├── .gitignore
├── LICENSE
├── README.md
├── dataloaders
    ├── dataloader.py
    ├── nyu.py
    └── transforms.py
├── deploy
    ├── data
    │   ├── depth.npy
    │   ├── depth.png
    │   ├── pred.npy
    │   ├── pred.png
    │   ├── rgb.npy
    │   ├── rgb.png
    │   └── visualize.py
    └── tx2_run_tvm.py
├── imagenet
    ├── __init__.py
    └── mobilenet.py
├── img
    ├── acc_fps_cpu.png
    ├── acc_fps_gpu.png
    └── visualization.png
├── main.py
├── metrics.py
├── models.py
├── tvm_compile
    └── tuning
    │   ├── tx2-cpu.mobilenet-nnconv5.trials=1000.stop=250.log
    │   ├── tx2-cpu.mobilenet-nnconv5dw-skipadd-pruned.trials=1000.stop=250.log
    │   ├── tx2-cpu.mobilenet-nnconv5dw-skipadd.trials=1000.stop=250.log
    │   ├── tx2-cpu.mobilenet-nnconv5dw.trials=1000.stop=250.log
    │   ├── tx2-gpu.mobilenet-nnconv5.trials=2000.stop=600.log
    │   ├── tx2-gpu.mobilenet-nnconv5dw-skipadd-pruned.trials=2000.stop=600.log
    │   ├── tx2-gpu.mobilenet-nnconv5dw-skipadd.trials=2000.stop=600.log
    │   └── tx2-gpu.mobilenet-nnconv5dw.trials=2000.stop=600.log
└── utils.py


/.gitignore:
--------------------------------------------------------------------------------
1 | __pycache__/
2 | *.pyc
3 | 


--------------------------------------------------------------------------------
/LICENSE:
--------------------------------------------------------------------------------
 1 | MIT License
 2 | 
 3 | Copyright (c) 2019 Diana Wofk
 4 | 
 5 | Permission is hereby granted, free of charge, to any person obtaining a copy
 6 | of this software and associated documentation files (the "Software"), to deal
 7 | in the Software without restriction, including without limitation the rights
 8 | to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
 9 | copies of the Software, and to permit persons to whom the Software is
10 | furnished to do so, subject to the following conditions:
11 | 
12 | The above copyright notice and this permission notice shall be included in all
13 | copies or substantial portions of the Software.
14 | 
15 | THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
16 | IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
17 | FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
18 | AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
19 | LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
20 | OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
21 | SOFTWARE.
22 | 


--------------------------------------------------------------------------------
/README.md:
--------------------------------------------------------------------------------
  1 | FastDepth
  2 | ============================
  3 | 
  4 | This repo offers trained models and evaluation code for the [FastDepth](http://fastdepth.mit.edu/) project at MIT.
  5 | 
  6 | <p align="center">
  7 | 	<img src="img/visualization.png" alt="photo not available" width="50%" height="50%">
  8 | </p>
  9 | 
 10 | ## Contents
 11 | 0. [Requirements](#requirements)
 12 | 0. [Trained Models](#trained-models)
 13 | 0. [Evaluation](#evaluation)
 14 | 0. [Deployment](#deployment)
 15 | 0. [Results](#results)
 16 | 0. [Citation](#citation)
 17 | 
 18 | ## Requirements
 19 | - Install [PyTorch](https://pytorch.org/) on a machine with a CUDA GPU. Our code was developed on a system running PyTorch v0.4.1.
 20 | - Install the [HDF5](https://en.wikipedia.org/wiki/Hierarchical_Data_Format) format libraries. Files in our pre-processed datasets are in HDF5 format.
 21 |   ```bash
 22 |   sudo apt-get update
 23 |   sudo apt-get install -y libhdf5-serial-dev hdf5-tools
 24 |   pip3 install h5py matplotlib imageio scikit-image opencv-python
 25 |   ```
 26 | - Download the preprocessed [NYU Depth V2](http://cs.nyu.edu/~silberman/datasets/nyu_depth_v2.html) dataset in HDF5 format and place it under a `data` folder outside the repo directory. The NYU dataset requires 32G of storage space.
 27 | 	```bash
 28 | 	mkdir data; cd data
 29 | 	wget http://datasets.lids.mit.edu/fastdepth/data/nyudepthv2.tar.gz
 30 | 	tar -xvf nyudepthv2.tar.gz && rm -f nyudepthv2.tar.gz
 31 | 	cd ..
 32 | 	```
 33 | 
 34 | ## Trained Models ##
 35 |   The following trained models can be found at [http://datasets.lids.mit.edu/fastdepth/results/](http://datasets.lids.mit.edu/fastdepth/results/).
 36 |   - MobileNet-NNConv5
 37 |   - MobileNet-NNConv5(depthwise)
 38 |   - MobileNet-NNConv5(depthwise), with additive skip connections
 39 |   - **MobileNet-NNConv5(depthwise), with additive skip connections, pruned**
 40 | 
 41 | Our final model is `mobilenet-nnconv5-skipadd-pruned`, i.e. a MobileNet-NNConv5 architecture with depthwise separable layers in the decoder, with additive skip connections between the encoder and decoder, and after network pruning using [NetAdapt](https://github.com/denru01/netadapt/). The other models are offered to provide insight into our approach.
 42 | 
 43 | When downloading, save models to a `results` folder outside the repo directory:
 44 |   ```bash
 45 |   mkdir results; cd results
 46 |   wget -r -np -nH --cut-dirs=2 --reject "index.html*" http://datasets.lids.mit.edu/fastdepth/results/
 47 |   cd ..
 48 |   ```
 49 | ### Pretrained MobileNet ###
 50 | 
 51 | The model file for the pretrained MobileNet used in our model definition can be downloaded from [http://datasets.lids.mit.edu/fastdepth/imagenet/](http://datasets.lids.mit.edu/fastdepth/imagenet/).
 52 | 
 53 | ## Evaluation ##
 54 | 
 55 | This step requires a valid PyTorch installation and a saved copy of the NYU Depth v2 dataset. It is meant to be performed on a host machine with a CUDA GPU, not on an embedded platform. Deployment on an embedded device is discussed in the [next section](#deployment).
 56 | 
 57 | To evaluate a model, navigate to the repo directory and run:
 58 | 
 59 | ```bash
 60 | python3 main.py --evaluate [path_to_trained_model]
 61 | ```
 62 | 
 63 | The evaluation code will report model accuracy in terms of the delta1 metric as well as RMSE in millimeters.
 64 | 
 65 | Note: This evaluation code was sourced and modified from [here](https://github.com/fangchangma/sparse-to-dense.pytorch).
 66 | 
 67 | ## Deployment ##
 68 | 
 69 | We use the [TVM compiler stack](https://tvm.ai/) to compile trained models for **deployment on an NVIDIA Jetson TX2**. Models are cross-compiled on a host machine and then deployed on the TX2. The `tvm-compile/tuning` folder in this repo contains the results of our [auto-tuning](https://docs.tvm.ai/tutorials/index.html#auto-tuning) the layers within our models for both the TX2 GPU and CPU. These can be used during the compilation process to achieve low model runtimes on the TX2. Outputs of TVM compilation for our trained models can be found at [http://datasets.lids.mit.edu/fastdepth/results/tvm_compiled/](http://datasets.lids.mit.edu/fastdepth/results/tvm_compiled/).
 70 | 
 71 | On the TX2, download the trained models as explained above in the section [Trained Models](#trained-models). The compiled model files should be located in `results/tvm_compiled`.
 72 | 
 73 | ### Installing the TVM Runtime ####
 74 | 
 75 | Deployment requires building the TVM runtime code on the target embedded device (that will be used solely for running a trained and compiled model). The following instructions are taken from [this TVM tutorial](https://docs.tvm.ai/tutorials/cross_compilation_and_rpc.html#build-tvm-runtime-on-device) and have been tested on a **TX2 with CUDA-8.0 and LLVM-4.0 installed**.
 76 | 
 77 | First, clone the TVM repo and modify config file:
 78 | ```bash
 79 | git clone --recursive https://github.com/dmlc/tvm
 80 | cd tvm
 81 | git reset --hard ab4946c8b80da510a5a518dca066d8159473345f
 82 | git submodule update --init
 83 | cp cmake/config.cmake .
 84 | ```
 85 | Make the following edits to the `config.cmake` file:
 86 | ```cmake
 87 | set(USE_CUDA OFF) -> set(USE_CUDA [path_to_cuda]) # e.g. /usr/local/cuda-8.0/
 88 | set(USE_LLVM OFF) -> set(USE_LLVM [path_to_llvm-config]) # e.g. /usr/lib/llvm-4.0/bin/llvm-config
 89 | ```
 90 | 
 91 | Then build the runtime:
 92 | ```bash
 93 | make runtime -j2
 94 | ```
 95 | Finally, update the `PYTHONPATH` environment variable:
 96 | ```bash
 97 | export PYTHONPATH=$PYTHONPATH:~/tvm/python
 98 | ```
 99 | ### Running a Compiled Model ####
100 | 
101 | To run a compiled model on the device, navigate to the `deploy` folder and run:
102 | 
103 | ```bash
104 | python3 tx2_run_tvm.py --input-fp [path_to_input_npy_file] --output-fp [path_to_output_npy_file] --model-dir [path_to_folder_with_tvm_compiled_model_files]
105 | ```
106 | 
107 | Note that when running a model compiled for the GPU, a `cuda` argument must be specified. For instance:
108 | 
109 | ```bash
110 | python3 tx2_run_tvm.py --input-fp data/rgb.npy --output-fp data/pred.npy --model-dir ../../results/tvm_compiled/tx2_cpu_mobilenet_nnconv5dw_skipadd_pruned/
111 | python3 tx2_run_tvm.py --input-fp data/rgb.npy --output-fp data/pred.npy --model-dir ../../results/tvm_compiled/tx2_gpu_mobilenet_nnconv5dw_skipadd_pruned/ --cuda True
112 | ```
113 | 
114 | Example RGB input, ground truth, and model prediction data (as numpy arrays) is provided in the `data` folder. To convert the `.npy` files into `.png` format, navigate into `data` and run `python3 visualize.py`.
115 | 
116 | ### Measuring Power Consumption ###
117 | 
118 | On the TX2, power consumption on the main VDD_IN rail can be measured by running the following command:
119 | 
120 | ```bash
121 | cat /sys/devices/3160000.i2c/i2c-0/0-0041/iio_device/in_power0_input
122 | ```
123 | 
124 | ## Results
125 | 
126 | Comparison against prior work. Runtimes were measured on an NVIDIA Jetson TX2 in max-N mode.
127 | 
128 | | on NYU Depth v2     |  Input Size  |  MACs [G]  | RMSE [m] | delta1 | CPU [ms] | GPU [ms] |
129 | |---------------------------------------------|:-----:|:-----:|:-----:|:-----:|:-----:|:--:|
130 | | [Eigen et al. [NIPS 2014]](https://papers.nips.cc/paper/5539-depth-map-prediction-from-a-single-image-using-a-multi-scale-deep-network.pdf)                | 228×304 | 2.06 | 0.907 | 0.611 | 300 | 23 |
131 | | [Eigen et al. [ICCV 2015]](https://www.cv-foundation.org/openaccess/content_iccv_2015/papers/Eigen_Predicting_Depth_Surface_ICCV_2015_paper.pdf) (with AlexNet) | 228×304 | 8.39 | 0.753 | 0.697 | 1400 | 96 |
132 | | [Eigen et al. [ICCV 2015]](https://www.cv-foundation.org/openaccess/content_iccv_2015/papers/Eigen_Predicting_Depth_Surface_ICCV_2015_paper.pdf) (with VGG)     | 228×304 | 23.4 | 0.641 | 0.769 | 2800 | 195 |
133 | | [Laina et al. [3DV 2016]](https://arxiv.org/pdf/1606.00373.pdf) (with UpConv)   | 228×304 | 22.9 | 0.604 | 0.789 | 2400 | 237 |
134 | | [Laina et al. [3DV 2016]](https://arxiv.org/pdf/1606.00373.pdf) (with UpProj)   | 228×304 | 42.7 | **0.573** | **0.811** | 3900 | 319 |
135 | | [Xian et al. [CVPR 2018]](http://openaccess.thecvf.com/content_cvpr_2018/papers/Xian_Monocular_Relative_Depth_CVPR_2018_paper.pdf) (with UpProj)   | 384×384 | 61.8 | 0.660 | 0.781 | 4400 | 283 |
136 | | This Work                                   | 224×224 | **0.37** | 0.604 | 0.771 | **37** | **5.6** |
137 | 
138 | "This Work" refers to MobileNet-NNConv5(depthwise), with additive skip connections, pruned.
139 | 
140 | <p float="left">
141 | <img src="img/acc_fps_gpu.png" alt="photo not available" width="375">
142 | <img src="img/acc_fps_cpu.png" alt="photo not available" width="375">
143 | </p>
144 | 
145 | ## Citation
146 | If you reference our work, please consider citing the following:
147 | 
148 | 	@inproceedings{icra_2019_fastdepth,
149 | 		author      = {{Wofk, Diana and Ma, Fangchang and Yang, Tien-Ju and Karaman, Sertac and Sze, Vivienne}},
150 | 		title       = {{FastDepth: Fast Monocular Depth Estimation on Embedded Systems}},
151 | 		booktitle   = {{IEEE International Conference on Robotics and Automation (ICRA)}},
152 | 		year        = {{2019}}
153 | 	}
154 | 


--------------------------------------------------------------------------------
/dataloaders/dataloader.py:
--------------------------------------------------------------------------------
  1 | import os
  2 | import os.path
  3 | import numpy as np
  4 | import torch.utils.data as data
  5 | import h5py
  6 | import dataloaders.transforms as transforms
  7 | 
  8 | def h5_loader(path):
  9 |     h5f = h5py.File(path, "r")
 10 |     rgb = np.array(h5f['rgb'])
 11 |     rgb = np.transpose(rgb, (1, 2, 0))
 12 |     depth = np.array(h5f['depth'])
 13 |     return rgb, depth
 14 | 
 15 | # def rgb2grayscale(rgb):
 16 | #     return rgb[:,:,0] * 0.2989 + rgb[:,:,1] * 0.587 + rgb[:,:,2] * 0.114
 17 | 
 18 | class MyDataloader(data.Dataset):
 19 |     modality_names = ['rgb']
 20 | 
 21 |     def is_image_file(self, filename):
 22 |         IMG_EXTENSIONS = ['.h5']
 23 |         return any(filename.endswith(extension) for extension in IMG_EXTENSIONS)
 24 | 
 25 |     def find_classes(self, dir):
 26 |         classes = [d for d in os.listdir(dir) if os.path.isdir(os.path.join(dir, d))]
 27 |         classes.sort()
 28 |         class_to_idx = {classes[i]: i for i in range(len(classes))}
 29 |         return classes, class_to_idx
 30 | 
 31 |     def make_dataset(self, dir, class_to_idx):
 32 |         images = []
 33 |         dir = os.path.expanduser(dir)
 34 |         for target in sorted(os.listdir(dir)):
 35 |             d = os.path.join(dir, target)
 36 |             if not os.path.isdir(d):
 37 |                 continue
 38 |             for root, _, fnames in sorted(os.walk(d)):
 39 |                 for fname in sorted(fnames):
 40 |                     if self.is_image_file(fname):
 41 |                         path = os.path.join(root, fname)
 42 |                         item = (path, class_to_idx[target])
 43 |                         images.append(item)
 44 |         return images
 45 | 
 46 |     color_jitter = transforms.ColorJitter(0.4, 0.4, 0.4)
 47 | 
 48 |     def __init__(self, root, split, modality='rgb', loader=h5_loader):
 49 |         classes, class_to_idx = self.find_classes(root)
 50 |         imgs = self.make_dataset(root, class_to_idx)
 51 |         assert len(imgs)>0, "Found 0 images in subfolders of: " + root + "\n"
 52 |         # print("Found {} images in {} folder.".format(len(imgs), split))
 53 |         self.root = root
 54 |         self.imgs = imgs
 55 |         self.classes = classes
 56 |         self.class_to_idx = class_to_idx
 57 |         if split == 'train':
 58 |             self.transform = self.train_transform
 59 |         elif split == 'holdout':
 60 |             self.transform = self.val_transform
 61 |         elif split == 'val':
 62 |             self.transform = self.val_transform
 63 |         else:
 64 |             raise (RuntimeError("Invalid dataset split: " + split + "\n"
 65 |                                 "Supported dataset splits are: train, val"))
 66 |         self.loader = loader
 67 | 
 68 |         assert (modality in self.modality_names), "Invalid modality split: " + modality + "\n" + \
 69 |                                 "Supported dataset splits are: " + ''.join(self.modality_names)
 70 |         self.modality = modality
 71 | 
 72 |     def train_transform(self, rgb, depth):
 73 |         raise (RuntimeError("train_transform() is not implemented. "))
 74 | 
 75 |     def val_transform(rgb, depth):
 76 |         raise (RuntimeError("val_transform() is not implemented."))
 77 | 
 78 |     def __getraw__(self, index):
 79 |         """
 80 |         Args:
 81 |             index (int): Index
 82 | 
 83 |         Returns:
 84 |             tuple: (rgb, depth) the raw data.
 85 |         """
 86 |         path, target = self.imgs[index]
 87 |         rgb, depth = self.loader(path)
 88 |         return rgb, depth
 89 | 
 90 |     def __getitem__(self, index):
 91 |         rgb, depth = self.__getraw__(index)
 92 |         if self.transform is not None:
 93 |             rgb_np, depth_np = self.transform(rgb, depth)
 94 |         else:
 95 |             raise(RuntimeError("transform not defined"))
 96 | 
 97 |         # color normalization
 98 |         # rgb_tensor = normalize_rgb(rgb_tensor)
 99 |         # rgb_np = normalize_np(rgb_np)
100 | 
101 |         if self.modality == 'rgb':
102 |             input_np = rgb_np
103 | 
104 |         to_tensor = transforms.ToTensor()
105 |         input_tensor = to_tensor(input_np)
106 |         while input_tensor.dim() < 3:
107 |             input_tensor = input_tensor.unsqueeze(0)
108 |         depth_tensor = to_tensor(depth_np)
109 |         depth_tensor = depth_tensor.unsqueeze(0)
110 | 
111 |         return input_tensor, depth_tensor
112 | 
113 |     def __len__(self):
114 |         return len(self.imgs)
115 | 


--------------------------------------------------------------------------------
/dataloaders/nyu.py:
--------------------------------------------------------------------------------
 1 | import numpy as np
 2 | import dataloaders.transforms as transforms
 3 | from dataloaders.dataloader import MyDataloader
 4 | 
 5 | iheight, iwidth = 480, 640 # raw image size
 6 | 
 7 | class NYUDataset(MyDataloader):
 8 |     def __init__(self, root, split, modality='rgb'):
 9 |         self.split = split
10 |         super(NYUDataset, self).__init__(root, split, modality)
11 |         self.output_size = (224, 224)
12 | 
13 |     def is_image_file(self, filename):
14 |         # IMG_EXTENSIONS = ['.h5']
15 |         if self.split == 'train':
16 |             return (filename.endswith('.h5') and \
17 |                 '00001.h5' not in filename and '00201.h5' not in filename)
18 |         elif self.split == 'holdout':
19 |             return ('00001.h5' in filename or '00201.h5' in filename)
20 |         elif self.split == 'val':
21 |             return (filename.endswith('.h5'))
22 |         else:
23 |             raise (RuntimeError("Invalid dataset split: " + split + "\n"
24 |                                 "Supported dataset splits are: train, val"))
25 | 
26 |     def train_transform(self, rgb, depth):
27 |         s = np.random.uniform(1.0, 1.5) # random scaling
28 |         depth_np = depth / s
29 |         angle = np.random.uniform(-5.0, 5.0) # random rotation degrees
30 |         do_flip = np.random.uniform(0.0, 1.0) < 0.5 # random horizontal flip
31 | 
32 |         # perform 1st step of data augmentation
33 |         transform = transforms.Compose([
34 |             transforms.Resize(250.0 / iheight), # this is for computational efficiency, since rotation can be slow
35 |             transforms.Rotate(angle),
36 |             transforms.Resize(s),
37 |             transforms.CenterCrop((228, 304)),
38 |             transforms.HorizontalFlip(do_flip),
39 |             transforms.Resize(self.output_size),
40 |         ])
41 |         rgb_np = transform(rgb)
42 |         rgb_np = self.color_jitter(rgb_np) # random color jittering
43 |         rgb_np = np.asfarray(rgb_np, dtype='float') / 255
44 |         depth_np = transform(depth_np)
45 | 
46 |         return rgb_np, depth_np
47 | 
48 |     def val_transform(self, rgb, depth):
49 |         depth_np = depth
50 |         transform = transforms.Compose([
51 |             transforms.Resize(250.0 / iheight),
52 |             transforms.CenterCrop((228, 304)),
53 |             transforms.Resize(self.output_size),
54 |         ])
55 |         rgb_np = transform(rgb)
56 |         rgb_np = np.asfarray(rgb_np, dtype='float') / 255
57 |         depth_np = transform(depth_np)
58 | 
59 |         return rgb_np, depth_np
60 | 


