.encoding', type=str, default=default_encoding, metavar='MODELENC', 46 | help="model encoding (default: {:s})".format(default_encoding)) 47 | # train args 48 | parser.add_argument('--train.epochs', type=int, default=10000, metavar='NEPOCHS', 49 | help='number of epochs to train (default: 10000)') 50 | parser.add_argument('--train.optim_method', type=str, default='Adam', metavar='OPTIM', 51 | help='optimization method (default: Adam)') 52 | parser.add_argument('--train.learning_rate', type=float, default=0.001, metavar='LR', 53 | help='learning rate (default: 0.0001)') 54 | parser.add_argument('--train.decay_every', type=int, default=20, metavar='LRDECAY', 55 | help='number of epochs after which to decay the learning rate') 56 | default_weight_decay = 0.0 57 | parser.add_argument('--train.weight_decay', type=float, default=default_weight_decay, metavar='WD', 58 | help="weight decay (default: {:f})".format(default_weight_decay)) 59 | parser.add_argument('--train.patience', type=int, default=200, metavar='PATIENCE', 60 | help='number of epochs to wait before validation improvement (default: 1000)') 61 | 62 | # log args 63 | default_fields = 'loss,acc' 64 | parser.add_argument('--log.fields', type=str, default=default_fields, metavar='FIELDS', 65 | help="fields to monitor during training (default: {:s})".format(default_fields)) 66 | default_exp_dir = 'results' 67 | parser.add_argument('--log.exp_dir', type=str, default=default_exp_dir, metavar='EXP_DIR', 68 | help="directory where experiments should be saved (default: {:s})".format(default_exp_dir)) 69 | 70 | # speech data args 71 | parser.add_argument('--speech.include_background', action='store_true', help="mix background noise with samples (default: False)") 72 | parser.add_argument('--speech.include_silence', action='store_true', help="one of the classes out of n should be silence (default: False)") 73 | parser.add_argument('--speech.include_unknown', action='store_true', help="one of the classes out of n should be unknown (default: False)") 74 | parser.add_argument('--speech.sample_rate', type=int, default=16000, help='desired sampling rate of the input') 75 | parser.add_argument('--speech.clip_duration', type=int, default=1000, help='clip duration in milliseconds') 76 | parser.add_argument('--speech.window_size', type=int, default=40) 77 | parser.add_argument('--speech.window_stride', type=int,default=20) 78 | parser.add_argument('--speech.num_features', type=int, default=40, help='Number of mfcc features to extract') 79 | parser.add_argument('--speech.time_shift', type=int, default=100, help='time shift the audio in milliseconds') 80 | parser.add_argument('--speech.bg_volume', type=float, default=0.1, help='background volumen to mix in between 0 and 1') 81 | parser.add_argument('--speech.bg_frequency', type=float, default=1.0, help='Amount of samples that should be mixed with background noise (between 0 and 1)') 82 | parser.add_argument('--speech.num_silence', type=int, default=1000, help='Number of silence samples to generate') 83 | parser.add_argument('--speech.foreground_volume', type=float, default=1) 84 | args = vars(parser.parse_args()) 85 | 86 | main(args) 87 | -------------------------------------------------------------------------------- /scripts/train/few_shot/run_trainval.py: -------------------------------------------------------------------------------- 1 | import argparse 2 | 3 | from trainval import main 4 | 5 | parser = argparse.ArgumentParser(description='Re-run prototypical networks training in trainval mode') 6 | 7 | parser.add_argument('--model.model_path', type=str, default='results/best_model.pt', metavar='MODELPATH', 8 | help="location of pretrained model to retrain in trainval mode") 9 | 10 | args = vars(parser.parse_args()) 11 | 12 | main(args) 13 | -------------------------------------------------------------------------------- /scripts/train/few_shot/train.py: -------------------------------------------------------------------------------- 1 | import os 2 | import json 3 | from functools import partial 4 | from tqdm import tqdm 5 | 6 | import numpy as np 7 | 8 | import torch 9 | import torch.optim as optim 10 | import torch.optim.lr_scheduler as lr_scheduler 11 | import torchvision 12 | import torchnet as tnt 13 | 14 | from protonets.engine import Engine 15 | 16 | import protonets.utils.data as data_utils 17 | import protonets.utils.model as model_utils 18 | import protonets.utils.log as log_utils 19 | from datetime import datetime, timedelta 20 | import time 21 | 22 | def main(opt): 23 | timestamp = datetime.now().strftime('%Y-%m-%d %H:%M:%S') 24 | opt['log.exp_dir'] = os.path.join('./results', opt['log.exp_dir'], timestamp) 25 | os.makedirs(opt['log.exp_dir']) 26 | 27 | # save opts 28 | with open(os.path.join(opt['log.exp_dir'], 'opt.json'), 'w') as f: 29 | json.dump(opt, f) 30 | f.write('\n') 31 | 32 | trace_file = os.path.join(opt['log.exp_dir'], 'trace.txt') 33 | 34 | # Postprocess arguments 35 | opt['model.x_dim'] = list(map(int, opt['model.x_dim'].split(','))) 36 | opt['log.fields'] = opt['log.fields'].split(',') 37 | 