--------------------------------------------------------------------------------
/dataloaders/transforms.py:
--------------------------------------------------------------------------------
  1 | from __future__ import division
  2 | import torch
  3 | import math
  4 | import random
  5 | 
  6 | from PIL import Image, ImageOps, ImageEnhance
  7 | try:
  8 |     import accimage
  9 | except ImportError:
 10 |     accimage = None
 11 | 
 12 | import numpy as np
 13 | import numbers
 14 | import types
 15 | import collections
 16 | import warnings
 17 | 
 18 | import scipy.ndimage.interpolation as itpl
 19 | import scipy.misc as misc
 20 | 
 21 | 
 22 | def _is_numpy_image(img):
 23 |     return isinstance(img, np.ndarray) and (img.ndim in {2, 3})
 24 | 
 25 | def _is_pil_image(img):
 26 |     if accimage is not None:
 27 |         return isinstance(img, (Image.Image, accimage.Image))
 28 |     else:
 29 |         return isinstance(img, Image.Image)
 30 | 
 31 | def _is_tensor_image(img):
 32 |     return torch.is_tensor(img) and img.ndimension() == 3
 33 | 
 34 | def adjust_brightness(img, brightness_factor):
 35 |     """Adjust brightness of an Image.
 36 | 
 37 |     Args:
 38 |         img (PIL Image): PIL Image to be adjusted.
 39 |         brightness_factor (float):  How much to adjust the brightness. Can be
 40 |             any non negative number. 0 gives a black image, 1 gives the
 41 |             original image while 2 increases the brightness by a factor of 2.
 42 | 
 43 |     Returns:
 44 |         PIL Image: Brightness adjusted image.
 45 |     """
 46 |     if not _is_pil_image(img):
 47 |         raise TypeError('img should be PIL Image. Got {}'.format(type(img)))
 48 | 
 49 |     enhancer = ImageEnhance.Brightness(img)
 50 |     img = enhancer.enhance(brightness_factor)
 51 |     return img
 52 | 
 53 | 
 54 | def adjust_contrast(img, contrast_factor):
 55 |     """Adjust contrast of an Image.
 56 | 
 57 |     Args:
 58 |         img (PIL Image): PIL Image to be adjusted.
 59 |         contrast_factor (float): How much to adjust the contrast. Can be any
 60 |             non negative number. 0 gives a solid gray image, 1 gives the
 61 |             original image while 2 increases the contrast by a factor of 2.
 62 | 
 63 |     Returns:
 64 |         PIL Image: Contrast adjusted image.
 65 |     """
 66 |     if not _is_pil_image(img):
 67 |         raise TypeError('img should be PIL Image. Got {}'.format(type(img)))
 68 | 
 69 |     enhancer = ImageEnhance.Contrast(img)
 70 |     img = enhancer.enhance(contrast_factor)
 71 |     return img
 72 | 
 73 | 
 74 | def adjust_saturation(img, saturation_factor):
 75 |     """Adjust color saturation of an image.
 76 | 
 77 |     Args:
 78 |         img (PIL Image): PIL Image to be adjusted.
 79 |         saturation_factor (float):  How much to adjust the saturation. 0 will
 80 |             give a black and white image, 1 will give the original image while
 81 |             2 will enhance the saturation by a factor of 2.
 82 | 
 83 |     Returns:
 84 |         PIL Image: Saturation adjusted image.
 85 |     """
 86 |     if not _is_pil_image(img):
 87 |         raise TypeError('img should be PIL Image. Got {}'.format(type(img)))
 88 | 
 89 |     enhancer = ImageEnhance.Color(img)
 90 |     img = enhancer.enhance(saturation_factor)
 91 |     return img
 92 | 
 93 | 
 94 | def adjust_hue(img, hue_factor):
 95 |     """Adjust hue of an image.
 96 | 
 97 |     The image hue is adjusted by converting the image to HSV and
 98 |     cyclically shifting the intensities in the hue channel (H).
 99 |     The image is then converted back to original image mode.
100 | 
101 |     `hue_factor` is the amount of shift in H channel and must be in the
102 |     interval `[-0.5, 0.5]`.
103 | 
104 |     See https://en.wikipedia.org/wiki/Hue for more details on Hue.
105 | 
106 |     Args:
107 |         img (PIL Image): PIL Image to be adjusted.
108 |         hue_factor (float):  How much to shift the hue channel. Should be in
109 |             [-0.5, 0.5]. 0.5 and -0.5 give complete reversal of hue channel in
110 |             HSV space in positive and negative direction respectively.
111 |             0 means no shift. Therefore, both -0.5 and 0.5 will give an image
112 |             with complementary colors while 0 gives the original image.
113 | 
114 |     Returns:
115 |         PIL Image: Hue adjusted image.
116 |     """
117 |     if not(-0.5 <= hue_factor <= 0.5):
118 |         raise ValueError('hue_factor is not in [-0.5, 0.5].'.format(hue_factor))
119 | 
120 |     if not _is_pil_image(img):
121 |         raise TypeError('img should be PIL Image. Got {}'.format(type(img)))
122 | 
123 |     input_mode = img.mode
124 |     if input_mode in {'L', '1', 'I', 'F'}:
125 |         return img
126 | 
127 |     h, s, v = img.convert('HSV').split()
128 | 
129 |     np_h = np.array(h, dtype=np.uint8)
130 |     # uint8 addition take cares of rotation across boundaries
131 |     with np.errstate(over='ignore'):
132 |         np_h += np.uint8(hue_factor * 255)
133 |     h = Image.fromarray(np_h, 'L')
134 | 
135 |     img = Image.merge('HSV', (h, s, v)).convert(input_mode)
136 |     return img
137 | 
138 | 
139 | def adjust_gamma(img, gamma, gain=1):
140 |     """Perform gamma correction on an image.
141 | 
142 |     Also known as Power Law Transform. Intensities in RGB mode are adjusted
143 |     based on the following equation:
144 | 
145 |         I_out = 255 * gain * ((I_in / 255) ** gamma)
146 | 
147 |     See https://en.wikipedia.org/wiki/Gamma_correction for more details.
148 | 
149 |     Args:
150 |         img (PIL Image): PIL Image to be adjusted.
151 |         gamma (float): Non negative real number. gamma larger than 1 make the
152 |             shadows darker, while gamma smaller than 1 make dark regions
153 |             lighter.
154 |         gain (float): The constant multiplier.
155 |     """
156 |     if not _is_pil_image(img):
157 |         raise TypeError('img should be PIL Image. Got {}'.format(type(img)))
158 | 
159 |     if gamma < 0:
160 |         raise ValueError('Gamma should be a non-negative real number')
161 | 
162 |     input_mode = img.mode
163 |     img = img.convert('RGB')
164 | 
165 |     np_img = np.array(img, dtype=np.float32)
166 |     np_img = 255 * gain * ((np_img / 255) ** gamma)
167 |     np_img = np.uint8(np.clip(np_img, 0, 255))
168 | 
169 |     img = Image.fromarray(np_img, 'RGB').convert(input_mode)
170 |     return img
171 | 
172 | 
173 | class Compose(object):
174 |     """Composes several transforms together.
175 | 
176 |     Args:
177 |         transforms (list of ``Transform`` objects): list of transforms to compose.
178 | 
179 |     Example:
180 |         >>> transforms.Compose([
181 |         >>>     transforms.CenterCrop(10),
182 |         >>>     transforms.ToTensor(),
183 |         >>> ])
184 |     """
185 | 
186 |     def __init__(self, transforms):
187 |         self.transforms = transforms
188 | 
189 |     def __call__(self, img):
190 |         for t in self.transforms:
191 |             img = t(img)
192 |         return img
193 | 
194 | 
195 | class ToTensor(object):
196 |     """Convert a ``numpy.ndarray`` to tensor.
197 | 
198 |     Converts a numpy.ndarray (H x W x C) to a torch.FloatTensor of shape (C x H x W).
199 |     """
200 | 
201 |     def __call__(self, img):
202 |         """Convert a ``numpy.ndarray`` to tensor.
203 | 
204 |         Args:
205 |             img (numpy.ndarray): Image to be converted to tensor.
206 | 
207 |         Returns:
208 |             Tensor: Converted image.
209 |         """
210 |         if not(_is_numpy_image(img)):
211 |             raise TypeError('img should be ndarray. Got {}'.format(type(img)))
212 | 
213 |         if isinstance(img, np.ndarray):
214 |             # handle numpy array
215 |             if img.ndim == 3:
216 |                 img = torch.from_numpy(img.transpose((2, 0, 1)).copy())
217 |             elif img.ndim == 2:
218 |                 img = torch.from_numpy(img.copy())
219 |             else:
220 |                 raise RuntimeError('img should be ndarray with 2 or 3 dimensions. Got {}'.format(img.ndim))
221 | 
222 |             # backward compatibility
223 |             # return img.float().div(255)
224 |             return img.float()
225 | 
226 | 
227 | class NormalizeNumpyArray(object):
228 |     """Normalize a ``numpy.ndarray`` with mean and standard deviation.
229 |     Given mean: ``(M1,...,Mn)`` and std: ``(M1,..,Mn)`` for ``n`` channels, this transform
230 |     will normalize each channel of the input ``numpy.ndarray`` i.e.
231 |     ``input[channel] = (input[channel] - mean[channel]) / std[channel]``
232 | 
233 |     Args:
234 |         mean (sequence): Sequence of means for each channel.
235 |         std (sequence): Sequence of standard deviations for each channel.
236 |     """
237 | 
238 |     def __init__(self, mean, std):
239 |         self.mean = mean
240 |         self.std = std
241 | 
242 |     def __call__(self, img):
243 |         """
244 |         Args:
245 |             img (numpy.ndarray): Image of size (H, W, C) to be normalized.
246 | 
247 |         Returns:
248 |             Tensor: Normalized image.
249 |         """
250 |         if not(_is_numpy_image(img)):
251 |             raise TypeError('img should be ndarray. Got {}'.format(type(img)))
252 |         # TODO: make efficient
253 |         print(img.shape)
254 |         for i in range(3):
255 |             img[:,:,i] = (img[:,:,i] - self.mean[i]) / self.std[i]
256 |         return img
257 | 
258 | class NormalizeTensor(object):
259 |     """Normalize an tensor image with mean and standard deviation.
260 |     Given mean: ``(M1,...,Mn)`` and std: ``(M1,..,Mn)`` for ``n`` channels, this transform
261 |     will normalize each channel of the input ``torch.*Tensor`` i.e.
262 |     ``input[channel] = (input[channel] - mean[channel]) / std[channel]``
263 | 
264 |     Args:
265 |         mean (sequence): Sequence of means for each channel.
266 |         std (sequence): Sequence of standard deviations for each channel.
267 |     """
268 | 
269 |     def __init__(self, mean, std):
270 |         self.mean = mean
271 |         self.std = std
272 | 
273 |     def __call__(self, tensor):
274 |         """
275 |         Args:
276 |             tensor (Tensor): Tensor image of size (C, H, W) to be normalized.
277 | 
278 |         Returns:
279 |             Tensor: Normalized Tensor image.
280 |         """
281 |         if not _is_tensor_image(tensor):
282 |             raise TypeError('tensor is not a torch image.')
283 |         # TODO: make efficient
284 |         for t, m, s in zip(tensor, self.mean, self.std):
285 |             t.sub_(m).div_(s)
286 |         return tensor
287 | 
288 | class Rotate(object):
289 |     """Rotates the given ``numpy.ndarray``.
290 | 
291 |     Args:
292 |         angle (float): The rotation angle in degrees.
293 |     """
294 | 
295 |     def __init__(self, angle):
296 |         self.angle = angle
297 | 
298 |     def __call__(self, img):
299 |         """
300 |         Args:
301 |             img (numpy.ndarray (C x H x W)): Image to be rotated.
302 | 
303 |         Returns:
304 |             img (numpy.ndarray (C x H x W)): Rotated image.
305 |         """
306 | 
307 |         # order=0 means nearest-neighbor type interpolation
308 |         return itpl.rotate(img, self.angle, reshape=False, prefilter=False, order=0)
309 | 
310 | 
311 | class Resize(object):
312 |     """Resize the the given ``numpy.ndarray`` to the given size.
313 |     Args:
314 |         size (sequence or int): Desired output size. If size is a sequence like
315 |             (h, w), output size will be matched to this. If size is an int,
316 |             smaller edge of the image will be matched to this number.
317 |             i.e, if height > width, then image will be rescaled to
318 |             (size * height / width, size)
319 |         interpolation (int, optional): Desired interpolation. Default is
320 |             ``PIL.Image.BILINEAR``
321 |     """
322 | 
323 |     def __init__(self, size, interpolation='nearest'):
324 |         assert isinstance(size, int) or isinstance(size, float) or \
325 |                (isinstance(size, collections.Iterable) and len(size) == 2)
326 |         self.size = size
327 |         self.interpolation = interpolation
328 | 
329 |     def __call__(self, img):
330 |         """
331 |         Args:
332 |             img (PIL Image): Image to be scaled.
333 |         Returns:
334 |             PIL Image: Rescaled image.
335 |         """
336 |         if img.ndim == 3:
337 |             return misc.imresize(img, self.size, self.interpolation)
338 |         elif img.ndim == 2:
339 |             return misc.imresize(img, self.size, self.interpolation, 'F')
340 |         else:
341 |             RuntimeError('img should be ndarray with 2 or 3 dimensions. Got {}'.format(img.ndim))
342 | 
343 | 
344 | class CenterCrop(object):
345 |     """Crops the given ``numpy.ndarray`` at the center.
346 | 
347 |     Args:
348 |         size (sequence or int): Desired output size of the crop. If size is an
349 |             int instead of sequence like (h, w), a square crop (size, size) is
350 |             made.
351 |     """
352 | 
353 |     def __init__(self, size):
354 |         if isinstance(size, numbers.Number):
355 |             self.size = (int(size), int(size))
356 |         else:
357 |             self.size = size
358 | 
359 |     @staticmethod
360 |     def get_params(img, output_size):
361 |         """Get parameters for ``crop`` for center crop.
362 | 
363 |         Args:
364 |             img (numpy.ndarray (C x H x W)): Image to be cropped.
365 |             output_size (tuple): Expected output size of the crop.
366 | 
367 |         Returns:
368 |             tuple: params (i, j, h, w) to be passed to ``crop`` for center crop.
369 |         """
370 |         h = img.shape[0]
371 |         w = img.shape[1]
372 |         th, tw = output_size
373 |         i = int(round((h - th) / 2.))
374 |         j = int(round((w - tw) / 2.))
375 | 
376 |         # # randomized cropping
377 |         # i = np.random.randint(i-3, i+4)
378 |         # j = np.random.randint(j-3, j+4)
379 | 
380 |         return i, j, th, tw
381 | 
382 |     def __call__(self, img):
383 |         """
384 |         Args:
385 |             img (numpy.ndarray (C x H x W)): Image to be cropped.
386 | 
387 |         Returns:
388 |             img (numpy.ndarray (C x H x W)): Cropped image.
389 |         """
390 |         i, j, h, w = self.get_params(img, self.size)
391 | 
392 |         """
393 |         i: Upper pixel coordinate.
394 |         j: Left pixel coordinate.
395 |         h: Height of the cropped image.
396 |         w: Width of the cropped image.
397 |         """
398 |         if not(_is_numpy_image(img)):
399 |             raise TypeError('img should be ndarray. Got {}'.format(type(img)))
400 |         if img.ndim == 3:
401 |             return img[i:i+h, j:j+w, :]
402 |         elif img.ndim == 2:
403 |             return img[i:i + h, j:j + w]
404 |         else:
405 |             raise RuntimeError('img should be ndarray with 2 or 3 dimensions. Got {}'.format(img.ndim))
406 | 
407 | class BottomCrop(object):
408 |     """Crops the given ``numpy.ndarray`` at the bottom.
409 | 
410 |     Args:
411 |         size (sequence or int): Desired output size of the crop. If size is an
412 |             int instead of sequence like (h, w), a square crop (size, size) is
413 |             made.
414 |     """
415 | 
416 |     def __init__(self, size):
417 |         if isinstance(size, numbers.Number):
418 |             self.size = (int(size), int(size))
419 |         else:
420 |             self.size = size
421 | 
422 |     @staticmethod
423 |     def get_params(img, output_size):
424 |         """Get parameters for ``crop`` for bottom crop.
425 | 
426 |         Args:
427 |             img (numpy.ndarray (C x H x W)): Image to be cropped.
428 |             output_size (tuple): Expected output size of the crop.
429 | 
430 |         Returns:
431 |             tuple: params (i, j, h, w) to be passed to ``crop`` for bottom crop.
432 |         """
433 |         h = img.shape[0]
434 |         w = img.shape[1]
435 |         th, tw = output_size
436 |         i = h - th
437 |         j = int(round((w - tw) / 2.))
438 | 
439 |         # randomized left and right cropping
440 |         # i = np.random.randint(i-3, i+4)
441 |         # j = np.random.randint(j-1, j+1)
442 | 
443 |         return i, j, th, tw
444 | 
445 |     def __call__(self, img):
446 |         """
447 |         Args:
448 |             img (numpy.ndarray (C x H x W)): Image to be cropped.
449 | 
450 |         Returns:
451 |             img (numpy.ndarray (C x H x W)): Cropped image.
452 |         """
453 |         i, j, h, w = self.get_params(img, self.size)
454 | 
455 |         """
456 |         i: Upper pixel coordinate.
457 |         j: Left pixel coordinate.
458 |         h: Height of the cropped image.
459 |         w: Width of the cropped image.
460 |         """
461 |         if not(_is_numpy_image(img)):
462 |             raise TypeError('img should be ndarray. Got {}'.format(type(img)))
463 |         if img.ndim == 3:
464 |             return img[i:i+h, j:j+w, :]
465 |         elif img.ndim == 2:
466 |             return img[i:i + h, j:j + w]
467 |         else:
468 |             raise RuntimeError('img should be ndarray with 2 or 3 dimensions. Got {}'.format(img.ndim))
469 | 
470 | class Lambda(object):
471 |     """Apply a user-defined lambda as a transform.
472 | 
473 |     Args:
474 |         lambd (function): Lambda/function to be used for transform.
475 |     """
476 | 
477 |     def __init__(self, lambd):
478 |         assert isinstance(lambd, types.LambdaType)
479 |         self.lambd = lambd
480 | 
481 |     def __call__(self, img):
482 |         return self.lambd(img)
483 | 
484 | 
485 | class HorizontalFlip(object):
486 |     """Horizontally flip the given ``numpy.ndarray``.
487 | 
488 |     Args:
489 |         do_flip (boolean): whether or not do horizontal flip.
490 | 
491 |     """
492 | 
493 |     def __init__(self, do_flip):
494 |         self.do_flip = do_flip
495 | 
496 |     def __call__(self, img):
497 |         """
498 |         Args:
499 |             img (numpy.ndarray (C x H x W)): Image to be flipped.
500 | 
501 |         Returns:
502 |             img (numpy.ndarray (C x H x W)): flipped image.
503 |         """
504 |         if not(_is_numpy_image(img)):
505 |             raise TypeError('img should be ndarray. Got {}'.format(type(img)))
506 | 
507 |         if self.do_flip:
508 |             return np.fliplr(img)
509 |         else:
510 |             return img
511 | 
512 | 
513 | class ColorJitter(object):
514 |     """Randomly change the brightness, contrast and saturation of an image.
515 | 
516 |     Args:
517 |         brightness (float): How much to jitter brightness. brightness_factor
518 |             is chosen uniformly from [max(0, 1 - brightness), 1 + brightness].
519 |         contrast (float): How much to jitter contrast. contrast_factor
520 |             is chosen uniformly from [max(0, 1 - contrast), 1 + contrast].
521 |         saturation (float): How much to jitter saturation. saturation_factor
522 |             is chosen uniformly from [max(0, 1 - saturation), 1 + saturation].
523 |         hue(float): How much to jitter hue. hue_factor is chosen uniformly from
524 |             [-hue, hue]. Should be >=0 and <= 0.5.
525 |     """
526 |     def __init__(self, brightness=0, contrast=0, saturation=0, hue=0):
527 |         self.brightness = brightness
528 |         self.contrast = contrast
529 |         self.saturation = saturation
530 |         self.hue = hue
531 | 
532 |     @staticmethod
533 |     def get_params(brightness, contrast, saturation, hue):
534 |         """Get a randomized transform to be applied on image.
535 | 
536 |         Arguments are same as that of __init__.
537 | 
538 |         Returns:
539 |             Transform which randomly adjusts brightness, contrast and
540 |             saturation in a random order.
541 |         """
542 |         transforms = []
543 |         if brightness > 0:
544 |             brightness_factor = np.random.uniform(max(0, 1 - brightness), 1 + brightness)
545 |             transforms.append(Lambda(lambda img: adjust_brightness(img, brightness_factor)))
546 | 
547 |         if contrast > 0:
548 |             contrast_factor = np.random.uniform(max(0, 1 - contrast), 1 + contrast)
549 |             transforms.append(Lambda(lambda img: adjust_contrast(img, contrast_factor)))
550 | 
551 |         if saturation > 0:
552 |             saturation_factor = np.random.uniform(max(0, 1 - saturation), 1 + saturation)
553 |             transforms.append(Lambda(lambda img: adjust_saturation(img, saturation_factor)))
554 | 
555 |         if hue > 0:
556 |             hue_factor = np.random.uniform(-hue, hue)
557 |             transforms.append(Lambda(lambda img: adjust_hue(img, hue_factor)))
558 | 
559 |         np.random.shuffle(transforms)
560 |         transform = Compose(transforms)
561 | 
562 |         return transform
563 | 
564 |     def __call__(self, img):
565 |         """
566 |         Args:
567 |             img (numpy.ndarray (C x H x W)): Input image.
568 | 
569 |         Returns:
570 |             img (numpy.ndarray (C x H x W)): Color jittered image.
571 |         """
572 |         if not(_is_numpy_image(img)):
573 |             raise TypeError('img should be ndarray. Got {}'.format(type(img)))
574 | 
575 |         pil = Image.fromarray(img)
576 |         transform = self.get_params(self.brightness, self.contrast,
577 |                                     self.saturation, self.hue)
578 |         return np.array(transform(pil))
579 | 
580 | class Crop(object):
581 |     """Crops the given PIL Image to a rectangular region based on a given
582 |     4-tuple defining the left, upper pixel coordinated, hight and width size.
583 | 
584 |     Args:
585 |         a tuple: (upper pixel coordinate, left pixel coordinate, hight, width)-tuple
586 |     """
587 | 
588 |     def __init__(self, i, j, h, w):
589 |         """
590 |         i: Upper pixel coordinate.
591 |         j: Left pixel coordinate.
592 |         h: Height of the cropped image.
593 |         w: Width of the cropped image.
594 |         """
595 |         self.i = i
596 |         self.j = j
597 |         self.h = h
598 |         self.w = w
599 | 
600 |     def __call__(self, img):
601 |         """
602 |         Args:
603 |             img (numpy.ndarray (C x H x W)): Image to be cropped.
604 |         Returns:
605 |             img (numpy.ndarray (C x H x W)): Cropped image.
606 |         """
607 | 
608 |         i, j, h, w = self.i, self.j, self.h, self.w
609 | 
610 |         if not(_is_numpy_image(img)):
611 |             raise TypeError('img should be ndarray. Got {}'.format(type(img)))
612 |         if img.ndim == 3:
613 |             return img[i:i + h, j:j + w, :]
614 |         elif img.ndim == 2:
615 |             return img[i:i + h, j:j + w]
616 |         else:
617 |             raise RuntimeError(
618 |                 'img should be ndarray with 2 or 3 dimensions. Got {}'.format(img.ndim))
619 | 
620 |     def __repr__(self):
621 |         return self.__class__.__name__ + '(i={0},j={1},h={2},w={3})'.format(
622 |             self.i, self.j, self.h, self.w)
623 | 


--------------------------------------------------------------------------------
/deploy/data/depth.npy:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/dwofk/fast-depth/e68492011609c9bfb7de6d402da5d1d201d95bd9/deploy/data/depth.npy


--------------------------------------------------------------------------------
/deploy/data/depth.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/dwofk/fast-depth/e68492011609c9bfb7de6d402da5d1d201d95bd9/deploy/data/depth.png


--------------------------------------------------------------------------------
/deploy/data/pred.npy:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/dwofk/fast-depth/e68492011609c9bfb7de6d402da5d1d201d95bd9/deploy/data/pred.npy


--------------------------------------------------------------------------------
/deploy/data/pred.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/dwofk/fast-depth/e68492011609c9bfb7de6d402da5d1d201d95bd9/deploy/data/pred.png


--------------------------------------------------------------------------------
/deploy/data/rgb.npy:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/dwofk/fast-depth/e68492011609c9bfb7de6d402da5d1d201d95bd9/deploy/data/rgb.npy


--------------------------------------------------------------------------------
/deploy/data/rgb.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/dwofk/fast-depth/e68492011609c9bfb7de6d402da5d1d201d95bd9/deploy/data/rgb.png


--------------------------------------------------------------------------------
/deploy/data/visualize.py:
--------------------------------------------------------------------------------
 1 | import numpy as np
 2 | import matplotlib as mp
 3 | 
 4 | mp.use("pdf")
 5 | import matplotlib.pyplot as plt
 6 | 
 7 | cmap = plt.cm.viridis
 8 | 
 9 | def colored_depthmap(depth, d_min=None, d_max=None):
10 |     if d_min is None:
11 |         d_min = np.min(depth)
12 |     if d_max is None:
13 |         d_max = np.max(depth)
14 |     depth_relative = (depth - d_min) / (d_max - d_min)
15 |     return 255 * cmap(depth_relative)[:,:,:3] # HWC
16 | 
17 | def save_rgb_image(rgb_npy_fp, filename):
18 |     rgb_np = np.load(rgb_npy_fp)
19 |     rgb_scaled = 255 * np.transpose(np.squeeze(rgb_np), (0,1,2)) # HWC
20 |     mp.image.imsave('rgb.png', rgb_scaled.astype('uint8'))
21 | 
22 | def save_depth_image(depth_npy_fp, filename):
23 |     depth_np = np.load(depth_npy_fp)
24 |     depth_np_color = colored_depthmap(depth_np)
25 |     mp.image.imsave('depth.png', depth_np_color.astype('uint8'))
26 | 
27 | def save_pred_image(pred_npy_fp, filename):
28 |     pred_np = np.load(pred_npy_fp)
29 |     pred_np_2d = pred_np[0,0,:,:] # HW
30 |     pred_np_color = colored_depthmap(pred_np_2d)
31 |     mp.image.imsave('pred.png', pred_np_color.astype('uint8'))
32 | 
33 | save_rgb_image('rgb.npy', 'rgb.png')
34 | save_depth_image('depth.npy', 'depth.png')
35 | save_pred_image('pred.npy', 'pred.png')
36 | 


--------------------------------------------------------------------------------
/deploy/tx2_run_tvm.py:
--------------------------------------------------------------------------------
 1 | import tvm
 2 | import numpy as np
 3 | import argparse
 4 | import os
 5 | import time
 6 | 
 7 | def run_model(model_dir, input_fp, output_fp, warmup_trials, run_trials, cuda, try_randin):
 8 |     # import compiled graph
 9 |     print("=> [TVM on TX2] using model files in {}".format(model_dir))
10 |     assert(os.path.isdir(model_dir))
11 | 
12 |     print("=> [TVM on TX2] loading model lib and ptx")
13 |     loaded_lib = tvm.module.load(os.path.join(model_dir, "deploy_lib.o"))
14 |     if cuda:
15 |         dev_lib = tvm.module.load(os.path.join(model_dir, "deploy_cuda.ptx"))
16 |         loaded_lib.import_module(dev_lib)
17 | 
18 |     print("=> [TVM on TX2] loading model graph and params")
19 |     loaded_graph = open(os.path.join(model_dir,"deploy_graph.json")).read()
20 |     loaded_params = bytearray(open(os.path.join(model_dir, "deploy_param.params"), "rb").read())
21 |     
22 |     print("=> [TVM on TX2] creating TVM runtime module")
23 |     fcreate = tvm.get_global_func("tvm.graph_runtime.create")
24 |     ctx = tvm.gpu(0) if cuda else tvm.cpu(0)
25 |     gmodule = fcreate(loaded_graph, loaded_lib, ctx.device_type, ctx.device_id)
26 |     set_input, get_output, run = gmodule["set_input"], gmodule["get_output"], gmodule["run"]
27 | 
28 |     print("=> [TVM on TX2] feeding inputs and params into TVM module")
29 |     rgb_np = np.load(input_fp) # HWC
30 |     x = np.zeros([1,3,224,224]) # NCHW
31 |     x[0,:,:,:] = np.transpose(rgb_np, (2,0,1))
32 |     set_input('0', tvm.nd.array(x.astype('float32')))
33 |     gmodule["load_params"](loaded_params)
34 | 
35 |     print("=> [TVM on TX2] running TVM module, saving output")
36 |     run() # not gmodule.run()
37 |     out_shape = (1, 1, 224, 224)
38 |     out = tvm.nd.empty(out_shape, "float32")
39 |     get_output(0, out)
40 |     np.save(output_fp, out.asnumpy())
41 | 
42 |     print("=> [TVM on TX2] benchmarking: {} warmup, {} run trials".format(warmup_trials, run_trials))
43 |     # run model several times as a warmup
44 |     for i in range(warmup_trials):
45 |         run()
46 |         ctx.sync()
47 | 
48 |     # profile runtime using TVM time evaluator
49 |     ftimer = gmodule.time_evaluator("run", ctx, number=1, repeat=run_trials)
50 |     profile_result = ftimer()
51 |     profiled_runtime = profile_result[0]
52 |     
53 |     print("=> [TVM on TX2] profiled runtime (in ms): {:.5f}".format(1000*profiled_runtime))
54 | 
55 |     # try randomizing input
56 |     if try_randin:
57 |         randin_runtime = 0 
58 |         for i in range(run_trials):
59 |             x = np.random.randn(1, 3, 224, 224)
60 |             set_input('0', tvm.nd.array(x.astype('float32')))
61 |             randin_ftimer = gmodule.time_evaluator("run", ctx, number=1, repeat=1)
62 |             randin_profile_result = randin_ftimer()
63 |             randin_runtime += randin_profile_result[0]
64 |         randomized_input_runtime = randin_runtime/run_trials
65 |         print("=> [TVM on TX2] with randomized input on every run, profiled runtime (in ms): {:.5f}".format(1000*randomized_input_runtime))
66 | 
67 | def main():
68 |     parser = argparse.ArgumentParser()
69 |     parser.add_argument('--model-dir', type=str, required=True, 
70 |         help='path to folder with TVM-compiled model files (required)')
71 |     parser.add_argument('--input-fp', type=str, default='data/rgb.npy', 
72 |         help='numpy file containing input rgb data (default: data/rgb.npy')
73 |     parser.add_argument('--output-fp', type=str, default='data/pred.npy',
74 |         help='numpy file to store output prediction data (default: data/pred.npy')
75 | 
76 | 
77 |     parser.add_argument('--warmup', type=int, default=10, 
78 |         help='number of inference warmup trials (default: 10)')
79 |     parser.add_argument('--run', type=int, default=100,
80 |         help='number of inference run trials (default: 100)')
81 |     parser.add_argument('--cuda', type=bool, default=False,
82 |         help='run with CUDA (default: False)')
83 | 
84 |     parser.add_argument('--randin', type=bool, default=False,
85 |         help='profile runtime while randomizing input on every run (default: False)')
86 | 
87 |     args = parser.parse_args()
88 |     run_model(args.model_dir, args.input_fp, args.output_fp, args.warmup, args.run, args.cuda,  try_randin=args.randin)
89 | 
90 | if __name__ == '__main__':
91 |     main()
92 | 
93 | 


--------------------------------------------------------------------------------
/imagenet/__init__.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/dwofk/fast-depth/e68492011609c9bfb7de6d402da5d1d201d95bd9/imagenet/__init__.py


--------------------------------------------------------------------------------
/imagenet/mobilenet.py:
--------------------------------------------------------------------------------
 1 | import os
 2 | import shutil
 3 | import time
 4 | 
 5 | import torch
 6 | import torch.nn as nn
 7 | import torch.nn.parallel
 8 | import torch.backends.cudnn as cudnn
 9 | import torch.optim
10 | import torch.utils.data
11 | 
12 | class MobileNet(nn.Module):
13 |     def __init__(self, relu6=True):
14 |         super(MobileNet, self).__init__()
15 | 
16 |         def relu(relu6):
17 |             if relu6:
18 |                 return nn.ReLU6(inplace=True)
19 |             else:
20 |                 return nn.ReLU(inplace=True)
21 | 
22 |         def conv_bn(inp, oup, stride, relu6):
23 |             return nn.Sequential(
24 |                 nn.Conv2d(inp, oup, 3, stride, 1, bias=False),
25 |                 nn.BatchNorm2d(oup),
26 |                 relu(relu6),
27 |             )
28 | 
29 |         def conv_dw(inp, oup, stride, relu6):
30 |             return nn.Sequential(
31 |                 nn.Conv2d(inp, inp, 3, stride, 1, groups=inp, bias=False),
32 |                 nn.BatchNorm2d(inp),
33 |                 relu(relu6),
34 |     
35 |                 nn.Conv2d(inp, oup, 1, 1, 0, bias=False),
36 |                 nn.BatchNorm2d(oup),
37 |                 relu(relu6),
38 |             )
39 | 
40 |         self.model = nn.Sequential(
41 |             conv_bn(  3,  32, 2, relu6), 
42 |             conv_dw( 32,  64, 1, relu6),
43 |             conv_dw( 64, 128, 2, relu6),
44 |             conv_dw(128, 128, 1, relu6),
45 |             conv_dw(128, 256, 2, relu6),
46 |             conv_dw(256, 256, 1, relu6),
47 |             conv_dw(256, 512, 2, relu6),
48 |             conv_dw(512, 512, 1, relu6),
49 |             conv_dw(512, 512, 1, relu6),
50 |             conv_dw(512, 512, 1, relu6),
51 |             conv_dw(512, 512, 1, relu6),
52 |             conv_dw(512, 512, 1, relu6),
53 |             conv_dw(512, 1024, 2, relu6),
54 |             conv_dw(1024, 1024, 1, relu6),
55 |             nn.AvgPool2d(7),
56 |         )
57 |         self.fc = nn.Linear(1024, 1000)
58 | 
59 |     def forward(self, x):
60 |         x = self.model(x)
61 |         x = x.view(-1, 1024)
62 |         x = self.fc(x)
63 |         return x
64 | 
65 | def main():
66 |     import torchvision.models
67 |     model = MobileNet(relu6=True)
68 |     model = torch.nn.DataParallel(model).cuda()
69 |     model_filename = os.path.join('results', 'imagenet.arch=mobilenet.lr=0.1.bs=256', 'model_best.pth.tar')
70 |     if os.path.isfile(model_filename):
71 |         print("=> loading Imagenet pretrained model '{}'".format(model_filename))
72 |         checkpoint = torch.load(model_filename)
73 |         epoch = checkpoint['epoch']
74 |         best_prec1 = checkpoint['best_prec1']
75 |         model.load_state_dict(checkpoint['state_dict'])
76 |         print("=> loaded Imagenet pretrained model '{}' (epoch {}). best_prec1={}".format(model_filename, epoch, best_prec1))
77 | 
78 | if __name__ == '__main__':
79 |     main()
80 | 


--------------------------------------------------------------------------------
/img/acc_fps_cpu.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/dwofk/fast-depth/e68492011609c9bfb7de6d402da5d1d201d95bd9/img/acc_fps_cpu.png


--------------------------------------------------------------------------------
/img/acc_fps_gpu.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/dwofk/fast-depth/e68492011609c9bfb7de6d402da5d1d201d95bd9/img/acc_fps_gpu.png


--------------------------------------------------------------------------------
/img/visualization.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/dwofk/fast-depth/e68492011609c9bfb7de6d402da5d1d201d95bd9/img/visualization.png


--------------------------------------------------------------------------------
/main.py:
--------------------------------------------------------------------------------
  1 | import os
  2 | import time
  3 | import csv
  4 | import numpy as np
  5 | 
  6 | import torch
  7 | import torch.nn.parallel
  8 | import torch.backends.cudnn as cudnn
  9 | import torch.optim
 10 | cudnn.benchmark = True
 11 | 
 12 | import models
 13 | from metrics import AverageMeter, Result
 14 | import utils
 15 | 
 16 | args = utils.parse_command()
 17 | print(args)
 18 | os.environ["CUDA_VISIBLE_DEVICES"] = args.gpu # Set the GPU.
 19 | 
 20 | fieldnames = ['rmse', 'mae', 'delta1', 'absrel',
 21 |             'lg10', 'mse', 'delta2', 'delta3', 'data_time', 'gpu_time']
 22 | best_fieldnames = ['best_epoch'] + fieldnames
 23 | best_result = Result()
 24 | best_result.set_to_worst()
 25 | 
 26 | def main():
 27 |     global args, best_result, output_directory, train_csv, test_csv
 28 | 
 29 |     # Data loading code
 30 |     print("=> creating data loaders...")
 31 |     valdir = os.path.join('..', 'data', args.data, 'val')
 32 | 
 33 |     if args.data == 'nyudepthv2':
 34 |         from dataloaders.nyu import NYUDataset
 35 |         val_dataset = NYUDataset(valdir, split='val', modality=args.modality)
 36 |     else:
 37 |         raise RuntimeError('Dataset not found.')
 38 | 
 39 |     # set batch size to be 1 for validation
 40 |     val_loader = torch.utils.data.DataLoader(val_dataset,
 41 |         batch_size=1, shuffle=False, num_workers=args.workers, pin_memory=True)
 42 |     print("=> data loaders created.")
 43 | 
 44 |     # evaluation mode
 45 |     if args.evaluate:
 46 |         assert os.path.isfile(args.evaluate), \
 47 |         "=> no model found at '{}'".format(args.evaluate)
 48 |         print("=> loading model '{}'".format(args.evaluate))
 49 |         checkpoint = torch.load(args.evaluate)
 50 |         if type(checkpoint) is dict:
 51 |             args.start_epoch = checkpoint['epoch']
 52 |             best_result = checkpoint['best_result']
 53 |             model = checkpoint['model']
 54 |             print("=> loaded best model (epoch {})".format(checkpoint['epoch']))
 55 |         else:
 56 |             model = checkpoint
 57 |             args.start_epoch = 0
 58 |         output_directory = os.path.dirname(args.evaluate)
 59 |         validate(val_loader, model, args.start_epoch, write_to_file=False)
 60 |         return
 61 | 
 62 | 
 63 | def validate(val_loader, model, epoch, write_to_file=True):
 64 |     average_meter = AverageMeter()
 65 |     model.eval() # switch to evaluate mode
 66 |     end = time.time()
 67 |     for i, (input, target) in enumerate(val_loader):
 68 |         input, target = input.cuda(), target.cuda()
 69 |         # torch.cuda.synchronize()
 70 |         data_time = time.time() - end
 71 | 
 72 |         # compute output
 73 |         end = time.time()
 74 |         with torch.no_grad():
 75 |             pred = model(input)
 76 |         # torch.cuda.synchronize()
 77 |         gpu_time = time.time() - end
 78 | 
 79 |         # measure accuracy and record loss
 80 |         result = Result()
 81 |         result.evaluate(pred.data, target.data)
 82 |         average_meter.update(result, gpu_time, data_time, input.size(0))
 83 |         end = time.time()
 84 | 
 85 |         # save 8 images for visualization
 86 |         skip = 50
 87 | 
 88 |         if args.modality == 'rgb':
 89 |             rgb = input
 90 | 
 91 |         if i == 0:
 92 |             img_merge = utils.merge_into_row(rgb, target, pred)
 93 |         elif (i < 8*skip) and (i % skip == 0):
 94 |             row = utils.merge_into_row(rgb, target, pred)
 95 |             img_merge = utils.add_row(img_merge, row)
 96 |         elif i == 8*skip:
 97 |             filename = output_directory + '/comparison_' + str(epoch) + '.png'
 98 |             utils.save_image(img_merge, filename)
 99 | 
100 |         if (i+1) % args.print_freq == 0:
101 |             print('Test: [{0}/{1}]\t'
102 |                   't_GPU={gpu_time:.3f}({average.gpu_time:.3f})\n\t'
103 |                   'RMSE={result.rmse:.2f}({average.rmse:.2f}) '
104 |                   'MAE={result.mae:.2f}({average.mae:.2f}) '
105 |                   'Delta1={result.delta1:.3f}({average.delta1:.3f}) '
106 |                   'REL={result.absrel:.3f}({average.absrel:.3f}) '
107 |                   'Lg10={result.lg10:.3f}({average.lg10:.3f}) '.format(
108 |                    i+1, len(val_loader), gpu_time=gpu_time, result=result, average=average_meter.average()))
109 | 
110 |     avg = average_meter.average()
111 | 
112 |     print('\n*\n'
113 |         'RMSE={average.rmse:.3f}\n'
114 |         'MAE={average.mae:.3f}\n'
115 |         'Delta1={average.delta1:.3f}\n'
116 |         'REL={average.absrel:.3f}\n'
117 |         'Lg10={average.lg10:.3f}\n'
118 |         't_GPU={time:.3f}\n'.format(
119 |         average=avg, time=avg.gpu_time))
120 | 
121 |     if write_to_file:
122 |         with open(test_csv, 'a') as csvfile:
123 |             writer = csv.DictWriter(csvfile, fieldnames=fieldnames)
124 |             writer.writerow({'mse': avg.mse, 'rmse': avg.rmse, 'absrel': avg.absrel, 'lg10': avg.lg10,
125 |                 'mae': avg.mae, 'delta1': avg.delta1, 'delta2': avg.delta2, 'delta3': avg.delta3,
126 |                 'data_time': avg.data_time, 'gpu_time': avg.gpu_time})
127 |     return avg, img_merge
128 | 
129 | if __name__ == '__main__':
130 |     main()
131 | 


--------------------------------------------------------------------------------
/metrics.py:
--------------------------------------------------------------------------------
 1 | import torch
 2 | import math
 3 | import numpy as np
 4 | 
 5 | def log10(x):
 6 |       """Convert a new tensor with the base-10 logarithm of the elements of x. """
 7 |       return torch.log(x) / math.log(10)
 8 | 
 9 | class Result(object):
10 |     def __init__(self):
11 |         self.irmse, self.imae = 0, 0
12 |         self.mse, self.rmse, self.mae = 0, 0, 0
13 |         self.absrel, self.lg10 = 0, 0
14 |         self.delta1, self.delta2, self.delta3 = 0, 0, 0
15 |         self.data_time, self.gpu_time = 0, 0
16 | 
17 |     def set_to_worst(self):
18 |         self.irmse, self.imae = np.inf, np.inf
19 |         self.mse, self.rmse, self.mae = np.inf, np.inf, np.inf
20 |         self.absrel, self.lg10 = np.inf, np.inf
21 |         self.delta1, self.delta2, self.delta3 = 0, 0, 0
22 |         self.data_time, self.gpu_time = 0, 0
23 | 
24 |     def update(self, irmse, imae, mse, rmse, mae, absrel, lg10, delta1, delta2, delta3, gpu_time, data_time):
25 |         self.irmse, self.imae = irmse, imae
26 |         self.mse, self.rmse, self.mae = mse, rmse, mae
27 |         self.absrel, self.lg10 = absrel, lg10
28 |         self.delta1, self.delta2, self.delta3 = delta1, delta2, delta3
29 |         self.data_time, self.gpu_time = data_time, gpu_time
30 | 
31 |     def evaluate(self, output, target):
32 |         valid_mask = ((target>0) + (output>0)) > 0
33 | 
34 |         output = 1e3 * output[valid_mask]
35 |         target = 1e3 * target[valid_mask]
36 |         abs_diff = (output - target).abs()
37 | 
38 |         self.mse = float((torch.pow(abs_diff, 2)).mean())
39 |         self.rmse = math.sqrt(self.mse)
40 |         self.mae = float(abs_diff.mean())
41 |         self.lg10 = float((log10(output) - log10(target)).abs().mean())
42 |         self.absrel = float((abs_diff / target).mean())
43 | 
44 |         maxRatio = torch.max(output / target, target / output)
45 |         self.delta1 = float((maxRatio < 1.25).float().mean())
46 |         self.delta2 = float((maxRatio < 1.25 ** 2).float().mean())
47 |         self.delta3 = float((maxRatio < 1.25 ** 3).float().mean())
48 |         self.data_time = 0
49 |         self.gpu_time = 0
50 | 
51 |         inv_output = 1 / output
52 |         inv_target = 1 / target
53 |         abs_inv_diff = (inv_output - inv_target).abs()
54 |         self.irmse = math.sqrt((torch.pow(abs_inv_diff, 2)).mean())
55 |         self.imae = float(abs_inv_diff.mean())
56 | 
57 | 
58 | class AverageMeter(object):
59 |     def __init__(self):
60 |         self.reset()
61 | 
62 |     def reset(self):
63 |         self.count = 0.0
64 | 
65 |         self.sum_irmse, self.sum_imae = 0, 0
66 |         self.sum_mse, self.sum_rmse, self.sum_mae = 0, 0, 0
67 |         self.sum_absrel, self.sum_lg10 = 0, 0
68 |         self.sum_delta1, self.sum_delta2, self.sum_delta3 = 0, 0, 0
69 |         self.sum_data_time, self.sum_gpu_time = 0, 0
70 | 
71 |     def update(self, result, gpu_time, data_time, n=1):
72 |         self.count += n
73 | 
74 |         self.sum_irmse += n*result.irmse
75 |         self.sum_imae += n*result.imae
76 |         self.sum_mse += n*result.mse
77 |         self.sum_rmse += n*result.rmse
78 |         self.sum_mae += n*result.mae
79 |         self.sum_absrel += n*result.absrel
80 |         self.sum_lg10 += n*result.lg10
81 |         self.sum_delta1 += n*result.delta1
82 |         self.sum_delta2 += n*result.delta2
83 |         self.sum_delta3 += n*result.delta3
84 |         self.sum_data_time += n*data_time
85 |         self.sum_gpu_time += n*gpu_time
86 | 
87 |     def average(self):
88 |         avg = Result()
89 |         avg.update(
90 |             self.sum_irmse / self.count, self.sum_imae / self.count,
91 |             self.sum_mse / self.count, self.sum_rmse / self.count, self.sum_mae / self.count,
92 |             self.sum_absrel / self.count, self.sum_lg10 / self.count,
93 |             self.sum_delta1 / self.count, self.sum_delta2 / self.count, self.sum_delta3 / self.count,
94 |             self.sum_gpu_time / self.count, self.sum_data_time / self.count)
95 |         return avg