38 | torch.manual_seed(1234) 39 | if opt['data.cuda']: 40 | torch.cuda.manual_seed(1234) 41 | 42 | if opt['data.trainval']: 43 | data = data_utils.load(opt, ['trainval']) 44 | train_loader = data['trainval'] 45 | val_loader = None 46 | else: 47 | data = data_utils.load(opt, ['train', 'val']) 48 | train_loader = data['train'] 49 | val_loader = data['val'] 50 | 51 | model = model_utils.load(opt) 52 | 53 | if opt['data.cuda']: 54 | model.cuda() 55 | 56 | engine = Engine() 57 | 58 | meters = { 'train': { field: tnt.meter.AverageValueMeter() for field in opt['log.fields'] } } 59 | 60 | if val_loader is not None: 61 | meters['val'] = { field: tnt.meter.AverageValueMeter() for field in opt['log.fields'] } 62 | 63 | def on_start(state): 64 | if os.path.isfile(trace_file): 65 | os.remove(trace_file) 66 | state['scheduler'] = lr_scheduler.StepLR(state['optimizer'], opt['train.decay_every'], gamma=0.5) 67 | engine.hooks['on_start'] = on_start 68 | 69 | def on_start_epoch(state): 70 | for split, split_meters in meters.items(): 71 | for field, meter in split_meters.items(): 72 | meter.reset() 73 | state['scheduler'].step() 74 | engine.hooks['on_start_epoch'] = on_start_epoch 75 | 76 | def on_update(state): 77 | for field, meter in meters['train'].items(): 78 | meter.add(state['output'][field]) 79 | engine.hooks['on_update'] = on_update 80 | 81 | def on_end_epoch(hook_state, state): 82 | if val_loader is not None: 83 | if 'best_loss' not in hook_state: 84 | hook_state['best_loss'] = np.inf 85 | if 'wait' not in hook_state: 86 | hook_state['wait'] = 0 87 | 88 | if val_loader is not None: 89 | model_utils.evaluate(state['model'], 90 | val_loader, 91 | meters['val'], 92 | desc="Epoch {:d} valid".format(state['epoch'])) 93 | 94 | meter_vals = log_utils.extract_meter_values(meters) 95 | print("Epoch {:02d}: {:s}".format(state['epoch'], log_utils.render_meter_values(meter_vals))) 96 | meter_vals['epoch'] = state['epoch'] 97 | with open(trace_file, 'a') as f: 98 | json.dump(meter_vals, f) 99 | f.write('\n') 100 | 101 | if val_loader is not None: 102 | if meter_vals['val']['loss'] < hook_state['best_loss']: 103 | hook_state['best_loss'] = meter_vals['val']['loss'] 104 | print("==> best model (loss = {:0.6f}), saving model...".format(hook_state['best_loss'])) 105 | 106 | state['model'].cpu() 107 | torch.save(state['model'], os.path.join(opt['log.exp_dir'], 'best_model.pt')) 108 | if opt['data.cuda']: 109 | state['model'].cuda() 110 | 111 | hook_state['wait'] = 0 112 | else: 113 | hook_state['wait'] += 1 114 | 115 | if hook_state['wait'] > opt['train.patience']: 116 | print("==> patience {:d} exceeded".format(opt['train.patience'])) 117 | state['stop'] = True 118 | else: 119 | state['model'].cpu() 120 | torch.save(state['model'], os.path.join(opt['log.exp_dir'], 'best_model.pt')) 121 | if opt['data.cuda']: 122 | state['model'].cuda() 123 | 124 | engine.hooks['on_end_epoch'] = partial(on_end_epoch, { }) 125 | 126 | start = time.time() 127 | engine.train( 128 | model = model, 129 | loader = train_loader, 130 | optim_method = getattr(optim, opt['train.optim_method']), 131 | optim_config = { 'lr': opt['train.learning_rate'], 132 | 'weight_decay': opt['train.weight_decay'] }, 133 | max_epoch = opt['train.epochs'] 134 | ) 135 | end = time.time() 136 | elapsed = str(timedelta(seconds= end-start)) 137 | print("Total Time: {}".format(elapsed)) 138 | -------------------------------------------------------------------------------- /scripts/train/few_shot/trainval.py: -------------------------------------------------------------------------------- 1 | import os 2 | import json 3 | import subprocess 4 | 5 | from protonets.utils import format_opts, merge_dict 6 | from protonets.utils.log import load_trace 7 | 8 | def main(opt): 9 | result_dir = os.path.dirname(opt['model.model_path']) 10 | 11 | # get target training loss to exceed 12 | trace_file = os.path.join(result_dir, 'trace.txt') 13 | trace_vals = load_trace(trace_file) 14 | best_epoch = trace_vals['val']['loss'].argmin() 15 | 16 | # load opts 17 | model_opt_file = os.path.join(os.path.dirname(opt['model.model_path']), 'opt.json') 18 | with open(model_opt_file, 'r') as f: 19 | model_opt = json.load(f) 20 | 21 | # override previous training ops 22 | model_opt = merge_dict(model_opt, { 23 | 'log.exp_dir': os.path.join(model_opt['log.exp_dir'], 'trainval'), 24 | 'data.trainval': True, 25 | 'train.epochs': best_epoch + model_opt['train.patience'], 26 | }) 27 | 28 | subprocess.call(['python', os.path.join(os.getcwd(), 'scripts/train/few_shot/run_train.py')] + format_opts(model_opt)) 29 | -------------------------------------------------------------------------------- /setup.py: -------------------------------------------------------------------------------- 1 | from setuptools import setup 2 | 3 | setup(name='protonets', 4 | version='0.0.1', 5 | author='Jake Snell', 6 | author_email='jsnell@cs.toronto.edu', 7 | license='MIT', 8 | packages=['protonets'], 9 | install_requires=[ 10 | 'torch', 11 | 'tqdm' 12 | ]) 13 | --------------------------------------------------------------------------------