--------------------------------------------------------------------------------
/tvm_compile/tuning/tx2-cpu.mobilenet-nnconv5.trials=1000.stop=250.log:
--------------------------------------------------------------------------------
 1 | {"i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 32, 224, 224], "float32"], ["TENSOR", [1, 32, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 32, 224, 224, "float32"], [1, 32, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"t": "direct", "i": 7410, "c": null, "e": [["tile_co", "sp", [1, 1]], ["tile_oh", "sp", [14, 16]], ["tile_ow", "sp", [16, 14]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 9, 7, 8]], ["ann_reduce", "an", ["none", "unroll"]], ["ann_spatial", "an", ["none", "vec", "none"]]]}], "v": 0.1, "r": [[0.0038207925], 0, 1.7591049671173096, 1549478477.5367444]}
 2 | {"i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 64, 112, 112], "float32"], ["TENSOR", [32, 64, 5, 5], "float32"], [1, 1], [2, 2], "NCHW", "float32"], {}, ["conv2d", [1, 64, 112, 112, "float32"], [32, 64, 5, 5, "float32"], [1, 1], [2, 2], "NCHW", "float32"], {"t": "direct", "i": 66074, "c": null, "e": [["tile_co", "sp", [8, 4]], ["tile_oh", "sp", [28, 4]], ["tile_ow", "sp", [56, 2]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["unroll", "unroll"]], ["ann_spatial", "an", ["unroll", "unroll", "vec"]]]}], "v": 0.1, "r": [[0.066266069], 0, 5.3768932819366455, 1549479743.947523]}
 3 | {"i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 128, 56, 56], "float32"], ["TENSOR", [64, 128, 5, 5], "float32"], [1, 1], [2, 2], "NCHW", "float32"], {}, ["conv2d", [1, 128, 56, 56, "float32"], [64, 128, 5, 5, "float32"], [1, 1], [2, 2], "NCHW", "float32"], {"t": "direct", "i": 49401, "c": null, "e": [["tile_co", "sp", [16, 4]], ["tile_oh", "sp", [28, 2]], ["tile_ow", "sp", [14, 4]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["unroll", "unroll"]], ["ann_spatial", "an", ["unroll", "unroll", "vec"]]]}], "v": 0.1, "r": [[0.051313814], 0, 9.66435980796814, 1549480240.59895]}
 4 | {"i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 256, 28, 28], "float32"], ["TENSOR", [128, 256, 5, 5], "float32"], [1, 1], [2, 2], "NCHW", "float32"], {}, ["conv2d", [1, 256, 28, 28, "float32"], [128, 256, 5, 5, "float32"], [1, 1], [2, 2], "NCHW", "float32"], {"t": "direct", "i": 31787, "c": null, "e": [["tile_co", "sp", [16, 8]], ["tile_oh", "sp", [14, 2]], ["tile_ow", "sp", [7, 4]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["unroll", "unroll"]], ["ann_spatial", "an", ["unroll", "unroll", "vec"]]]}], "v": 0.1, "r": [[0.04352342825], 0, 3.9346649646759033, 1549480763.6560743]}
 5 | {"i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 512, 14, 14], "float32"], ["TENSOR", [256, 512, 5, 5], "float32"], [1, 1], [2, 2], "NCHW", "float32"], {}, ["conv2d", [1, 512, 14, 14, "float32"], [256, 512, 5, 5, "float32"], [1, 1], [2, 2], "NCHW", "float32"], {"t": "direct", "i": 6734, "c": null, "e": [["tile_co", "sp", [64, 4]], ["tile_oh", "sp", [14, 1]], ["tile_ow", "sp", [1, 14]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["unroll", "unroll"]], ["ann_spatial", "an", ["none", "unroll", "vec"]]]}], "v": 0.1, "r": [[0.03991823], 0, 3.566746950149536, 1549481391.9678566]}
 6 | {"i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 1024, 7, 7], "float32"], ["TENSOR", [512, 1024, 5, 5], "float32"], [1, 1], [2, 2], "NCHW", "float32"], {}, ["conv2d", [1, 1024, 7, 7, "float32"], [512, 1024, 5, 5, "float32"], [1, 1], [2, 2], "NCHW", "float32"], {"t": "direct", "i": 3463, "c": null, "e": [["tile_co", "sp", [64, 8]], ["tile_oh", "sp", [7, 1]], ["tile_ow", "sp", [1, 7]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["unroll", "unroll"]], ["ann_spatial", "an", ["unroll", "none", "vec"]]]}], "v": 0.1, "r": [[0.036470917], 0, 5.302130937576294, 1549481777.5447953]}
 7 | {"i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 1024, 7, 7], "float32"], ["TENSOR", [1024, 1024, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 1024, 7, 7, "float32"], [1024, 1024, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"t": "direct", "i": 4591, "c": null, "e": [["tile_co", "sp", [64, 16]], ["tile_oh", "sp", [1, 7]], ["tile_ow", "sp", [7, 1]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["none", "none"]], ["ann_spatial", "an", ["unroll", "unroll", "vec"]]]}], "v": 0.1, "r": [[0.0043714415], 0, 2.8076863288879395, 1549482462.154945]}
 8 | {"i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 1024, 7, 7], "float32"], ["TENSOR", [1024, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 1024, 7, 7, "float32"], [1024, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"t": "direct", "i": 455, "c": null, "e": [["tile_c", "sp", [64, 16]], ["tile_h", "sp", [1, 7]], ["tile_w", "sp", [7, 1]], ["ann", "an", ["unroll", "none", "vec"]]]}], "v": 0.1, "r": [[0.000124567], 0, 0.8551723957061768, 1549482638.7107537]}
 9 | {"i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 512, 7, 7], "float32"], ["TENSOR", [1024, 512, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 512, 7, 7, "float32"], [1024, 512, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"t": "direct", "i": 3667, "c": null, "e": [["tile_co", "sp", [64, 16]], ["tile_oh", "sp", [1, 7]], ["tile_ow", "sp", [7, 1]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 9, 7, 8]], ["ann_reduce", "an", ["none", "unroll"]], ["ann_spatial", "an", ["unroll", "none", "vec"]]]}], "v": 0.1, "r": [[0.0021189205], 0, 1.0700788497924805, 1549482899.913476]}
10 | {"i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 512, 14, 14], "float32"], ["TENSOR", [512, 1, 3, 3], "float32"], [2, 2], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 512, 14, 14, "float32"], [512, 1, 3, 3, "float32"], [2, 2], [1, 1], "float32"], {"t": "direct", "i": 412, "c": null, "e": [["tile_c", "sp", [128, 4]], ["tile_h", "sp", [1, 7]], ["tile_w", "sp", [7, 1]], ["ann", "an", ["unroll", "none", "vec"]]]}], "v": 0.1, "r": [[0.0001378475], 0, 1.3542444705963135, 1549483028.7694135]}
11 | {"i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 512, 14, 14], "float32"], ["TENSOR", [512, 512, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 512, 14, 14, "float32"], [512, 512, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"t": "direct", "i": 12884, "c": null, "e": [["tile_co", "sp", [32, 16]], ["tile_oh", "sp", [14, 1]], ["tile_ow", "sp", [2, 7]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["none", "none"]], ["ann_spatial", "an", ["unroll", "none", "vec"]]]}], "v": 0.1, "r": [[0.00439193], 0, 2.0783283710479736, 1549483215.359623]}
12 | {"i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 512, 14, 14], "float32"], ["TENSOR", [512, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 512, 14, 14, "float32"], [512, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"t": "direct", "i": 2112, "c": null, "e": [["tile_c", "sp", [128, 4]], ["tile_h", "sp", [1, 14]], ["tile_w", "sp", [14, 1]], ["ann", "an", ["unroll", "unroll", "vec"]]]}], "v": 0.1, "r": [[0.0002192565], 0, 0.5895946025848389, 1549483345.4555154]}
13 | {"i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 256, 14, 14], "float32"], ["TENSOR", [512, 256, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 256, 14, 14, "float32"], [512, 256, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"t": "direct", "i": 13364, "c": null, "e": [["tile_co", "sp", [32, 16]], ["tile_oh", "sp", [14, 1]], ["tile_ow", "sp", [2, 7]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 9, 7, 8]], ["ann_reduce", "an", ["none", "unroll"]], ["ann_spatial", "an", ["unroll", "none", "vec"]]]}], "v": 0.1, "r": [[0.00211282625], 0, 0.6594874858856201, 1549483478.8246555]}
14 | {"i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 256, 28, 28], "float32"], ["TENSOR", [256, 1, 3, 3], "float32"], [2, 2], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 256, 28, 28, "float32"], [256, 1, 3, 3, "float32"], [2, 2], [1, 1], "float32"], {"t": "direct", "i": 1928, "c": null, "e": [["tile_c", "sp", [64, 4]], ["tile_h", "sp", [2, 7]], ["tile_w", "sp", [7, 2]], ["ann", "an", ["unroll", "unroll", "vec"]]]}], "v": 0.1, "r": [[0.000224376], 0, 0.7611992359161377, 1549483607.5230222]}
15 | {"i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 256, 28, 28], "float32"], ["TENSOR", [256, 256, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 256, 28, 28, "float32"], [256, 256, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"t": "direct", "i": 6646, "c": null, "e": [["tile_co", "sp", [16, 16]], ["tile_oh", "sp", [28, 1]], ["tile_ow", "sp", [4, 7]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["unroll", "none"]], ["ann_spatial", "an", ["none", "none", "vec"]]]}], "v": 0.1, "r": [[0.004570014], 0, 1.2378997802734375, 1549483753.9382432]}
16 | {"i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 256, 28, 28], "float32"], ["TENSOR", [256, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 256, 28, 28, "float32"], [256, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"t": "direct", "i": 3332, "c": null, "e": [["tile_c", "sp", [64, 4]], ["tile_h", "sp", [2, 14]], ["tile_w", "sp", [14, 2]], ["ann", "an", ["unroll", "none", "vec"]]]}], "v": 0.1, "r": [[0.0005125755], 0, 0.9667601585388184, 1549483890.7859952]}
17 | {"i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 128, 28, 28], "float32"], ["TENSOR", [256, 128, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 128, 28, 28, "float32"], [256, 128, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"t": "direct", "i": 14692, "c": null, "e": [["tile_co", "sp", [16, 16]], ["tile_oh", "sp", [28, 1]], ["tile_ow", "sp", [7, 4]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 9, 7, 8]], ["ann_reduce", "an", ["unroll", "none"]], ["ann_spatial", "an", ["none", "unroll", "vec"]]]}], "v": 0.1, "r": [[0.00236881275], 0, 0.5344700813293457, 1549484047.5860903]}
18 | {"i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 128, 56, 56], "float32"], ["TENSOR", [128, 1, 3, 3], "float32"], [2, 2], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 128, 56, 56, "float32"], [128, 1, 3, 3, "float32"], [2, 2], [1, 1], "float32"], {"t": "direct", "i": 3818, "c": null, "e": [["tile_c", "sp", [32, 4]], ["tile_h", "sp", [4, 7]], ["tile_w", "sp", [14, 2]], ["ann", "an", ["unroll", "unroll", "vec"]]]}], "v": 0.1, "r": [[0.00043479925], 0, 0.5477871894836426, 1549484179.4244192]}
19 | {"i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 128, 56, 56], "float32"], ["TENSOR", [128, 128, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 128, 56, 56, "float32"], [128, 128, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"t": "direct", "i": 21315, "c": null, "e": [["tile_co", "sp", [16, 8]], ["tile_oh", "sp", [56, 1]], ["tile_ow", "sp", [4, 14]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 9, 7, 8]], ["ann_reduce", "an", ["none", "none"]], ["ann_spatial", "an", ["none", "unroll", "vec"]]]}], "v": 0.1, "r": [[0.005667829], 0, 1.7964444160461426, 1549484435.3359888]}
20 | {"i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 128, 56, 56], "float32"], ["TENSOR", [128, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 128, 56, 56, "float32"], [128, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"t": "direct", "i": 6698, "c": null, "e": [["tile_c", "sp", [32, 4]], ["tile_h", "sp", [4, 14]], ["tile_w", "sp", [56, 1]], ["ann", "an", ["unroll", "unroll", "vec"]]]}], "v": 0.1, "r": [[0.0010919125], 0, 1.353485107421875, 1549484632.265847]}
21 | {"i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 64, 56, 56], "float32"], ["TENSOR", [128, 64, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 64, 56, 56, "float32"], [128, 64, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"t": "direct", "i": 53955, "c": null, "e": [["tile_co", "sp", [16, 8]], ["tile_oh", "sp", [56, 1]], ["tile_ow", "sp", [8, 7]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 9, 7, 8]], ["ann_reduce", "an", ["none", "none"]], ["ann_spatial", "an", ["unroll", "unroll", "vec"]]]}], "v": 0.1, "r": [[0.00302072925], 0, 1.6285948753356934, 1549484777.1761692]}
22 | {"i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 64, 112, 112], "float32"], ["TENSOR", [64, 1, 3, 3], "float32"], [2, 2], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 64, 112, 112, "float32"], [64, 1, 3, 3, "float32"], [2, 2], [1, 1], "float32"], {"t": "direct", "i": 2417, "c": null, "e": [["tile_c", "sp", [16, 4]], ["tile_h", "sp", [28, 2]], ["tile_w", "sp", [8, 7]], ["ann", "an", ["none", "unroll", "vec"]]]}], "v": 0.1, "r": [[0.00089098925], 0, 1.8967888355255127, 1549484944.7167633]}
23 | {"i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 32, 112, 112], "float32"], ["TENSOR", [64, 32, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 32, 112, 112, "float32"], [64, 32, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"t": "direct", "i": 75618, "c": null, "e": [["tile_co", "sp", [4, 16]], ["tile_oh", "sp", [28, 4]], ["tile_ow", "sp", [112, 1]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["unroll", "none"]], ["ann_spatial", "an", ["unroll", "unroll", "vec"]]]}], "v": 0.1, "r": [[0.0042586435], 0, 0.5511445999145508, 1549485179.0085921]}
24 | {"i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 32, 112, 112], "float32"], ["TENSOR", [32, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 32, 112, 112, "float32"], [32, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"t": "direct", "i": 6086, "c": null, "e": [["tile_c", "sp", [8, 4]], ["tile_h", "sp", [14, 8]], ["tile_w", "sp", [56, 2]], ["ann", "an", ["unroll", "none", "vec"]]]}], "v": 0.1, "r": [[0.0013867625], 0, 0.6528422832489014, 1549485331.7410119]}
25 | {"i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 3, 224, 224], "float32"], ["TENSOR", [32, 3, 3, 3], "float32"], [2, 2], [1, 1], "NCHW", "float32"], {}, ["conv2d", [1, 3, 224, 224, "float32"], [32, 3, 3, 3, "float32"], [2, 2], [1, 1], "NCHW", "float32"], {"t": "direct", "i": 49329, "c": null, "e": [["tile_co", "sp", [4, 8]], ["tile_oh", "sp", [56, 2]], ["tile_ow", "sp", [28, 4]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["none", "unroll"]], ["ann_spatial", "an", ["unroll", "none", "vec"]]]}], "v": 0.1, "r": [[0.00171880075], 0, 0.7888402938842773, 1549485642.3346002]}
26 | 


--------------------------------------------------------------------------------
/tvm_compile/tuning/tx2-cpu.mobilenet-nnconv5dw-skipadd-pruned.trials=1000.stop=250.log:
--------------------------------------------------------------------------------
 1 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 16, 224, 224], "float32"], ["TENSOR", [1, 16, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 16, 224, 224, "float32"], [1, 16, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"t": "direct", "c": null, "e": [["tile_co", "sp", [1, 1]], ["tile_oh", "sp", [1, 224]], ["tile_ow", "sp", [14, 16]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 9, 7, 8]], ["ann_reduce", "an", ["unroll", "unroll"]], ["ann_spatial", "an", ["none", "vec", "unroll"]]], "i": 9155}], "r": [[0.00116569925], 0, 1.1003811359405518, 1548189982.8573406]}
 2 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 56, 112, 112], "float32"], ["TENSOR", [16, 56, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 56, 112, 112, "float32"], [16, 56, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"t": "direct", "c": null, "e": [["tile_co", "sp", [4, 4]], ["tile_oh", "sp", [112, 1]], ["tile_ow", "sp", [7, 16]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 9, 7, 8]], ["ann_reduce", "an", ["none", "unroll"]], ["ann_spatial", "an", ["unroll", "unroll", "vec"]]], "i": 53802}], "r": [[0.00162019275], 0, 0.8826901912689209, 1548190186.063045]}
 3 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 56, 112, 112], "float32"], ["TENSOR", [56, 1, 5, 5], "float32"], [1, 1], [2, 2], "float32"], {}, ["depthwise_conv2d_nchw", [1, 56, 112, 112, "float32"], [56, 1, 5, 5, "float32"], [1, 1], [2, 2], "float32"], {"t": "direct", "c": null, "e": [["tile_c", "sp", [14, 4]], ["tile_h", "sp", [7, 16]], ["tile_w", "sp", [56, 2]], ["ann", "an", ["unroll", "unroll", "vec"]]], "i": 10530}], "r": [[0.002066204], 0, 2.2892937660217285, 1548190458.9113562]}
 4 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 120, 56, 56], "float32"], ["TENSOR", [56, 120, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 120, 56, 56, "float32"], [56, 120, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"t": "direct", "c": null, "e": [["tile_co", "sp", [7, 8]], ["tile_oh", "sp", [56, 1]], ["tile_ow", "sp", [4, 14]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 9, 7, 8]], ["ann_reduce", "an", ["none", "none"]], ["ann_spatial", "an", ["unroll", "unroll", "vec"]]], "i": 54084}], "r": [[0.001725823], 0, 0.8528237342834473, 1548190698.1614182]}
 5 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 120, 56, 56], "float32"], ["TENSOR", [120, 1, 5, 5], "float32"], [1, 1], [2, 2], "float32"], {}, ["depthwise_conv2d_nchw", [1, 120, 56, 56, "float32"], [120, 1, 5, 5, "float32"], [1, 1], [2, 2], "float32"], {"t": "direct", "c": null, "e": [["tile_c", "sp", [30, 4]], ["tile_h", "sp", [4, 14]], ["tile_w", "sp", [56, 1]], ["ann", "an", ["unroll", "unroll", "vec"]]], "i": 13395}], "r": [[0.00095593075], 0, 1.2980289459228516, 1548190892.7128842]}
 6 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 256, 28, 28], "float32"], ["TENSOR", [120, 256, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 256, 28, 28, "float32"], [120, 256, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"t": "direct", "c": null, "e": [["tile_co", "sp", [15, 8]], ["tile_oh", "sp", [14, 2]], ["tile_ow", "sp", [4, 7]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 9, 7, 8]], ["ann_reduce", "an", ["unroll", "unroll"]], ["ann_spatial", "an", ["unroll", "none", "vec"]]], "i": 50422}], "r": [[0.00122721775], 0, 1.411975622177124, 1548191121.6755435]}
 7 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 256, 28, 28], "float32"], ["TENSOR", [256, 1, 5, 5], "float32"], [1, 1], [2, 2], "float32"], {}, ["depthwise_conv2d_nchw", [1, 256, 28, 28, "float32"], [256, 1, 5, 5, "float32"], [1, 1], [2, 2], "float32"], {"t": "direct", "c": null, "e": [["tile_c", "sp", [32, 8]], ["tile_h", "sp", [7, 4]], ["tile_w", "sp", [14, 2]], ["ann", "an", ["unroll", "none", "vec"]]], "i": 3315}], "r": [[0.000616226], 0, 0.5349960327148438, 1548191327.3241053]}
 8 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 200, 14, 14], "float32"], ["TENSOR", [256, 200, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 200, 14, 14, "float32"], [256, 200, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"t": "direct", "c": null, "e": [["tile_co", "sp", [64, 4]], ["tile_oh", "sp", [14, 1]], ["tile_ow", "sp", [1, 14]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["none", "unroll"]], ["ann_spatial", "an", ["unroll", "unroll", "vec"]]], "i": 15374}], "r": [[0.00046795525], 0, 1.549518346786499, 1548191591.9854326]}
 9 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 200, 14, 14], "float32"], ["TENSOR", [200, 1, 5, 5], "float32"], [1, 1], [2, 2], "float32"], {}, ["depthwise_conv2d_nchw", [1, 200, 14, 14, "float32"], [200, 1, 5, 5, "float32"], [1, 1], [2, 2], "float32"], {"t": "direct", "c": null, "e": [["tile_c", "sp", [50, 4]], ["tile_h", "sp", [1, 14]], ["tile_w", "sp", [7, 2]], ["ann", "an", ["unroll", "none", "vec"]]], "i": 2006}], "r": [[0.0001491905], 0, 1.5461997985839844, 1548191767.1845741]}
10 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 512, 7, 7], "float32"], ["TENSOR", [200, 512, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 512, 7, 7, "float32"], [200, 512, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"t": "direct", "c": null, "e": [["tile_co", "sp", [25, 8]], ["tile_oh", "sp", [7, 1]], ["tile_ow", "sp", [1, 7]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 9, 7, 8]], ["ann_reduce", "an", ["none", "none"]], ["ann_spatial", "an", ["unroll", "unroll", "vec"]]], "i": 5068}], "r": [[0.0003510605], 0, 0.4742095470428467, 1548191892.1782413]}
11 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 512, 7, 7], "float32"], ["TENSOR", [512, 1, 5, 5], "float32"], [1, 1], [2, 2], "float32"], {}, ["depthwise_conv2d_nchw", [1, 512, 7, 7, "float32"], [512, 1, 5, 5, "float32"], [1, 1], [2, 2], "float32"], {"t": "direct", "c": null, "e": [["tile_c", "sp", [64, 8]], ["tile_h", "sp", [1, 7]], ["tile_w", "sp", [7, 1]], ["ann", "an", ["unroll", "unroll", "vec"]]], "i": 533}], "r": [[0.0001124635], 0, 0.676560640335083, 1548191991.434562]}
12 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 480, 7, 7], "float32"], ["TENSOR", [512, 480, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 480, 7, 7, "float32"], [512, 480, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"t": "direct", "c": null, "e": [["tile_co", "sp", [32, 16]], ["tile_oh", "sp", [7, 1]], ["tile_ow", "sp", [1, 7]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["none", "none"]], ["ann_spatial", "an", ["none", "unroll", "vec"]]], "i": 1624}], "r": [[0.0005328295], 0, 1.406437635421753, 1548192119.4241006]}
13 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 480, 7, 7], "float32"], ["TENSOR", [480, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 480, 7, 7, "float32"], [480, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"t": "direct", "c": null, "e": [["tile_c", "sp", [30, 16]], ["tile_h", "sp", [1, 7]], ["tile_w", "sp", [7, 1]], ["ann", "an", ["unroll", "none", "vec"]]], "i": 994}], "r": [[6.130375e-05], 0, 1.4277822971343994, 1548192234.7198637]}
14 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 328, 7, 7], "float32"], ["TENSOR", [480, 328, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 328, 7, 7, "float32"], [480, 328, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"t": "direct", "c": null, "e": [["tile_co", "sp", [30, 16]], ["tile_oh", "sp", [7, 1]], ["tile_ow", "sp", [1, 7]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["unroll", "none"]], ["ann_spatial", "an", ["unroll", "unroll", "vec"]]], "i": 10426}], "r": [[0.00031776475], 0, 0.4927072525024414, 1548192479.2059495]}
15 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 328, 14, 14], "float32"], ["TENSOR", [328, 1, 3, 3], "float32"], [2, 2], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 328, 14, 14, "float32"], [328, 1, 3, 3, "float32"], [2, 2], [1, 1], "float32"], {"t": "direct", "c": null, "e": [["tile_c", "sp", [82, 4]], ["tile_h", "sp", [1, 7]], ["tile_w", "sp", [7, 1]], ["ann", "an", ["unroll", "none", "vec"]]], "i": 330}], "r": [[8.818325e-05], 0, 0.7043302059173584, 1548192581.0602925]}
16 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 296, 14, 14], "float32"], ["TENSOR", [328, 296, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 296, 14, 14, "float32"], [328, 296, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"t": "direct", "c": null, "e": [["tile_co", "sp", [41, 8]], ["tile_oh", "sp", [14, 1]], ["tile_ow", "sp", [1, 14]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["none", "none"]], ["ann_spatial", "an", ["none", "unroll", "vec"]]], "i": 5219}], "r": [[0.00074605375], 0, 2.0092272758483887, 1548192829.0489042]}
17 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 296, 14, 14], "float32"], ["TENSOR", [296, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 296, 14, 14, "float32"], [296, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"t": "direct", "c": null, "e": [["tile_c", "sp", [74, 4]], ["tile_h", "sp", [1, 14]], ["tile_w", "sp", [7, 2]], ["ann", "an", ["unroll", "none", "vec"]]], "i": 1338}], "r": [[0.0001298155], 0, 0.9202358722686768, 1548192962.0136871]}
18 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 288, 14, 14], "float32"], ["TENSOR", [296, 288, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 288, 14, 14, "float32"], [296, 288, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"t": "direct", "c": null, "e": [["tile_co", "sp", [37, 8]], ["tile_oh", "sp", [7, 2]], ["tile_ow", "sp", [2, 7]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 9, 7, 8]], ["ann_reduce", "an", ["unroll", "unroll"]], ["ann_spatial", "an", ["unroll", "none", "vec"]]], "i": 11211}], "r": [[0.00079630925], 0, 0.8467371463775635, 1548193166.7224784]}
19 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 288, 14, 14], "float32"], ["TENSOR", [288, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 288, 14, 14, "float32"], [288, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"t": "direct", "c": null, "e": [["tile_c", "sp", [72, 4]], ["tile_h", "sp", [1, 14]], ["tile_w", "sp", [7, 2]], ["ann", "an", ["unroll", "none", "vec"]]], "i": 3009}], "r": [[0.0001182315], 0, 0.7226011753082275, 1548193289.163559]}
20 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 272, 14, 14], "float32"], ["TENSOR", [288, 272, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 272, 14, 14, "float32"], [288, 272, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"t": "direct", "c": null, "e": [["tile_co", "sp", [36, 8]], ["tile_oh", "sp", [1, 14]], ["tile_ow", "sp", [7, 2]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 9, 7, 8]], ["ann_reduce", "an", ["none", "none"]], ["ann_spatial", "an", ["unroll", "none", "vec"]]], "i": 23459}], "r": [[0.00073303275], 0, 0.9118449687957764, 1548193427.542316]}
21 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 272, 14, 14], "float32"], ["TENSOR", [272, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 272, 14, 14, "float32"], [272, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"t": "direct", "c": null, "e": [["tile_c", "sp", [68, 4]], ["tile_h", "sp", [1, 14]], ["tile_w", "sp", [7, 2]], ["ann", "an", ["unroll", "none", "vec"]]], "i": 1672}], "r": [[0.0001212955], 0, 1.2856500148773193, 1548193581.368387]}
22 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 376, 14, 14], "float32"], ["TENSOR", [272, 376, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 376, 14, 14, "float32"], [272, 376, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"t": "direct", "c": null, "e": [["tile_co", "sp", [34, 8]], ["tile_oh", "sp", [7, 2]], ["tile_ow", "sp", [2, 7]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 9, 7, 8]], ["ann_reduce", "an", ["none", "unroll"]], ["ann_spatial", "an", ["none", "unroll", "vec"]]], "i": 6973}], "r": [[0.0008803325], 0, 0.6495048999786377, 1548193731.0867894]}
23 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 376, 14, 14], "float32"], ["TENSOR", [376, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 376, 14, 14, "float32"], [376, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"t": "direct", "c": null, "e": [["tile_c", "sp", [94, 4]], ["tile_h", "sp", [1, 14]], ["tile_w", "sp", [14, 1]], ["ann", "an", ["unroll", "none", "vec"]]], "i": 1306}], "r": [[0.0001601915], 0, 1.889883279800415, 1548193863.9992387]}
24 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 408, 14, 14], "float32"], ["TENSOR", [376, 408, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 408, 14, 14, "float32"], [376, 408, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"t": "direct", "c": null, "e": [["tile_co", "sp", [47, 8]], ["tile_oh", "sp", [7, 2]], ["tile_ow", "sp", [2, 7]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 9, 7, 8]], ["ann_reduce", "an", ["none", "none"]], ["ann_spatial", "an", ["unroll", "unroll", "vec"]]], "i": 13515}], "r": [[0.0011498275], 0, 0.4511895179748535, 1548193991.0625558]}
25 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 408, 14, 14], "float32"], ["TENSOR", [408, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 408, 14, 14, "float32"], [408, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"t": "direct", "c": null, "e": [["tile_c", "sp", [102, 4]], ["tile_h", "sp", [1, 14]], ["tile_w", "sp", [7, 2]], ["ann", "an", ["unroll", "none", "vec"]]], "i": 2675}], "r": [[0.000155239], 0, 0.7136528491973877, 1548194156.493113]}
26 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 256, 14, 14], "float32"], ["TENSOR", [408, 256, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 256, 14, 14, "float32"], [408, 256, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"t": "direct", "c": null, "e": [["tile_co", "sp", [51, 8]], ["tile_oh", "sp", [14, 1]], ["tile_ow", "sp", [1, 14]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 9, 7, 8]], ["ann_reduce", "an", ["none", "none"]], ["ann_spatial", "an", ["none", "unroll", "vec"]]], "i": 10693}], "r": [[0.00085011425], 0, 1.063706874847412, 1548194341.4088423]}
27 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 256, 28, 28], "float32"], ["TENSOR", [256, 1, 3, 3], "float32"], [2, 2], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 256, 28, 28, "float32"], [256, 1, 3, 3, "float32"], [2, 2], [1, 1], "float32"], {"t": "direct", "c": null, "e": [["tile_c", "sp", [32, 8]], ["tile_h", "sp", [7, 2]], ["tile_w", "sp", [2, 7]], ["ann", "an", ["unroll", "unroll", "vec"]]], "i": 1956}], "r": [[0.00021250325], 0, 0.9065120220184326, 1548194495.7803543]}
28 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 144, 28, 28], "float32"], ["TENSOR", [256, 144, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 144, 28, 28, "float32"], [256, 144, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"t": "direct", "c": null, "e": [["tile_co", "sp", [32, 8]], ["tile_oh", "sp", [28, 1]], ["tile_ow", "sp", [2, 14]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["none", "none"]], ["ann_spatial", "an", ["unroll", "unroll", "vec"]]], "i": 33915}], "r": [[0.00119772325], 0, 0.663311243057251, 1548194752.2460217]}
29 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 144, 28, 28], "float32"], ["TENSOR", [144, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 144, 28, 28, "float32"], [144, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"t": "direct", "c": null, "e": [["tile_c", "sp", [36, 4]], ["tile_h", "sp", [4, 7]], ["tile_w", "sp", [7, 4]], ["ann", "an", ["unroll", "unroll", "vec"]]], "i": 7248}], "r": [[0.00026295], 0, 1.8018908500671387, 1548194896.5325284]}
30 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 120, 28, 28], "float32"], ["TENSOR", [144, 120, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 120, 28, 28, "float32"], [144, 120, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"t": "direct", "c": null, "e": [["tile_co", "sp", [18, 8]], ["tile_oh", "sp", [28, 1]], ["tile_ow", "sp", [4, 7]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["unroll", "none"]], ["ann_spatial", "an", ["none", "unroll", "vec"]]], "i": 24035}], "r": [[0.00057663775], 0, 0.8634822368621826, 1548195053.5468194]}
31 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 120, 56, 56], "float32"], ["TENSOR", [120, 1, 3, 3], "float32"], [2, 2], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 120, 56, 56, "float32"], [120, 1, 3, 3, "float32"], [2, 2], [1, 1], "float32"], {"t": "direct", "c": null, "e": [["tile_c", "sp", [30, 4]], ["tile_h", "sp", [4, 7]], ["tile_w", "sp", [7, 4]], ["ann", "an", ["unroll", "unroll", "vec"]]], "i": 7731}], "r": [[0.00031674125], 0, 1.1324889659881592, 1548195190.1497178]}
32 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 88, 56, 56], "float32"], ["TENSOR", [120, 88, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 88, 56, 56, "float32"], [120, 88, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"t": "direct", "c": null, "e": [["tile_co", "sp", [15, 8]], ["tile_oh", "sp", [56, 1]], ["tile_ow", "sp", [7, 8]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 9, 7, 8]], ["ann_reduce", "an", ["unroll", "none"]], ["ann_spatial", "an", ["unroll", "unroll", "vec"]]], "i": 112134}], "r": [[0.001929271], 0, 1.1447086334228516, 1548195327.1973894]}
33 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 88, 56, 56], "float32"], ["TENSOR", [88, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 88, 56, 56, "float32"], [88, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"t": "direct", "c": null, "e": [["tile_c", "sp", [22, 4]], ["tile_h", "sp", [56, 1]], ["tile_w", "sp", [14, 4]], ["ann", "an", ["none", "unroll", "vec"]]], "i": 2690}], "r": [[0.00058562175], 0, 1.124485731124878, 1548195501.234142]}
34 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 56, 56, 56], "float32"], ["TENSOR", [88, 56, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 56, 56, 56, "float32"], [88, 56, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"t": "direct", "c": null, "e": [["tile_co", "sp", [11, 8]], ["tile_oh", "sp", [56, 1]], ["tile_ow", "sp", [7, 8]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 9, 7, 8]], ["ann_reduce", "an", ["unroll", "none"]], ["ann_spatial", "an", ["none", "none", "vec"]]], "i": 11011}], "r": [[0.000900928], 0, 1.7441248893737793, 1548195753.743164]}
35 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 56, 112, 112], "float32"], ["TENSOR", [56, 1, 3, 3], "float32"], [2, 2], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 56, 112, 112, "float32"], [56, 1, 3, 3, "float32"], [2, 2], [1, 1], "float32"], {"t": "direct", "c": null, "e": [["tile_c", "sp", [14, 4]], ["tile_h", "sp", [28, 2]], ["tile_w", "sp", [28, 2]], ["ann", "an", ["unroll", "unroll", "vec"]]], "i": 6730}], "r": [[0.0007634405], 0, 0.5247933864593506, 1548195883.5107229]}
36 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 16, 112, 112], "float32"], ["TENSOR", [56, 16, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 16, 112, 112, "float32"], [56, 16, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"t": "direct", "c": null, "e": [["tile_co", "sp", [14, 4]], ["tile_oh", "sp", [112, 1]], ["tile_ow", "sp", [16, 7]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["none", "unroll"]], ["ann_spatial", "an", ["none", "unroll", "vec"]]], "i": 33842}], "r": [[0.00138687475], 0, 0.5438559055328369, 1548196184.7103753]}
37 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 16, 112, 112], "float32"], ["TENSOR", [16, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 16, 112, 112, "float32"], [16, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"t": "direct", "c": null, "e": [["tile_c", "sp", [4, 4]], ["tile_h", "sp", [56, 2]], ["tile_w", "sp", [28, 4]], ["ann", "an", ["unroll", "unroll", "vec"]]], "i": 6607}], "r": [[0.0006212455], 0, 0.6399178504943848, 1548196338.593711]}
38 | {"v": 0.1, "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 3, 224, 224], "float32"], ["TENSOR", [16, 3, 3, 3], "float32"], [2, 2], [1, 1], "NCHW", "float32"], {}, ["conv2d", [1, 3, 224, 224, "float32"], [16, 3, 3, 3, "float32"], [2, 2], [1, 1], "NCHW", "float32"], {"t": "direct", "c": null, "e": [["tile_co", "sp", [4, 4]], ["tile_oh", "sp", [56, 2]], ["tile_ow", "sp", [28, 4]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 9, 7, 8]], ["ann_reduce", "an", ["unroll", "unroll"]], ["ann_spatial", "an", ["none", "none", "vec"]]], "i": 11607}], "r": [[0.00103108375], 0, 0.8736927509307861, 1548196616.90433]}
39 | 


--------------------------------------------------------------------------------
/tvm_compile/tuning/tx2-cpu.mobilenet-nnconv5dw-skipadd.trials=1000.stop=250.log:
--------------------------------------------------------------------------------
 1 | {"r": [[0.00218976625], 0, 0.38044118881225586, 1547737329.8211834], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 32, 224, 224], "float32"], ["TENSOR", [1, 32, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 32, 224, 224, "float32"], [1, 32, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"i": 7283, "e": [["tile_co", "sp", [1, 1]], ["tile_oh", "sp", [1, 224]], ["tile_ow", "sp", [14, 16]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["none", "unroll"]], ["ann_spatial", "an", ["none", "vec", "none"]]], "t": "direct", "c": null}], "v": 0.1}
 2 | {"r": [[0.002453024], 0, 0.8830351829528809, 1547737670.3910737], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 64, 112, 112], "float32"], ["TENSOR", [32, 64, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 64, 112, 112, "float32"], [32, 64, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"i": 65168, "e": [["tile_co", "sp", [8, 4]], ["tile_oh", "sp", [56, 2]], ["tile_ow", "sp", [7, 16]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["unroll", "none"]], ["ann_spatial", "an", ["unroll", "unroll", "vec"]]], "t": "direct", "c": null}], "v": 0.1}
 3 | {"r": [[0.00222503425], 0, 1.423985481262207, 1547737855.3445117], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 64, 112, 112], "float32"], ["TENSOR", [64, 1, 5, 5], "float32"], [1, 1], [2, 2], "float32"], {}, ["depthwise_conv2d_nchw", [1, 64, 112, 112, "float32"], [64, 1, 5, 5, "float32"], [1, 1], [2, 2], "float32"], {"i": 9130, "e": [["tile_c", "sp", [16, 4]], ["tile_h", "sp", [14, 8]], ["tile_w", "sp", [112, 1]], ["ann", "an", ["unroll", "unroll", "vec"]]], "t": "direct", "c": null}], "v": 0.1}
 4 | {"r": [[0.0017814385], 0, 0.6939797401428223, 1547738053.2760224], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 128, 56, 56], "float32"], ["TENSOR", [64, 128, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 128, 56, 56, "float32"], [64, 128, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"i": 8292, "e": [["tile_co", "sp", [4, 16]], ["tile_oh", "sp", [56, 1]], ["tile_ow", "sp", [7, 8]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["none", "unroll"]], ["ann_spatial", "an", ["none", "none", "vec"]]], "t": "direct", "c": null}], "v": 0.1}
 5 | {"r": [[0.001141443], 0, 0.5493149757385254, 1547738328.2206848], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 128, 56, 56], "float32"], ["TENSOR", [128, 1, 5, 5], "float32"], [1, 1], [2, 2], "float32"], {}, ["depthwise_conv2d_nchw", [1, 128, 56, 56, "float32"], [128, 1, 5, 5, "float32"], [1, 1], [2, 2], "float32"], {"i": 5386, "e": [["tile_c", "sp", [32, 4]], ["tile_h", "sp", [28, 2]], ["tile_w", "sp", [7, 8]], ["ann", "an", ["unroll", "none", "vec"]]], "t": "direct", "c": null}], "v": 0.1}
 6 | {"r": [[0.0012462585], 0, 1.180959701538086, 1547738477.9994175], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 256, 28, 28], "float32"], ["TENSOR", [128, 256, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 256, 28, 28, "float32"], [128, 256, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"i": 31299, "e": [["tile_co", "sp", [16, 8]], ["tile_oh", "sp", [28, 1]], ["tile_ow", "sp", [2, 14]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["unroll", "none"]], ["ann_spatial", "an", ["unroll", "unroll", "vec"]]], "t": "direct", "c": null}], "v": 0.1}
 7 | {"r": [[0.000699989], 0, 1.2548935413360596, 1547738647.2521586], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 256, 28, 28], "float32"], ["TENSOR", [256, 1, 5, 5], "float32"], [1, 1], [2, 2], "float32"], {}, ["depthwise_conv2d_nchw", [1, 256, 28, 28, "float32"], [256, 1, 5, 5, "float32"], [1, 1], [2, 2], "float32"], {"i": 3377, "e": [["tile_c", "sp", [64, 4]], ["tile_h", "sp", [4, 7]], ["tile_w", "sp", [7, 4]], ["ann", "an", ["unroll", "none", "vec"]]], "t": "direct", "c": null}], "v": 0.1}
 8 | {"r": [[0.001157523], 0, 0.510570764541626, 1547738817.8810546], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 512, 14, 14], "float32"], ["TENSOR", [256, 512, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 512, 14, 14, "float32"], [256, 512, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"i": 6735, "e": [["tile_co", "sp", [32, 8]], ["tile_oh", "sp", [14, 1]], ["tile_ow", "sp", [1, 14]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["unroll", "unroll"]], ["ann_spatial", "an", ["none", "unroll", "vec"]]], "t": "direct", "c": null}], "v": 0.1}
 9 | {"r": [[0.00032969425], 0, 1.0283777713775635, 1547738958.5527909], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 512, 14, 14], "float32"], ["TENSOR", [512, 1, 5, 5], "float32"], [1, 1], [2, 2], "float32"], {}, ["depthwise_conv2d_nchw", [1, 512, 14, 14, "float32"], [512, 1, 5, 5, "float32"], [1, 1], [2, 2], "float32"], {"i": 2112, "e": [["tile_c", "sp", [128, 4]], ["tile_h", "sp", [1, 14]], ["tile_w", "sp", [14, 1]], ["ann", "an", ["unroll", "unroll", "vec"]]], "t": "direct", "c": null}], "v": 0.1}
10 | {"r": [[0.001088833], 0, 1.3669013977050781, 1547739167.7177014], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 1024, 7, 7], "float32"], ["TENSOR", [512, 1024, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 1024, 7, 7, "float32"], [512, 1024, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"i": 942, "e": [["tile_co", "sp", [128, 4]], ["tile_oh", "sp", [7, 1]], ["tile_ow", "sp", [1, 7]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 9, 7, 8]], ["ann_reduce", "an", ["unroll", "unroll"]], ["ann_spatial", "an", ["none", "none", "vec"]]], "t": "direct", "c": null}], "v": 0.1}
11 | {"r": [[0.0002016545], 0, 1.0431005954742432, 1547739286.0030966], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 1024, 7, 7], "float32"], ["TENSOR", [1024, 1, 5, 5], "float32"], [1, 1], [2, 2], "float32"], {}, ["depthwise_conv2d_nchw", [1, 1024, 7, 7, "float32"], [1024, 1, 5, 5, "float32"], [1, 1], [2, 2], "float32"], {"i": 587, "e": [["tile_c", "sp", [64, 16]], ["tile_h", "sp", [1, 7]], ["tile_w", "sp", [7, 1]], ["ann", "an", ["unroll", "unroll", "vec"]]], "t": "direct", "c": null}], "v": 0.1}
12 | {"r": [[0.00204971875], 0, 1.4387884140014648, 1547739409.8011386], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 1024, 7, 7], "float32"], ["TENSOR", [1024, 1024, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 1024, 7, 7, "float32"], [1024, 1024, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"i": 729, "e": [["tile_co", "sp", [128, 8]], ["tile_oh", "sp", [7, 1]], ["tile_ow", "sp", [1, 7]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["none", "none"]], ["ann_spatial", "an", ["none", "none", "vec"]]], "t": "direct", "c": null}], "v": 0.1}
13 | {"r": [[0.000125911], 0, 0.5861396789550781, 1547739528.3804939], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 1024, 7, 7], "float32"], ["TENSOR", [1024, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 1024, 7, 7, "float32"], [1024, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"i": 455, "e": [["tile_c", "sp", [64, 16]], ["tile_h", "sp", [1, 7]], ["tile_w", "sp", [7, 1]], ["ann", "an", ["unroll", "none", "vec"]]], "t": "direct", "c": null}], "v": 0.1}
14 | {"r": [[0.0010378515], 0, 1.5314667224884033, 1547739659.6202457], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 512, 7, 7], "float32"], ["TENSOR", [1024, 512, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 512, 7, 7, "float32"], [1024, 512, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"i": 3799, "e": [["tile_co", "sp", [64, 16]], ["tile_oh", "sp", [1, 7]], ["tile_ow", "sp", [7, 1]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["unroll", "unroll"]], ["ann_spatial", "an", ["unroll", "none", "vec"]]], "t": "direct", "c": null}], "v": 0.1}
15 | {"r": [[0.0001344795], 0, 1.138458490371704, 1547739772.782865], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 512, 14, 14], "float32"], ["TENSOR", [512, 1, 3, 3], "float32"], [2, 2], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 512, 14, 14, "float32"], [512, 1, 3, 3, "float32"], [2, 2], [1, 1], "float32"], {"i": 412, "e": [["tile_c", "sp", [128, 4]], ["tile_h", "sp", [1, 7]], ["tile_w", "sp", [7, 1]], ["ann", "an", ["unroll", "none", "vec"]]], "t": "direct", "c": null}], "v": 0.1}
16 | {"r": [[0.00178856], 0, 0.4340789318084717, 1547740023.9774854], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 512, 14, 14], "float32"], ["TENSOR", [512, 512, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 512, 14, 14, "float32"], [512, 512, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"i": 16993, "e": [["tile_co", "sp", [64, 8]], ["tile_oh", "sp", [1, 14]], ["tile_ow", "sp", [14, 1]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["none", "unroll"]], ["ann_spatial", "an", ["unroll", "unroll", "vec"]]], "t": "direct", "c": null}], "v": 0.1}
17 | {"r": [[0.000210239], 0, 1.2972590923309326, 1547740177.1609933], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 512, 14, 14], "float32"], ["TENSOR", [512, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 512, 14, 14, "float32"], [512, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"i": 1672, "e": [["tile_c", "sp", [128, 4]], ["tile_h", "sp", [1, 14]], ["tile_w", "sp", [7, 2]], ["ann", "an", ["unroll", "none", "vec"]]], "t": "direct", "c": null}], "v": 0.1}
18 | {"r": [[0.0009993475], 0, 0.8285174369812012, 1547740306.637762], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 256, 14, 14], "float32"], ["TENSOR", [512, 256, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 256, 14, 14, "float32"], [512, 256, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"i": 17693, "e": [["tile_co", "sp", [64, 8]], ["tile_oh", "sp", [7, 2]], ["tile_ow", "sp", [2, 7]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["unroll", "unroll"]], ["ann_spatial", "an", ["unroll", "unroll", "vec"]]], "t": "direct", "c": null}], "v": 0.1}
19 | {"r": [[0.00019471025], 0, 1.4993979930877686, 1547740464.7630417], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 256, 28, 28], "float32"], ["TENSOR", [256, 1, 3, 3], "float32"], [2, 2], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 256, 28, 28, "float32"], [256, 1, 3, 3, "float32"], [2, 2], [1, 1], "float32"], {"i": 1489, "e": [["tile_c", "sp", [16, 16]], ["tile_h", "sp", [7, 2]], ["tile_w", "sp", [7, 2]], ["ann", "an", ["unroll", "none", "vec"]]], "t": "direct", "c": null}], "v": 0.1}
20 | {"r": [[0.00211812525], 0, 1.3814301490783691, 1547740591.591791], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 256, 28, 28], "float32"], ["TENSOR", [256, 256, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 256, 28, 28, "float32"], [256, 256, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"i": 28030, "e": [["tile_co", "sp", [16, 16]], ["tile_oh", "sp", [28, 1]], ["tile_ow", "sp", [4, 7]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["unroll", "unroll"]], ["ann_spatial", "an", ["unroll", "none", "vec"]]], "t": "direct", "c": null}], "v": 0.1}
21 | {"r": [[0.00051201125], 0, 1.5567152500152588, 1547740743.3866448], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 256, 28, 28], "float32"], ["TENSOR", [256, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 256, 28, 28, "float32"], [256, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"i": 3323, "e": [["tile_c", "sp", [64, 4]], ["tile_h", "sp", [4, 7]], ["tile_w", "sp", [14, 2]], ["ann", "an", ["unroll", "none", "vec"]]], "t": "direct", "c": null}], "v": 0.1}
22 | {"r": [[0.001125614], 0, 1.1552152633666992, 1547740871.2665403], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 128, 28, 28], "float32"], ["TENSOR", [256, 128, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 128, 28, 28, "float32"], [256, 128, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"i": 14692, "e": [["tile_co", "sp", [16, 16]], ["tile_oh", "sp", [28, 1]], ["tile_ow", "sp", [7, 4]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 9, 7, 8]], ["ann_reduce", "an", ["unroll", "none"]], ["ann_spatial", "an", ["none", "unroll", "vec"]]], "t": "direct", "c": null}], "v": 0.1}
23 | {"r": [[0.00040230925], 0, 1.6127750873565674, 1547741005.5860286], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 128, 56, 56], "float32"], ["TENSOR", [128, 1, 3, 3], "float32"], [2, 2], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 128, 56, 56, "float32"], [128, 1, 3, 3, "float32"], [2, 2], [1, 1], "float32"], {"i": 3802, "e": [["tile_c", "sp", [32, 4]], ["tile_h", "sp", [14, 2]], ["tile_w", "sp", [14, 2]], ["ann", "an", ["unroll", "unroll", "vec"]]], "t": "direct", "c": null}], "v": 0.1}
24 | {"r": [[0.00276929975], 0, 1.7560184001922607, 1547741243.1428356], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 128, 56, 56], "float32"], ["TENSOR", [128, 128, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 128, 56, 56, "float32"], [128, 128, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"i": 54595, "e": [["tile_co", "sp", [16, 8]], ["tile_oh", "sp", [56, 1]], ["tile_ow", "sp", [4, 14]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["none", "unroll"]], ["ann_spatial", "an", ["unroll", "unroll", "vec"]]], "t": "direct", "c": null}], "v": 0.1}
25 | {"r": [[0.00082235325], 0, 1.116337537765503, 1547741456.8146865], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 128, 56, 56], "float32"], ["TENSOR", [128, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 128, 56, 56, "float32"], [128, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"i": 2690, "e": [["tile_c", "sp", [32, 4]], ["tile_h", "sp", [56, 1]], ["tile_w", "sp", [14, 4]], ["ann", "an", ["none", "unroll", "vec"]]], "t": "direct", "c": null}], "v": 0.1}
26 | {"r": [[0.00145455625], 0, 1.3054676055908203, 1547741597.89791], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 64, 56, 56], "float32"], ["TENSOR", [128, 64, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 64, 56, 56, "float32"], [128, 64, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"i": 22787, "e": [["tile_co", "sp", [16, 8]], ["tile_oh", "sp", [56, 1]], ["tile_ow", "sp", [7, 8]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["unroll", "none"]], ["ann_spatial", "an", ["none", "unroll", "vec"]]], "t": "direct", "c": null}], "v": 0.1}
27 | {"r": [[0.00079617225], 0, 1.41221022605896, 1547741742.9475193], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 64, 112, 112], "float32"], ["TENSOR", [64, 1, 3, 3], "float32"], [2, 2], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 64, 112, 112, "float32"], [64, 1, 3, 3, "float32"], [2, 2], [1, 1], "float32"], {"i": 6001, "e": [["tile_c", "sp", [16, 4]], ["tile_h", "sp", [28, 2]], ["tile_w", "sp", [8, 7]], ["ann", "an", ["unroll", "unroll", "vec"]]], "t": "direct", "c": null}], "v": 0.1}
28 | {"r": [[0.002438363], 0, 0.6403403282165527, 1547742139.42199], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 32, 112, 112], "float32"], ["TENSOR", [64, 32, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 32, 112, 112, "float32"], [64, 32, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"i": 29122, "e": [["tile_co", "sp", [16, 4]], ["tile_oh", "sp", [112, 1]], ["tile_ow", "sp", [7, 16]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 9, 7, 8]], ["ann_reduce", "an", ["none", "none"]], ["ann_spatial", "an", ["none", "unroll", "vec"]]], "t": "direct", "c": null}], "v": 0.1}
29 | {"r": [[0.00110020375], 0, 1.101663589477539, 1547742312.38935], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 32, 112, 112], "float32"], ["TENSOR", [32, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 32, 112, 112, "float32"], [32, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"i": 7862, "e": [["tile_c", "sp", [8, 4]], ["tile_h", "sp", [112, 1]], ["tile_w", "sp", [56, 2]], ["ann", "an", ["unroll", "unroll", "vec"]]], "t": "direct", "c": null}], "v": 0.1}
30 | {"r": [[0.00148117875], 0, 1.003594160079956, 1547742463.732416], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 3, 224, 224], "float32"], ["TENSOR", [32, 3, 3, 3], "float32"], [2, 2], [1, 1], "NCHW", "float32"], {}, ["conv2d", [1, 3, 224, 224, "float32"], [32, 3, 3, 3, "float32"], [2, 2], [1, 1], "NCHW", "float32"], {"i": 65048, "e": [["tile_co", "sp", [8, 4]], ["tile_oh", "sp", [56, 2]], ["tile_ow", "sp", [14, 8]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["unroll", "none"]], ["ann_spatial", "an", ["unroll", "unroll", "vec"]]], "t": "direct", "c": null}], "v": 0.1}
31 | 


--------------------------------------------------------------------------------
/tvm_compile/tuning/tx2-cpu.mobilenet-nnconv5dw.trials=1000.stop=250.log:
--------------------------------------------------------------------------------
 1 | {"v": 0.1, "r": [[0.00211120525], 0, 1.6490120887756348, 1549463275.3477194], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 32, 224, 224], "float32"], ["TENSOR", [1, 32, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 32, 224, 224, "float32"], [1, 32, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"c": null, "i": 7706, "t": "direct", "e": [["tile_co", "sp", [1, 1]], ["tile_oh", "sp", [56, 4]], ["tile_ow", "sp", [14, 16]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 9, 7, 8]], ["ann_reduce", "an", ["unroll", "none"]], ["ann_spatial", "an", ["none", "vec", "none"]]]}]}
 2 | {"v": 0.1, "r": [[0.002641528], 0, 1.8992490768432617, 1549463522.9436731], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 64, 112, 112], "float32"], ["TENSOR", [32, 64, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 64, 112, 112, "float32"], [32, 64, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"c": null, "i": 62768, "t": "direct", "e": [["tile_co", "sp", [8, 4]], ["tile_oh", "sp", [56, 2]], ["tile_ow", "sp", [7, 16]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["none", "none"]], ["ann_spatial", "an", ["unroll", "unroll", "vec"]]]}]}
 3 | {"v": 0.1, "r": [[0.002179639], 0, 0.5747640132904053, 1549463794.6300278], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 64, 112, 112], "float32"], ["TENSOR", [64, 1, 5, 5], "float32"], [1, 1], [2, 2], "float32"], {}, ["depthwise_conv2d_nchw", [1, 64, 112, 112, "float32"], [64, 1, 5, 5, "float32"], [1, 1], [2, 2], "float32"], {"c": null, "i": 9193, "t": "direct", "e": [["tile_c", "sp", [16, 4]], ["tile_h", "sp", [16, 7]], ["tile_w", "sp", [56, 2]], ["ann", "an", ["unroll", "unroll", "vec"]]]}]}
 4 | {"v": 0.1, "r": [[0.00177205575], 0, 0.9056341648101807, 1549464008.0640798], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 128, 56, 56], "float32"], ["TENSOR", [64, 128, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 128, 56, 56, "float32"], [64, 128, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"c": null, "i": 8292, "t": "direct", "e": [["tile_co", "sp", [4, 16]], ["tile_oh", "sp", [56, 1]], ["tile_ow", "sp", [7, 8]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["none", "unroll"]], ["ann_spatial", "an", ["none", "none", "vec"]]]}]}
 5 | {"v": 0.1, "r": [[0.00111706975], 0, 1.5156757831573486, 1549464225.2041118], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 128, 56, 56], "float32"], ["TENSOR", [128, 1, 5, 5], "float32"], [1, 1], [2, 2], "float32"], {}, ["depthwise_conv2d_nchw", [1, 128, 56, 56, "float32"], [128, 1, 5, 5, "float32"], [1, 1], [2, 2], "float32"], {"c": null, "i": 5218, "t": "direct", "e": [["tile_c", "sp", [32, 4]], ["tile_h", "sp", [7, 8]], ["tile_w", "sp", [28, 2]], ["ann", "an", ["unroll", "none", "vec"]]]}]}
 6 | {"v": 0.1, "r": [[0.00118750475], 0, 1.0649268627166748, 1549464359.02914], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 256, 28, 28], "float32"], ["TENSOR", [128, 256, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 256, 28, 28, "float32"], [128, 256, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"c": null, "i": 31827, "t": "direct", "e": [["tile_co", "sp", [16, 8]], ["tile_oh", "sp", [28, 1]], ["tile_ow", "sp", [4, 7]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["unroll", "unroll"]], ["ann_spatial", "an", ["unroll", "unroll", "vec"]]]}]}
 7 | {"v": 0.1, "r": [[0.00073739475], 0, 0.8508305549621582, 1549464569.377289], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 256, 28, 28], "float32"], ["TENSOR", [256, 1, 5, 5], "float32"], [1, 1], [2, 2], "float32"], {}, ["depthwise_conv2d_nchw", [1, 256, 28, 28, "float32"], [256, 1, 5, 5, "float32"], [1, 1], [2, 2], "float32"], {"c": null, "i": 3377, "t": "direct", "e": [["tile_c", "sp", [64, 4]], ["tile_h", "sp", [4, 7]], ["tile_w", "sp", [7, 4]], ["ann", "an", ["unroll", "none", "vec"]]]}]}
 8 | {"v": 0.1, "r": [[0.00099514], 0, 0.7310960292816162, 1549464698.1474411], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 512, 14, 14], "float32"], ["TENSOR", [256, 512, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 512, 14, 14, "float32"], [256, 512, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"c": null, "i": 11892, "t": "direct", "e": [["tile_co", "sp", [32, 8]], ["tile_oh", "sp", [7, 2]], ["tile_ow", "sp", [2, 7]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["none", "unroll"]], ["ann_spatial", "an", ["unroll", "none", "vec"]]]}]}
 9 | {"v": 0.1, "r": [[0.0003016135], 0, 0.6347737312316895, 1549464859.3759325], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 512, 14, 14], "float32"], ["TENSOR", [512, 1, 5, 5], "float32"], [1, 1], [2, 2], "float32"], {}, ["depthwise_conv2d_nchw", [1, 512, 14, 14, "float32"], [512, 1, 5, 5, "float32"], [1, 1], [2, 2], "float32"], {"c": null, "i": 952, "t": "direct", "e": [["tile_c", "sp", [128, 4]], ["tile_h", "sp", [1, 14]], ["tile_w", "sp", [1, 14]], ["ann", "an", ["none", "unroll", "vec"]]]}]}
10 | {"v": 0.1, "r": [[0.0011650785], 0, 1.9754915237426758, 1549465000.6372852], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 1024, 7, 7], "float32"], ["TENSOR", [512, 1024, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 1024, 7, 7, "float32"], [512, 1024, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"c": null, "i": 813, "t": "direct", "e": [["tile_co", "sp", [64, 8]], ["tile_oh", "sp", [1, 7]], ["tile_ow", "sp", [7, 1]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["unroll", "none"]], ["ann_spatial", "an", ["none", "none", "vec"]]]}]}
11 | {"v": 0.1, "r": [[0.0002012395], 0, 0.5135416984558105, 1549465119.4674332], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 1024, 7, 7], "float32"], ["TENSOR", [1024, 1, 5, 5], "float32"], [1, 1], [2, 2], "float32"], {}, ["depthwise_conv2d_nchw", [1, 1024, 7, 7, "float32"], [1024, 1, 5, 5, "float32"], [1, 1], [2, 2], "float32"], {"c": null, "i": 587, "t": "direct", "e": [["tile_c", "sp", [64, 16]], ["tile_h", "sp", [1, 7]], ["tile_w", "sp", [7, 1]], ["ann", "an", ["unroll", "unroll", "vec"]]]}]}
12 | {"v": 0.1, "r": [[0.001770547], 0, 1.3441791534423828, 1549465333.001983], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 1024, 7, 7], "float32"], ["TENSOR", [1024, 1024, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 1024, 7, 7, "float32"], [1024, 1024, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"c": null, "i": 4733, "t": "direct", "e": [["tile_co", "sp", [128, 8]], ["tile_oh", "sp", [7, 1]], ["tile_ow", "sp", [1, 7]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 9, 7, 8]], ["ann_reduce", "an", ["none", "unroll"]], ["ann_spatial", "an", ["unroll", "unroll", "vec"]]]}]}
13 | {"v": 0.1, "r": [[0.000129191], 0, 0.7576065063476562, 1549465462.6113133], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 1024, 7, 7], "float32"], ["TENSOR", [1024, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 1024, 7, 7, "float32"], [1024, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"c": null, "i": 587, "t": "direct", "e": [["tile_c", "sp", [64, 16]], ["tile_h", "sp", [1, 7]], ["tile_w", "sp", [7, 1]], ["ann", "an", ["unroll", "unroll", "vec"]]]}]}
14 | {"v": 0.1, "r": [[0.001034831], 0, 1.3871076107025146, 1549465588.4565344], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 512, 7, 7], "float32"], ["TENSOR", [1024, 512, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 512, 7, 7, "float32"], [1024, 512, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"c": null, "i": 3711, "t": "direct", "e": [["tile_co", "sp", [64, 16]], ["tile_oh", "sp", [1, 7]], ["tile_ow", "sp", [7, 1]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["unroll", "none"]], ["ann_spatial", "an", ["unroll", "none", "vec"]]]}]}
15 | {"v": 0.1, "r": [[0.00013741475], 0, 0.7633135318756104, 1549465716.011089], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 512, 14, 14], "float32"], ["TENSOR", [512, 1, 3, 3], "float32"], [2, 2], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 512, 14, 14, "float32"], [512, 1, 3, 3, "float32"], [2, 2], [1, 1], "float32"], {"c": null, "i": 532, "t": "direct", "e": [["tile_c", "sp", [128, 4]], ["tile_h", "sp", [1, 7]], ["tile_w", "sp", [7, 1]], ["ann", "an", ["unroll", "unroll", "vec"]]]}]}
16 | {"v": 0.1, "r": [[0.0018755755], 0, 1.5863986015319824, 1549465865.7965143], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 512, 14, 14], "float32"], ["TENSOR", [512, 512, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 512, 14, 14, "float32"], [512, 512, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"c": null, "i": 7163, "t": "direct", "e": [["tile_co", "sp", [64, 8]], ["tile_oh", "sp", [14, 1]], ["tile_ow", "sp", [1, 14]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["unroll", "none"]], ["ann_spatial", "an", ["none", "unroll", "vec"]]]}]}
17 | {"v": 0.1, "r": [[0.00021360725], 0, 1.4937946796417236, 1549466005.5232635], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 512, 14, 14], "float32"], ["TENSOR", [512, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 512, 14, 14, "float32"], [512, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"c": null, "i": 1672, "t": "direct", "e": [["tile_c", "sp", [128, 4]], ["tile_h", "sp", [1, 14]], ["tile_w", "sp", [7, 2]], ["ann", "an", ["unroll", "none", "vec"]]]}]}
18 | {"v": 0.1, "r": [[0.0008690005], 0, 0.654231071472168, 1549466252.5401561], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 256, 14, 14], "float32"], ["TENSOR", [512, 256, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 256, 14, 14, "float32"], [512, 256, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"c": null, "i": 3604, "t": "direct", "e": [["tile_co", "sp", [32, 16]], ["tile_oh", "sp", [14, 1]], ["tile_ow", "sp", [2, 7]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["unroll", "unroll"]], ["ann_spatial", "an", ["none", "none", "vec"]]]}]}
19 | {"v": 0.1, "r": [[0.00021412725], 0, 1.4013373851776123, 1549466444.3549302], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 256, 28, 28], "float32"], ["TENSOR", [256, 1, 3, 3], "float32"], [2, 2], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 256, 28, 28, "float32"], [256, 1, 3, 3, "float32"], [2, 2], [1, 1], "float32"], {"c": null, "i": 1505, "t": "direct", "e": [["tile_c", "sp", [64, 4]], ["tile_h", "sp", [1, 14]], ["tile_w", "sp", [7, 2]], ["ann", "an", ["unroll", "none", "vec"]]]}]}
20 | {"v": 0.1, "r": [[0.0020150435], 0, 0.5900120735168457, 1549466623.6702547], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 256, 28, 28], "float32"], ["TENSOR", [256, 256, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 256, 28, 28, "float32"], [256, 256, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"c": null, "i": 27382, "t": "direct", "e": [["tile_co", "sp", [16, 16]], ["tile_oh", "sp", [28, 1]], ["tile_ow", "sp", [4, 7]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["unroll", "none"]], ["ann_spatial", "an", ["unroll", "none", "vec"]]]}]}
21 | {"v": 0.1, "r": [[0.00048243825], 0, 0.46898937225341797, 1549466769.5930836], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 256, 28, 28], "float32"], ["TENSOR", [256, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 256, 28, 28, "float32"], [256, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"c": null, "i": 3278, "t": "direct", "e": [["tile_c", "sp", [64, 4]], ["tile_h", "sp", [2, 14]], ["tile_w", "sp", [28, 1]], ["ann", "an", ["unroll", "none", "vec"]]]}]}
22 | {"v": 0.1, "r": [[0.00114010975], 0, 1.0067620277404785, 1549466900.7609394], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 128, 28, 28], "float32"], ["TENSOR", [256, 128, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 128, 28, 28, "float32"], [256, 128, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"c": null, "i": 26032, "t": "direct", "e": [["tile_co", "sp", [16, 16]], ["tile_oh", "sp", [28, 1]], ["tile_ow", "sp", [7, 4]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["none", "none"]], ["ann_spatial", "an", ["unroll", "none", "vec"]]]}]}
23 | {"v": 0.1, "r": [[0.00035789325], 0, 0.7935357093811035, 1549467113.2244818], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 128, 56, 56], "float32"], ["TENSOR", [128, 1, 3, 3], "float32"], [2, 2], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 128, 56, 56, "float32"], [128, 1, 3, 3, "float32"], [2, 2], [1, 1], "float32"], {"c": null, "i": 3819, "t": "direct", "e": [["tile_c", "sp", [16, 8]], ["tile_h", "sp", [4, 7]], ["tile_w", "sp", [14, 2]], ["ann", "an", ["unroll", "unroll", "vec"]]]}]}
24 | {"v": 0.1, "r": [[0.00285133], 0, 1.7403016090393066, 1549467283.6627488], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 128, 56, 56], "float32"], ["TENSOR", [128, 128, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 128, 56, 56, "float32"], [128, 128, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"c": null, "i": 53571, "t": "direct", "e": [["tile_co", "sp", [16, 8]], ["tile_oh", "sp", [56, 1]], ["tile_ow", "sp", [4, 14]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["none", "none"]], ["ann_spatial", "an", ["unroll", "unroll", "vec"]]]}]}
25 | {"v": 0.1, "r": [[0.0008334615], 0, 0.5077493190765381, 1549467438.3719728], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 128, 56, 56], "float32"], ["TENSOR", [128, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 128, 56, 56, "float32"], [128, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"c": null, "i": 5386, "t": "direct", "e": [["tile_c", "sp", [32, 4]], ["tile_h", "sp", [28, 2]], ["tile_w", "sp", [7, 8]], ["ann", "an", ["unroll", "none", "vec"]]]}]}
26 | {"v": 0.1, "r": [[0.00140673975], 0, 0.7620842456817627, 1549467661.36005], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 64, 56, 56], "float32"], ["TENSOR", [128, 64, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 64, 56, 56, "float32"], [128, 64, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"c": null, "i": 43267, "t": "direct", "e": [["tile_co", "sp", [16, 8]], ["tile_oh", "sp", [56, 1]], ["tile_ow", "sp", [7, 8]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["unroll", "none"]], ["ann_spatial", "an", ["unroll", "none", "vec"]]]}]}
27 | {"v": 0.1, "r": [[0.0008076695], 0, 1.599564790725708, 1549467809.7291658], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 64, 112, 112], "float32"], ["TENSOR", [64, 1, 3, 3], "float32"], [2, 2], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 64, 112, 112, "float32"], [64, 1, 3, 3, "float32"], [2, 2], [1, 1], "float32"], {"c": null, "i": 4601, "t": "direct", "e": [["tile_c", "sp", [16, 4]], ["tile_h", "sp", [28, 2]], ["tile_w", "sp", [14, 4]], ["ann", "an", ["unroll", "none", "vec"]]]}]}
28 | {"v": 0.1, "r": [[0.00245742775], 0, 0.8367326259613037, 1549467959.4589097], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 32, 112, 112], "float32"], ["TENSOR", [64, 32, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 32, 112, 112, "float32"], [64, 32, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"c": null, "i": 73222, "t": "direct", "e": [["tile_co", "sp", [16, 4]], ["tile_oh", "sp", [112, 1]], ["tile_ow", "sp", [7, 16]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["none", "none"]], ["ann_spatial", "an", ["unroll", "unroll", "vec"]]]}]}
29 | {"v": 0.1, "r": [[0.0010120925], 0, 0.7711925506591797, 1549468147.8392646], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 32, 112, 112], "float32"], ["TENSOR", [32, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 32, 112, 112, "float32"], [32, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"c": null, "i": 1273, "t": "direct", "e": [["tile_c", "sp", [16, 2]], ["tile_h", "sp", [28, 4]], ["tile_w", "sp", [56, 2]], ["ann", "an", ["none", "none", "vec"]]]}]}
30 | {"v": 0.1, "r": [[0.001093943], 0, 1.7501792907714844, 1549468286.1650174], "i": ["llvm -device=arm_cpu -target=aarch64-linux-gnu", "topi_nn_conv2d", [["TENSOR", [1, 3, 224, 224], "float32"], ["TENSOR", [32, 3, 3, 3], "float32"], [2, 2], [1, 1], "NCHW", "float32"], {}, ["conv2d", [1, 3, 224, 224, "float32"], [32, 3, 3, 3, "float32"], [2, 2], [1, 1], "NCHW", "float32"], {"c": null, "i": 64940, "t": "direct", "e": [["tile_co", "sp", [8, 4]], ["tile_oh", "sp", [16, 7]], ["tile_ow", "sp", [28, 4]], ["reorder_0", "re", [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], ["ann_reduce", "an", ["unroll", "none"]], ["ann_spatial", "an", ["unroll", "unroll", "vec"]]]}]}
31 | 


--------------------------------------------------------------------------------
/tvm_compile/tuning/tx2-gpu.mobilenet-nnconv5.trials=2000.stop=600.log:
--------------------------------------------------------------------------------
 1 | {"v": 0.1, "r": [[0.000195583], 0, 2.7407631874084473, 1549520386.8809931], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 32, 224, 224], "float32"], ["TENSOR", [1, 32, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 32, 224, 224, "float32"], [1, 32, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"c": null, "e": [["tile_f", "sp", [1, 1, 1, 1]], ["tile_y", "sp", [112, 1, 2, 1]], ["tile_x", "sp", [1, 2, 112, 1]], ["tile_rc", "sp", [16, 2]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 512], ["unroll_explicit", "ot", 0]], "i": 364908, "t": "direct"}]}
 2 | {"v": 0.1, "r": [[0.003038952], 0, 2.0225138664245605, 1549522084.273167], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 64, 112, 112], "float32"], ["TENSOR", [32, 64, 5, 5], "float32"], [1, 1], [2, 2], "NCHW", "float32"], {}, ["conv2d", [1, 64, 112, 112, "float32"], [32, 64, 5, 5, "float32"], [1, 1], [2, 2], "NCHW", "float32"], {"c": null, "e": [["tile_f", "sp", [1, 1, 8, 4]], ["tile_y", "sp", [7, 2, 4, 2]], ["tile_x", "sp", [14, 1, 4, 2]], ["tile_rc", "sp", [32, 2]], ["tile_ry", "sp", [1, 5]], ["tile_rx", "sp", [5, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 0]], "i": 70712365, "t": "direct"}]}
 3 | {"v": 0.1, "r": [[0.00275695625], 0, 5.14209771156311, 1549523356.497182], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 128, 56, 56], "float32"], ["TENSOR", [64, 128, 5, 5], "float32"], [1, 1], [2, 2], "NCHW", "float32"], {}, ["conv2d", [1, 128, 56, 56, "float32"], [64, 128, 5, 5, "float32"], [1, 1], [2, 2], "NCHW", "float32"], {"c": null, "e": [["tile_f", "sp", [4, 1, 16, 1]], ["tile_y", "sp", [2, 2, 2, 7]], ["tile_x", "sp", [7, 1, 4, 2]], ["tile_rc", "sp", [128, 1]], ["tile_ry", "sp", [1, 5]], ["tile_rx", "sp", [1, 5]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 0]], "i": 47582830, "t": "direct"}]}
 4 | {"v": 0.1, "r": [[0.00238482725], 0, 7.642461776733398, 1549525645.12028], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 256, 28, 28], "float32"], ["TENSOR", [128, 256, 5, 5], "float32"], [1, 1], [2, 2], "NCHW", "float32"], {}, ["conv2d", [1, 256, 28, 28, "float32"], [128, 256, 5, 5, "float32"], [1, 1], [2, 2], "NCHW", "float32"], {"c": null, "e": [["tile_f", "sp", [8, 1, 16, 1]], ["tile_y", "sp", [1, 1, 4, 7]], ["tile_x", "sp", [2, 1, 2, 7]], ["tile_rc", "sp", [64, 4]], ["tile_ry", "sp", [1, 5]], ["tile_rx", "sp", [1, 5]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "i": 40290626, "t": "direct"}]}
 5 | {"v": 0.1, "r": [[0.002864893], 0, 7.662211894989014, 1549526803.3654418], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 512, 14, 14], "float32"], ["TENSOR", [256, 512, 5, 5], "float32"], [1, 1], [2, 2], "NCHW", "float32"], {}, ["conv2d", [1, 512, 14, 14, "float32"], [256, 512, 5, 5, "float32"], [1, 1], [2, 2], "NCHW", "float32"], {"c": null, "e": [["tile_f", "sp", [4, 1, 64, 1]], ["tile_y", "sp", [1, 1, 2, 7]], ["tile_x", "sp", [2, 1, 1, 7]], ["tile_rc", "sp", [256, 2]], ["tile_ry", "sp", [1, 5]], ["tile_rx", "sp", [1, 5]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "i": 9791469, "t": "direct"}]}
 6 | {"v": 0.1, "r": [[0.003657288], 0, 5.146381616592407, 1549528047.2862632], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 1024, 7, 7], "float32"], ["TENSOR", [512, 1024, 5, 5], "float32"], [1, 1], [2, 2], "NCHW", "float32"], {}, ["conv2d", [1, 1024, 7, 7, "float32"], [512, 1024, 5, 5, "float32"], [1, 1], [2, 2], "NCHW", "float32"], {"c": null, "e": [["tile_f", "sp", [16, 2, 16, 1]], ["tile_y", "sp", [1, 1, 7, 1]], ["tile_x", "sp", [1, 1, 1, 7]], ["tile_rc", "sp", [512, 2]], ["tile_ry", "sp", [1, 5]], ["tile_rx", "sp", [1, 5]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "i": 897195, "t": "direct"}]}
 7 | {"v": 0.1, "r": [[0.0005521495], 0, 2.424633264541626, 1549529378.1834185], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 1024, 7, 7], "float32"], ["TENSOR", [1024, 1024, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 1024, 7, 7, "float32"], [1024, 1024, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"c": null, "e": [["tile_f", "sp", [16, 1, 64, 1]], ["tile_y", "sp", [1, 7, 1, 1]], ["tile_x", "sp", [1, 7, 1, 1]], ["tile_rc", "sp", [128, 8]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 512], ["unroll_explicit", "ot", 1]], "i": 216553, "t": "direct"}]}
 8 | {"v": 0.1, "r": [[4.76715e-05], 0, 1.9822640419006348, 1549530309.6870823], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 1024, 7, 7], "float32"], ["TENSOR", [1024, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 1024, 7, 7, "float32"], [1024, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"c": null, "e": [["tile_f", "sp", [128, 2, 1, 4]], ["tile_y", "sp", [1, 1, 7, 1]], ["tile_x", "sp", [1, 1, 7, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "i": 25862, "t": "direct"}]}
 9 | {"v": 0.1, "r": [[0.000293095], 0, 1.1059186458587646, 1549531195.1452584], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 512, 7, 7], "float32"], ["TENSOR", [1024, 512, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 512, 7, 7, "float32"], [1024, 512, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"c": null, "e": [["tile_f", "sp", [32, 2, 8, 2]], ["tile_y", "sp", [1, 1, 1, 7]], ["tile_x", "sp", [1, 1, 7, 1]], ["tile_rc", "sp", [64, 8]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 0], ["unroll_explicit", "ot", 1]], "i": 154248, "t": "direct"}]}
10 | {"v": 0.1, "r": [[4.919975e-05], 0, 2.612053871154785, 1549531758.9769905], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 512, 14, 14], "float32"], ["TENSOR", [512, 1, 3, 3], "float32"], [2, 2], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 512, 14, 14, "float32"], [512, 1, 3, 3, "float32"], [2, 2], [1, 1], "float32"], {"c": null, "e": [["tile_f", "sp", [64, 1, 4, 2]], ["tile_y", "sp", [1, 1, 7, 1]], ["tile_x", "sp", [1, 1, 7, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "i": 19872, "t": "direct"}]}
11 | {"v": 0.1, "r": [[0.0003131495], 0, 2.075115203857422, 1549532492.7750611], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 512, 14, 14], "float32"], ["TENSOR", [512, 512, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 512, 14, 14, "float32"], [512, 512, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"c": null, "e": [["tile_f", "sp", [8, 8, 8, 1]], ["tile_y", "sp", [2, 7, 1, 1]], ["tile_x", "sp", [1, 1, 14, 1]], ["tile_rc", "sp", [64, 8]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 0]], "i": 1323990, "t": "direct"}]}
12 | {"v": 0.1, "r": [[5.81515e-05], 0, 3.1690168380737305, 1549533246.0323243], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 512, 14, 14], "float32"], ["TENSOR", [512, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 512, 14, 14, "float32"], [512, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"c": null, "e": [["tile_f", "sp", [128, 1, 2, 2]], ["tile_y", "sp", [1, 2, 7, 1]], ["tile_x", "sp", [1, 1, 14, 1]], ["auto_unroll_max_step", "ot", 256], ["unroll_explicit", "ot", 1]], "i": 255044, "t": "direct"}]}
13 | {"v": 0.1, "r": [[0.00017847925], 0, 2.3081159591674805, 1549533918.8541245], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 256, 14, 14], "float32"], ["TENSOR", [512, 256, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 256, 14, 14, "float32"], [512, 256, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"c": null, "e": [["tile_f", "sp", [8, 4, 8, 2]], ["tile_y", "sp", [2, 1, 1, 7]], ["tile_x", "sp", [1, 1, 14, 1]], ["tile_rc", "sp", [32, 8]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 512], ["unroll_explicit", "ot", 0]], "i": 706721, "t": "direct"}]}
14 | {"v": 0.1, "r": [[7.766325e-05], 0, 2.9981043338775635, 1549534785.984523], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 256, 28, 28], "float32"], ["TENSOR", [256, 1, 3, 3], "float32"], [2, 2], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 256, 28, 28, "float32"], [256, 1, 3, 3, "float32"], [2, 2], [1, 1], "float32"], {"c": null, "e": [["tile_f", "sp", [64, 1, 4, 1]], ["tile_y", "sp", [1, 1, 7, 2]], ["tile_x", "sp", [2, 1, 7, 1]], ["auto_unroll_max_step", "ot", 256], ["unroll_explicit", "ot", 1]], "i": 186632, "t": "direct"}]}
15 | {"v": 0.1, "r": [[0.00035476675], 0, 2.6585144996643066, 1549535835.6294103], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 256, 28, 28], "float32"], ["TENSOR", [256, 256, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 256, 28, 28, "float32"], [256, 256, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"c": null, "e": [["tile_f", "sp", [4, 8, 8, 1]], ["tile_y", "sp", [7, 1, 1, 4]], ["tile_x", "sp", [1, 2, 14, 1]], ["tile_rc", "sp", [32, 8]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 512], ["unroll_explicit", "ot", 0]], "i": 3278082, "t": "direct"}]}
16 | {"v": 0.1, "r": [[8.427175e-05], 0, 1.690295934677124, 1549536635.6460495], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 256, 28, 28], "float32"], ["TENSOR", [256, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 256, 28, 28, "float32"], [256, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"c": null, "e": [["tile_f", "sp", [256, 1, 1, 1]], ["tile_y", "sp", [1, 1, 4, 7]], ["tile_x", "sp", [1, 1, 28, 1]], ["auto_unroll_max_step", "ot", 256], ["unroll_explicit", "ot", 1]], "i": 1173975, "t": "direct"}]}
17 | {"v": 0.1, "r": [[0.00021443025], 0, 4.84061861038208, 1549537368.9350924], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 128, 28, 28], "float32"], ["TENSOR", [256, 128, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 128, 28, 28, "float32"], [256, 128, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"c": null, "e": [["tile_f", "sp", [4, 1, 8, 8]], ["tile_y", "sp", [7, 4, 1, 1]], ["tile_x", "sp", [1, 2, 14, 1]], ["tile_rc", "sp", [16, 8]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 0]], "i": 5122054, "t": "direct"}]}
18 | {"v": 0.1, "r": [[0.0001168395], 0, 1.664238691329956, 1549538332.091168], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 128, 56, 56], "float32"], ["TENSOR", [128, 1, 3, 3], "float32"], [2, 2], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 128, 56, 56, "float32"], [128, 1, 3, 3, "float32"], [2, 2], [1, 1], "float32"], {"c": null, "e": [["tile_f", "sp", [128, 1, 1, 1]], ["tile_y", "sp", [1, 1, 14, 2]], ["tile_x", "sp", [1, 1, 28, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "i": 1044720, "t": "direct"}]}
19 | {"v": 0.1, "r": [[0.00031253325], 0, 2.2684757709503174, 1549539680.4623065], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 128, 56, 56], "float32"], ["TENSOR", [128, 128, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 128, 56, 56, "float32"], [128, 128, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"c": null, "e": [["tile_f", "sp", [2, 1, 8, 8]], ["tile_y", "sp", [28, 1, 2, 1]], ["tile_x", "sp", [1, 7, 8, 1]], ["tile_rc", "sp", [16, 8]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 512], ["unroll_explicit", "ot", 1]], "i": 27101857, "t": "direct"}]}
20 | {"v": 0.1, "r": [[0.0001404955], 0, 1.7230756282806396, 1549540462.374373], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 128, 56, 56], "float32"], ["TENSOR", [128, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 128, 56, 56, "float32"], [128, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"c": null, "e": [["tile_f", "sp", [128, 1, 1, 1]], ["tile_y", "sp", [1, 1, 8, 7]], ["tile_x", "sp", [1, 1, 56, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "i": 4126320, "t": "direct"}]}
21 | {"v": 0.1, "r": [[0.00020255025], 0, 1.367140293121338, 1549541573.3382344], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 64, 56, 56], "float32"], ["TENSOR", [128, 64, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 64, 56, 56, "float32"], [128, 64, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"c": null, "e": [["tile_f", "sp", [2, 1, 8, 8]], ["tile_y", "sp", [28, 1, 2, 1]], ["tile_x", "sp", [1, 7, 8, 1]], ["tile_rc", "sp", [8, 8]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 0], ["unroll_explicit", "ot", 1]], "i": 18653857, "t": "direct"}]}
22 | {"v": 0.1, "r": [[0.00021456725], 0, 1.408292531967163, 1549542340.7373421], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 64, 112, 112], "float32"], ["TENSOR", [64, 1, 3, 3], "float32"], [2, 2], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 64, 112, 112, "float32"], [64, 1, 3, 3, "float32"], [2, 2], [1, 1], "float32"], {"c": null, "e": [["tile_f", "sp", [64, 1, 1, 1]], ["tile_y", "sp", [7, 2, 4, 1]], ["tile_x", "sp", [1, 1, 56, 1]], ["auto_unroll_max_step", "ot", 256], ["unroll_explicit", "ot", 1]], "i": 2346540, "t": "direct"}]}
23 | {"v": 0.1, "r": [[0.000223671], 0, 3.9969608783721924, 1549543261.223412], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 32, 112, 112], "float32"], ["TENSOR", [64, 32, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 32, 112, 112, "float32"], [64, 32, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"c": null, "e": [["tile_f", "sp", [2, 2, 2, 8]], ["tile_y", "sp", [56, 1, 2, 1]], ["tile_x", "sp", [1, 7, 16, 1]], ["tile_rc", "sp", [8, 4]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 0], ["unroll_explicit", "ot", 0]], "i": 3740589, "t": "direct"}]}
24 | {"v": 0.1, "r": [[0.00013695075], 0, 2.816110610961914, 1549544188.9147506], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 32, 112, 112], "float32"], ["TENSOR", [32, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 32, 112, 112, "float32"], [32, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"c": null, "e": [["tile_f", "sp", [32, 1, 1, 1]], ["tile_y", "sp", [4, 1, 4, 7]], ["tile_x", "sp", [1, 1, 112, 1]], ["auto_unroll_max_step", "ot", 256], ["unroll_explicit", "ot", 1]], "i": 4741072, "t": "direct"}]}
25 | {"v": 0.1, "r": [[0.00019237425], 0, 2.6703319549560547, 1549545517.7321806], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 3, 224, 224], "float32"], ["TENSOR", [32, 3, 3, 3], "float32"], [2, 2], [1, 1], "NCHW", "float32"], {}, ["conv2d", [1, 3, 224, 224, "float32"], [32, 3, 3, 3, "float32"], [2, 2], [1, 1], "NCHW", "float32"], {"c": null, "e": [["tile_f", "sp", [1, 1, 8, 4]], ["tile_y", "sp", [56, 1, 1, 2]], ["tile_x", "sp", [2, 1, 56, 1]], ["tile_rc", "sp", [1, 3]], ["tile_ry", "sp", [1, 3]], ["tile_rx", "sp", [1, 3]], ["auto_unroll_max_step", "ot", 512], ["unroll_explicit", "ot", 1]], "i": 43138245, "t": "direct"}]}
26 | 


--------------------------------------------------------------------------------
/tvm_compile/tuning/tx2-gpu.mobilenet-nnconv5dw-skipadd-pruned.trials=2000.stop=600.log:
--------------------------------------------------------------------------------
 1 | {"i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 16, 224, 224], "float32"], ["TENSOR", [1, 16, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 16, 224, 224, "float32"], [1, 16, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"i": 314944, "e": [["tile_f", "sp", [1, 1, 1, 1]], ["tile_y", "sp", [224, 1, 1, 1]], ["tile_x", "sp", [1, 1, 224, 1]], ["tile_rc", "sp", [8, 2]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 512], ["unroll_explicit", "ot", 0]], "t": "direct", "c": null}], "r": [[0.0001044715], 0, 1.65395188331604, 1548646492.0790715], "v": 0.1}
 2 | {"i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 56, 112, 112], "float32"], ["TENSOR", [16, 56, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 56, 112, 112, "float32"], [16, 56, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"i": 13920913, "e": [["tile_f", "sp", [1, 2, 8, 1]], ["tile_y", "sp", [112, 1, 1, 1]], ["tile_x", "sp", [1, 4, 28, 1]], ["tile_rc", "sp", [7, 8]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 0]], "t": "direct", "c": null}], "r": [[0.000170911], 0, 2.5805726051330566, 1548647107.7181647], "v": 0.1}
 3 | {"i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 56, 112, 112], "float32"], ["TENSOR", [56, 1, 5, 5], "float32"], [1, 1], [2, 2], "float32"], {}, ["depthwise_conv2d_nchw", [1, 56, 112, 112, "float32"], [56, 1, 5, 5, "float32"], [1, 1], [2, 2], "float32"], {"i": 6693600, "e": [["tile_f", "sp", [56, 1, 1, 1]], ["tile_y", "sp", [2, 1, 14, 4]], ["tile_x", "sp", [7, 1, 16, 1]], ["auto_unroll_max_step", "ot", 256], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null}], "r": [[0.0002963425], 0, 2.6830427646636963, 1548648186.9543624], "v": 0.1}
 4 | {"i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 120, 56, 56], "float32"], ["TENSOR", [56, 120, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 120, 56, 56, "float32"], [56, 120, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"i": 9907345, "e": [["tile_f", "sp", [1, 2, 4, 7]], ["tile_y", "sp", [28, 2, 1, 1]], ["tile_x", "sp", [1, 2, 28, 1]], ["tile_rc", "sp", [30, 4]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 512], ["unroll_explicit", "ot", 0]], "t": "direct", "c": null}], "r": [[0.00020867025], 0, 1.3627350330352783, 1548649315.0133436], "v": 0.1}
 5 | {"i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 120, 56, 56], "float32"], ["TENSOR", [120, 1, 5, 5], "float32"], [1, 1], [2, 2], "float32"], {}, ["depthwise_conv2d_nchw", [1, 120, 56, 56, "float32"], [120, 1, 5, 5, "float32"], [1, 1], [2, 2], "float32"], {"i": 11456640, "e": [["tile_f", "sp", [120, 1, 1, 1]], ["tile_y", "sp", [4, 1, 7, 2]], ["tile_x", "sp", [1, 1, 28, 2]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null}], "r": [[0.0001770705], 0, 2.9571971893310547, 1548650130.5959702], "v": 0.1}
 6 | {"i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 256, 28, 28], "float32"], ["TENSOR", [120, 256, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 256, 28, 28, "float32"], [120, 256, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"i": 15587259, "e": [["tile_f", "sp", [5, 3, 8, 1]], ["tile_y", "sp", [4, 1, 1, 7]], ["tile_x", "sp", [1, 1, 28, 1]], ["tile_rc", "sp", [32, 8]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 0], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null}], "r": [[0.000176463], 0, 1.957888126373291, 1548652038.6319573], "v": 0.1}
 7 | {"i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 256, 28, 28], "float32"], ["TENSOR", [256, 1, 5, 5], "float32"], [1, 1], [2, 2], "float32"], {}, ["depthwise_conv2d_nchw", [1, 256, 28, 28, "float32"], [256, 1, 5, 5, "float32"], [1, 1], [2, 2], "float32"], {"i": 1495890, "e": [["tile_f", "sp", [256, 1, 1, 1]], ["tile_y", "sp", [1, 1, 14, 2]], ["tile_x", "sp", [1, 1, 14, 2]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null}], "r": [[0.000116847], 0, 2.5286483764648438, 1548652840.9417276], "v": 0.1}
 8 | {"i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 200, 14, 14], "float32"], ["TENSOR", [256, 200, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 200, 14, 14, "float32"], [256, 200, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"i": 697884, "e": [["tile_f", "sp", [8, 1, 8, 4]], ["tile_y", "sp", [1, 7, 2, 1]], ["tile_x", "sp", [1, 1, 14, 1]], ["tile_rc", "sp", [25, 8]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 512], ["unroll_explicit", "ot", 0]], "t": "direct", "c": null}], "r": [[7.766325e-05], 0, 1.3543050289154053, 1548653774.0519202], "v": 0.1}
 9 | {"i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 200, 14, 14], "float32"], ["TENSOR", [200, 1, 5, 5], "float32"], [1, 1], [2, 2], "float32"], {}, ["depthwise_conv2d_nchw", [1, 200, 14, 14, "float32"], [200, 1, 5, 5, "float32"], [1, 1], [2, 2], "float32"], {"i": 273621, "e": [["tile_f", "sp", [50, 1, 4, 1]], ["tile_y", "sp", [1, 1, 14, 1]], ["tile_x", "sp", [1, 7, 2, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null}], "r": [[3.818375e-05], 0, 1.5769789218902588, 1548654909.867864], "v": 0.1}
10 | {"i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 512, 7, 7], "float32"], ["TENSOR", [200, 512, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 512, 7, 7, "float32"], [200, 512, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"i": 43436, "e": [["tile_f", "sp", [5, 5, 8, 1]], ["tile_y", "sp", [1, 7, 1, 1]], ["tile_x", "sp", [1, 1, 7, 1]], ["tile_rc", "sp", [64, 8]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 512], ["unroll_explicit", "ot", 0]], "t": "direct", "c": null}], "r": [[9.72875e-05], 0, 2.179405450820923, 1548655486.6924798], "v": 0.1}
11 | {"i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 512, 7, 7], "float32"], ["TENSOR", [512, 1, 5, 5], "float32"], [1, 1], [2, 2], "float32"], {}, ["depthwise_conv2d_nchw", [1, 512, 7, 7, "float32"], [512, 1, 5, 5, "float32"], [1, 1], [2, 2], "float32"], {"i": 19900, "e": [["tile_f", "sp", [128, 1, 1, 4]], ["tile_y", "sp", [1, 1, 7, 1]], ["tile_x", "sp", [1, 1, 7, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null}], "r": [[3.839175e-05], 0, 2.85693097114563, 1548656199.745016], "v": 0.1}
12 | {"i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 480, 7, 7], "float32"], ["TENSOR", [512, 480, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 480, 7, 7, "float32"], [512, 480, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"i": 287630, "e": [["tile_f", "sp", [8, 1, 32, 2]], ["tile_y", "sp", [1, 1, 1, 7]], ["tile_x", "sp", [1, 1, 7, 1]], ["tile_rc", "sp", [32, 15]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 0], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null}], "r": [[0.00013743925], 0, 1.9187650680541992, 1548657065.7516673], "v": 0.1}
13 | {"i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 480, 7, 7], "float32"], ["TENSOR", [480, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 480, 7, 7, "float32"], [480, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"i": 23621, "e": [["tile_f", "sp", [80, 2, 1, 3]], ["tile_y", "sp", [1, 1, 7, 1]], ["tile_x", "sp", [1, 1, 7, 1]], ["auto_unroll_max_step", "ot", 256], ["unroll_explicit", "ot", 0]], "t": "direct", "c": null}], "r": [[3.796775e-05], 0, 2.5406243801116943, 1548658554.4247355], "v": 0.1}
14 | {"i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 328, 7, 7], "float32"], ["TENSOR", [480, 328, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 328, 7, 7, "float32"], [480, 328, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"i": 51212, "e": [["tile_f", "sp", [8, 3, 20, 1]], ["tile_y", "sp", [1, 7, 1, 1]], ["tile_x", "sp", [1, 1, 7, 1]], ["tile_rc", "sp", [41, 8]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 0], ["unroll_explicit", "ot", 0]], "t": "direct", "c": null}], "r": [[0.000105911], 0, 1.2114758491516113, 1548659160.2813828], "v": 0.1}
15 | {"i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 328, 14, 14], "float32"], ["TENSOR", [328, 1, 3, 3], "float32"], [2, 2], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 328, 14, 14, "float32"], [328, 1, 3, 3, "float32"], [2, 2], [1, 1], "float32"], {"i": 7218, "e": [["tile_f", "sp", [41, 1, 8, 1]], ["tile_y", "sp", [1, 1, 7, 1]], ["tile_x", "sp", [1, 1, 7, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null}], "r": [[3.831175e-05], 0, 1.9223744869232178, 1548659793.643159], "v": 0.1}
16 | {"i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 296, 14, 14], "float32"], ["TENSOR", [328, 296, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 296, 14, 14, "float32"], [328, 296, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"i": 69322, "e": [["tile_f", "sp", [4, 1, 41, 2]], ["tile_y", "sp", [2, 7, 1, 1]], ["tile_x", "sp", [2, 1, 7, 1]], ["tile_rc", "sp", [37, 8]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 0], ["unroll_explicit", "ot", 0]], "t": "direct", "c": null}], "r": [[0.000235302], 0, 2.302255630493164, 1548660675.0394955], "v": 0.1}
17 | {"i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 296, 14, 14], "float32"], ["TENSOR", [296, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 296, 14, 14, "float32"], [296, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"i": 112248, "e": [["tile_f", "sp", [148, 1, 2, 1]], ["tile_y", "sp", [1, 1, 7, 2]], ["tile_x", "sp", [1, 2, 7, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null}], "r": [[3.78955e-05], 0, 1.5320491790771484, 1548661622.8226697], "v": 0.1}
18 | {"i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 288, 14, 14], "float32"], ["TENSOR", [296, 288, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 288, 14, 14, "float32"], [296, 288, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"i": 130741, "e": [["tile_f", "sp", [4, 2, 37, 1]], ["tile_y", "sp", [2, 7, 1, 1]], ["tile_x", "sp", [2, 1, 7, 1]], ["tile_rc", "sp", [32, 9]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 0], ["unroll_explicit", "ot", 0]], "t": "direct", "c": null}], "r": [[0.00021435875], 0, 1.6423635482788086, 1548662485.159752], "v": 0.1}
19 | {"i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 288, 14, 14], "float32"], ["TENSOR", [288, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 288, 14, 14, "float32"], [288, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"i": 676499, "e": [["tile_f", "sp", [72, 2, 2, 1]], ["tile_y", "sp", [1, 1, 14, 1]], ["tile_x", "sp", [1, 1, 7, 2]], ["auto_unroll_max_step", "ot", 256], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null}], "r": [[3.786375e-05], 0, 1.390625238418579, 1548663463.9980013], "v": 0.1}
20 | {"i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 272, 14, 14], "float32"], ["TENSOR", [288, 272, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 272, 14, 14, "float32"], [288, 272, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"i": 4805431, "e": [["tile_f", "sp", [6, 6, 8, 1]], ["tile_y", "sp", [1, 7, 2, 1]], ["tile_x", "sp", [1, 1, 14, 1]], ["tile_rc", "sp", [34, 8]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 0], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null}], "r": [[0.00011731875], 0, 2.9426913261413574, 1548664977.3522089], "v": 0.1}
21 | {"i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 272, 14, 14], "float32"], ["TENSOR", [272, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 272, 14, 14, "float32"], [272, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"i": 160044, "e": [["tile_f", "sp", [34, 1, 8, 1]], ["tile_y", "sp", [1, 2, 7, 1]], ["tile_x", "sp", [1, 2, 7, 1]], ["auto_unroll_max_step", "ot", 256], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null}], "r": [[3.8048e-05], 0, 1.3805427551269531, 1548666124.2337215], "v": 0.1}
22 | {"i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 376, 14, 14], "float32"], ["TENSOR", [272, 376, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 376, 14, 14, "float32"], [272, 376, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"i": 986321, "e": [["tile_f", "sp", [4, 17, 4, 1]], ["tile_y", "sp", [1, 7, 2, 1]], ["tile_x", "sp", [1, 1, 14, 1]], ["tile_rc", "sp", [47, 8]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 0], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null}], "r": [[0.00017648625], 0, 2.901970386505127, 1548666802.045066], "v": 0.1}
23 | {"i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 376, 14, 14], "float32"], ["TENSOR", [376, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 376, 14, 14, "float32"], [376, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"i": 31608, "e": [["tile_f", "sp", [188, 1, 2, 1]], ["tile_y", "sp", [1, 1, 7, 2]], ["tile_x", "sp", [1, 1, 14, 1]], ["auto_unroll_max_step", "ot", 256], ["unroll_explicit", "ot", 0]], "t": "direct", "c": null}], "r": [[5.722375e-05], 0, 1.8894000053405762, 1548667848.8311331], "v": 0.1}
24 | {"i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 408, 14, 14], "float32"], ["TENSOR", [376, 408, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 408, 14, 14, "float32"], [376, 408, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"i": 1399221, "e": [["tile_f", "sp", [4, 2, 47, 1]], ["tile_y", "sp", [2, 7, 1, 1]], ["tile_x", "sp", [1, 7, 2, 1]], ["tile_rc", "sp", [68, 6]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 512], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null}], "r": [[0.000361302], 0, 1.5910117626190186, 1548668723.9305615], "v": 0.1}
25 | {"i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 408, 14, 14], "float32"], ["TENSOR", [408, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 408, 14, 14, "float32"], [408, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"i": 371374, "e": [["tile_f", "sp", [102, 1, 1, 4]], ["tile_y", "sp", [1, 1, 14, 1]], ["tile_x", "sp", [1, 1, 14, 1]], ["auto_unroll_max_step", "ot", 256], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null}], "r": [[5.76475e-05], 0, 2.138455390930176, 1548669478.0002713], "v": 0.1}
26 | {"i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 256, 14, 14], "float32"], ["TENSOR", [408, 256, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 256, 14, 14, "float32"], [408, 256, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"i": 3237695, "e": [["tile_f", "sp", [6, 1, 4, 17]], ["tile_y", "sp", [1, 7, 2, 1]], ["tile_x", "sp", [1, 1, 14, 1]], ["tile_rc", "sp", [32, 8]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 512], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null}], "r": [[0.00014766275], 0, 1.5402188301086426, 1548670679.6535883], "v": 0.1}
27 | {"i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 256, 28, 28], "float32"], ["TENSOR", [256, 1, 3, 3], "float32"], [2, 2], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 256, 28, 28, "float32"], [256, 1, 3, 3, "float32"], [2, 2], [1, 1], "float32"], {"i": 233311, "e": [["tile_f", "sp", [128, 2, 1, 1]], ["tile_y", "sp", [2, 1, 7, 1]], ["tile_x", "sp", [1, 1, 14, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null}], "r": [[7.772775e-05], 0, 2.0413858890533447, 1548671505.4201894], "v": 0.1}
28 | {"i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 144, 28, 28], "float32"], ["TENSOR", [256, 144, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 144, 28, 28, "float32"], [256, 144, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"i": 1432721, "e": [["tile_f", "sp", [8, 4, 8, 1]], ["tile_y", "sp", [4, 7, 1, 1]], ["tile_x", "sp", [1, 1, 28, 1]], ["tile_rc", "sp", [18, 8]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 0], ["unroll_explicit", "ot", 0]], "t": "direct", "c": null}], "r": [[0.000216294], 0, 1.6201062202453613, 1548672455.3859437], "v": 0.1}
29 | {"i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 144, 28, 28], "float32"], ["TENSOR", [144, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 144, 28, 28, "float32"], [144, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"i": 3046490, "e": [["tile_f", "sp", [72, 1, 1, 2]], ["tile_y", "sp", [2, 1, 7, 2]], ["tile_x", "sp", [1, 1, 28, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null}], "r": [[5.779175e-05], 0, 2.2191083431243896, 1548673391.4412506], "v": 0.1}
30 | {"i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 120, 28, 28], "float32"], ["TENSOR", [144, 120, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 120, 28, 28, "float32"], [144, 120, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"i": 3047741, "e": [["tile_f", "sp", [2, 1, 8, 9]], ["tile_y", "sp", [7, 1, 1, 4]], ["tile_x", "sp", [1, 1, 28, 1]], ["tile_rc", "sp", [20, 6]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 0], ["unroll_explicit", "ot", 0]], "t": "direct", "c": null}], "r": [[0.0001163435], 0, 1.2421369552612305, 1548675456.083652], "v": 0.1}
31 | {"i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 120, 56, 56], "float32"], ["TENSOR", [120, 1, 3, 3], "float32"], [2, 2], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 120, 56, 56, "float32"], [120, 1, 3, 3, "float32"], [2, 2], [1, 1], "float32"], {"i": 2269760, "e": [["tile_f", "sp", [120, 1, 1, 1]], ["tile_y", "sp", [2, 2, 7, 1]], ["tile_x", "sp", [1, 1, 28, 1]], ["auto_unroll_max_step", "ot", 256], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null}], "r": [[0.00011662325], 0, 1.597099781036377, 1548676329.4550035], "v": 0.1}
32 | {"i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 88, 56, 56], "float32"], ["TENSOR", [120, 88, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 88, 56, 56, "float32"], [120, 88, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"i": 53975012, "e": [["tile_f", "sp", [2, 1, 4, 15]], ["tile_y", "sp", [14, 2, 1, 2]], ["tile_x", "sp", [1, 2, 28, 1]], ["tile_rc", "sp", [22, 4]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 0], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null}], "r": [[0.00027512525], 0, 1.369657278060913, 1548677300.418834], "v": 0.1}
33 | {"i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 88, 56, 56], "float32"], ["TENSOR", [88, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 88, 56, 56, "float32"], [88, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"i": 2750720, "e": [["tile_f", "sp", [88, 1, 1, 1]], ["tile_y", "sp", [2, 1, 4, 7]], ["tile_x", "sp", [1, 1, 56, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null}], "r": [[9.8191e-05], 0, 1.5733342170715332, 1548678185.3843915], "v": 0.1}
34 | {"i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 56, 56, 56], "float32"], ["TENSOR", [88, 56, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 56, 56, 56, "float32"], [88, 56, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"i": 13501456, "e": [["tile_f", "sp", [2, 11, 4, 1]], ["tile_y", "sp", [14, 1, 1, 4]], ["tile_x", "sp", [1, 1, 56, 1]], ["tile_rc", "sp", [14, 4]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 0], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null}], "r": [[0.0001444785], 0, 1.7148137092590332, 1548679401.0876276], "v": 0.1}
35 | {"i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 56, 112, 112], "float32"], ["TENSOR", [56, 1, 3, 3], "float32"], [2, 2], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 56, 112, 112, "float32"], [56, 1, 3, 3, "float32"], [2, 2], [1, 1], "float32"], {"i": 2748800, "e": [["tile_f", "sp", [56, 1, 1, 1]], ["tile_y", "sp", [7, 1, 4, 2]], ["tile_x", "sp", [1, 1, 56, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null}], "r": [[0.0001829665], 0, 1.6487436294555664, 1548680224.10838], "v": 0.1}
36 | {"i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 16, 112, 112], "float32"], ["TENSOR", [56, 16, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 16, 112, 112, "float32"], [56, 16, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"i": 28698016, "e": [["tile_f", "sp", [2, 7, 4, 1]], ["tile_y", "sp", [56, 1, 1, 2]], ["tile_x", "sp", [2, 1, 56, 1]], ["tile_rc", "sp", [2, 8]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 0], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null}], "r": [[0.000139111], 0, 1.4088413715362549, 1548681204.6446605], "v": 0.1}
37 | {"i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 16, 112, 112], "float32"], ["TENSOR", [16, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 16, 112, 112, "float32"], [16, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"i": 3638775, "e": [["tile_f", "sp", [16, 1, 1, 1]], ["tile_y", "sp", [7, 1, 4, 4]], ["tile_x", "sp", [2, 1, 56, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null}], "r": [[7.75195e-05], 0, 1.5174787044525146, 1548682004.0559516], "v": 0.1}
38 | {"i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 3, 224, 224], "float32"], ["TENSOR", [16, 3, 3, 3], "float32"], [2, 2], [1, 1], "NCHW", "float32"], {}, ["conv2d", [1, 3, 224, 224, "float32"], [16, 3, 3, 3, "float32"], [2, 2], [1, 1], "NCHW", "float32"], {"i": 20787688, "e": [["tile_f", "sp", [1, 1, 2, 8]], ["tile_y", "sp", [28, 1, 2, 2]], ["tile_x", "sp", [2, 1, 56, 1]], ["tile_rc", "sp", [3, 1]], ["tile_ry", "sp", [1, 3]], ["tile_rx", "sp", [1, 3]], ["auto_unroll_max_step", "ot", 0], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null}], "r": [[0.00013608675], 0, 3.140852451324463, 1548683091.1406922], "v": 0.1}
39 | 


--------------------------------------------------------------------------------
/tvm_compile/tuning/tx2-gpu.mobilenet-nnconv5dw-skipadd.trials=2000.stop=600.log:
--------------------------------------------------------------------------------
 1 | {"v": 0.1, "r": [[0.00019598275], 0, 1.5730626583099365, 1548562540.6536844], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 32, 224, 224], "float32"], ["TENSOR", [1, 32, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 32, 224, 224, "float32"], [1, 32, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"e": [["tile_f", "sp", [1, 1, 1, 1]], ["tile_y", "sp", [28, 2, 4, 1]], ["tile_x", "sp", [1, 1, 224, 1]], ["tile_rc", "sp", [32, 1]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 0], ["unroll_explicit", "ot", 0]], "t": "direct", "c": null, "i": 13911}]}
 2 | {"v": 0.1, "r": [[0.0002421905], 0, 3.849091053009033, 1548564960.4150162], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 64, 112, 112], "float32"], ["TENSOR", [32, 64, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 64, 112, 112, "float32"], [32, 64, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"e": [["tile_f", "sp", [1, 1, 8, 4]], ["tile_y", "sp", [56, 1, 1, 2]], ["tile_x", "sp", [1, 7, 16, 1]], ["tile_rc", "sp", [16, 4]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 0], ["unroll_explicit", "ot", 0]], "t": "direct", "c": null, "i": 2495685}]}
 3 | {"v": 0.1, "r": [[0.00033507775], 0, 1.67569899559021, 1548565783.6229427], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 64, 112, 112], "float32"], ["TENSOR", [64, 1, 5, 5], "float32"], [1, 1], [2, 2], "float32"], {}, ["depthwise_conv2d_nchw", [1, 64, 112, 112, "float32"], [64, 1, 5, 5, "float32"], [1, 1], [2, 2], "float32"], {"e": [["tile_f", "sp", [64, 1, 1, 1]], ["tile_y", "sp", [2, 1, 14, 4]], ["tile_x", "sp", [7, 1, 16, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 8674680}]}
 4 | {"v": 0.1, "r": [[0.00019454275], 0, 1.5050935745239258, 1548566644.4516716], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 128, 56, 56], "float32"], ["TENSOR", [64, 128, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 128, 56, 56, "float32"], [64, 128, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"e": [["tile_f", "sp", [2, 4, 8, 1]], ["tile_y", "sp", [28, 1, 1, 2]], ["tile_x", "sp", [1, 7, 8, 1]], ["tile_rc", "sp", [32, 4]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 512], ["unroll_explicit", "ot", 0]], "t": "direct", "c": null, "i": 5533100}]}
 5 | {"v": 0.1, "r": [[0.00019157425], 0, 3.138878345489502, 1548567436.7737145], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 128, 56, 56], "float32"], ["TENSOR", [128, 1, 5, 5], "float32"], [1, 1], [2, 2], "float32"], {}, ["depthwise_conv2d_nchw", [1, 128, 56, 56, "float32"], [128, 1, 5, 5, "float32"], [1, 1], [2, 2], "float32"], {"e": [["tile_f", "sp", [128, 1, 1, 1]], ["tile_y", "sp", [2, 2, 7, 2]], ["tile_x", "sp", [1, 1, 28, 2]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 4296360}]}
 6 | {"v": 0.1, "r": [[0.0001684785], 0, 1.6706435680389404, 1548568288.6996257], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 256, 28, 28], "float32"], ["TENSOR", [128, 256, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 256, 28, 28, "float32"], [128, 256, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"e": [["tile_f", "sp", [2, 1, 8, 8]], ["tile_y", "sp", [7, 1, 4, 1]], ["tile_x", "sp", [1, 7, 4, 1]], ["tile_rc", "sp", [32, 8]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 512], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 7542097}]}
 7 | {"v": 0.1, "r": [[0.000109935], 0, 1.4281127452850342, 1548568975.349207], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 256, 28, 28], "float32"], ["TENSOR", [256, 1, 5, 5], "float32"], [1, 1], [2, 2], "float32"], {}, ["depthwise_conv2d_nchw", [1, 256, 28, 28, "float32"], [256, 1, 5, 5, "float32"], [1, 1], [2, 2], "float32"], {"e": [["tile_f", "sp", [256, 1, 1, 1]], ["tile_y", "sp", [2, 1, 2, 7]], ["tile_x", "sp", [1, 1, 28, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 1437645}]}
 8 | {"v": 0.1, "r": [[0.000167879], 0, 1.5164942741394043, 1548570086.3063977], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 512, 14, 14], "float32"], ["TENSOR", [256, 512, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 512, 14, 14, "float32"], [256, 512, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"e": [["tile_f", "sp", [4, 8, 8, 1]], ["tile_y", "sp", [2, 7, 1, 1]], ["tile_x", "sp", [1, 1, 14, 1]], ["tile_rc", "sp", [64, 8]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 512], ["unroll_explicit", "ot", 0]], "t": "direct", "c": null, "i": 570597}]}
 9 | {"v": 0.1, "r": [[7.83035e-05], 0, 3.1352176666259766, 1548570831.7553627], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 512, 14, 14], "float32"], ["TENSOR", [512, 1, 5, 5], "float32"], [1, 1], [2, 2], "float32"], {}, ["depthwise_conv2d_nchw", [1, 512, 14, 14, "float32"], [512, 1, 5, 5, "float32"], [1, 1], [2, 2], "float32"], {"e": [["tile_f", "sp", [512, 1, 1, 1]], ["tile_y", "sp", [1, 1, 7, 2]], ["tile_x", "sp", [1, 2, 7, 1]], ["auto_unroll_max_step", "ot", 256], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 252340}]}
10 | {"v": 0.1, "r": [[0.0002710375], 0, 2.723203420639038, 1548571856.4664495], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 1024, 7, 7], "float32"], ["TENSOR", [512, 1024, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 1024, 7, 7, "float32"], [512, 1024, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"e": [["tile_f", "sp", [8, 4, 16, 1]], ["tile_y", "sp", [1, 1, 7, 1]], ["tile_x", "sp", [1, 7, 1, 1]], ["tile_rc", "sp", [64, 16]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 209036}]}
11 | {"v": 0.1, "r": [[7.43915e-05], 0, 1.9873511791229248, 1548572384.122891], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 1024, 7, 7], "float32"], ["TENSOR", [1024, 1, 5, 5], "float32"], [1, 1], [2, 2], "float32"], {}, ["depthwise_conv2d_nchw", [1, 1024, 7, 7, "float32"], [1024, 1, 5, 5, "float32"], [1, 1], [2, 2], "float32"], {"e": [["tile_f", "sp", [256, 1, 1, 4]], ["tile_y", "sp", [1, 1, 7, 1]], ["tile_x", "sp", [1, 1, 7, 1]], ["auto_unroll_max_step", "ot", 256], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 21285}]}
12 | {"v": 0.1, "r": [[0.00056902825], 0, 3.107520818710327, 1548573097.8638668], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 1024, 7, 7], "float32"], ["TENSOR", [1024, 1024, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 1024, 7, 7, "float32"], [1024, 1024, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"e": [["tile_f", "sp", [16, 2, 32, 1]], ["tile_y", "sp", [1, 1, 1, 7]], ["tile_x", "sp", [1, 1, 7, 1]], ["tile_rc", "sp", [64, 16]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 0], ["unroll_explicit", "ot", 0]], "t": "direct", "c": null, "i": 21496}]}
13 | {"v": 0.1, "r": [[4.303975e-05], 0, 1.3151054382324219, 1548573702.4212985], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 1024, 7, 7], "float32"], ["TENSOR", [1024, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 1024, 7, 7, "float32"], [1024, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"e": [["tile_f", "sp", [128, 1, 8, 1]], ["tile_y", "sp", [1, 1, 1, 7]], ["tile_x", "sp", [1, 1, 7, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 26056}]}
14 | {"v": 0.1, "r": [[0.000293478], 0, 1.6003823280334473, 1548574300.4192855], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 512, 7, 7], "float32"], ["TENSOR", [1024, 512, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 512, 7, 7, "float32"], [1024, 512, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"e": [["tile_f", "sp", [32, 2, 8, 2]], ["tile_y", "sp", [1, 7, 1, 1]], ["tile_x", "sp", [1, 1, 7, 1]], ["tile_rc", "sp", [64, 8]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 0], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 153676}]}
15 | {"v": 0.1, "r": [[5.61915e-05], 0, 1.6853110790252686, 1548574886.3446763], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 512, 14, 14], "float32"], ["TENSOR", [512, 1, 3, 3], "float32"], [2, 2], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 512, 14, 14, "float32"], [512, 1, 3, 3, "float32"], [2, 2], [1, 1], "float32"], {"e": [["tile_f", "sp", [32, 1, 8, 2]], ["tile_y", "sp", [1, 1, 7, 1]], ["tile_x", "sp", [1, 1, 7, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 19879}]}
16 | {"v": 0.1, "r": [[0.000312902], 0, 2.8101165294647217, 1548575397.150415], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 512, 14, 14], "float32"], ["TENSOR", [512, 512, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 512, 14, 14, "float32"], [512, 512, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"e": [["tile_f", "sp", [8, 4, 8, 2]], ["tile_y", "sp", [2, 1, 1, 7]], ["tile_x", "sp", [1, 1, 14, 1]], ["tile_rc", "sp", [64, 8]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 0]], "t": "direct", "c": null, "i": 1326241}]}
17 | {"v": 0.1, "r": [[5.80155e-05], 0, 2.5316574573516846, 1548576505.3454807], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 512, 14, 14], "float32"], ["TENSOR", [512, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 512, 14, 14, "float32"], [512, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"e": [["tile_f", "sp", [128, 1, 4, 1]], ["tile_y", "sp", [1, 7, 2, 1]], ["tile_x", "sp", [1, 1, 14, 1]], ["auto_unroll_max_step", "ot", 0], ["unroll_explicit", "ot", 0]], "t": "direct", "c": null, "i": 29279}]}
18 | {"v": 0.1, "r": [[0.00017500625], 0, 3.4283621311187744, 1548577207.0975342], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 256, 14, 14], "float32"], ["TENSOR", [512, 256, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 256, 14, 14, "float32"], [512, 256, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"e": [["tile_f", "sp", [8, 4, 8, 2]], ["tile_y", "sp", [2, 7, 1, 1]], ["tile_x", "sp", [1, 1, 14, 1]], ["tile_rc", "sp", [32, 8]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 512], ["unroll_explicit", "ot", 0]], "t": "direct", "c": null, "i": 704521}]}
19 | {"v": 0.1, "r": [[7.74235e-05], 0, 1.7664618492126465, 1548578059.412158], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 256, 28, 28], "float32"], ["TENSOR", [256, 1, 3, 3], "float32"], [2, 2], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 256, 28, 28, "float32"], [256, 1, 3, 3, "float32"], [2, 2], [1, 1], "float32"], {"e": [["tile_f", "sp", [32, 1, 4, 2]], ["tile_y", "sp", [2, 1, 7, 1]], ["tile_x", "sp", [1, 1, 14, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 233370}]}
20 | {"v": 0.1, "r": [[0.000316189], 0, 1.5466210842132568, 1548578728.0832877], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 256, 28, 28], "float32"], ["TENSOR", [256, 256, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 256, 28, 28, "float32"], [256, 256, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"e": [["tile_f", "sp", [4, 1, 8, 8]], ["tile_y", "sp", [7, 1, 4, 1]], ["tile_x", "sp", [1, 7, 4, 1]], ["tile_rc", "sp", [32, 8]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 12746374}]}
21 | {"v": 0.1, "r": [[7.8927e-05], 0, 1.5628066062927246, 1548579523.0229578], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 256, 28, 28], "float32"], ["TENSOR", [256, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 256, 28, 28, "float32"], [256, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"e": [["tile_f", "sp", [256, 1, 1, 1]], ["tile_y", "sp", [1, 1, 4, 7]], ["tile_x", "sp", [1, 1, 28, 1]], ["auto_unroll_max_step", "ot", 256], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 1173975}]}
22 | {"v": 0.1, "r": [[0.00019647875], 0, 2.4384193420410156, 1548580509.127791], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 128, 28, 28], "float32"], ["TENSOR", [256, 128, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 128, 28, 28, "float32"], [256, 128, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"e": [["tile_f", "sp", [4, 1, 8, 8]], ["tile_y", "sp", [7, 1, 4, 1]], ["tile_x", "sp", [1, 7, 4, 1]], ["tile_rc", "sp", [16, 8]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 512], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 9314374}]}
23 | {"v": 0.1, "r": [[0.00011714325], 0, 1.3188731670379639, 1548581309.9010916], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 128, 56, 56], "float32"], ["TENSOR", [128, 1, 3, 3], "float32"], [2, 2], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 128, 56, 56, "float32"], [128, 1, 3, 3, "float32"], [2, 2], [1, 1], "float32"], {"e": [["tile_f", "sp", [128, 1, 1, 1]], ["tile_y", "sp", [2, 1, 14, 1]], ["tile_x", "sp", [1, 2, 14, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 1038600}]}
24 | {"v": 0.1, "r": [[0.00031403], 0, 1.9832639694213867, 1548581896.1469371], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 128, 56, 56], "float32"], ["TENSOR", [128, 128, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 128, 56, 56, "float32"], [128, 128, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"e": [["tile_f", "sp", [2, 2, 8, 4]], ["tile_y", "sp", [28, 1, 2, 1]], ["tile_x", "sp", [1, 7, 8, 1]], ["tile_rc", "sp", [16, 8]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 512], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 27101840}]}
25 | {"v": 0.1, "r": [[0.000137879], 0, 1.4539649486541748, 1548582835.3741553], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 128, 56, 56], "float32"], ["TENSOR", [128, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 128, 56, 56, "float32"], [128, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"e": [["tile_f", "sp", [128, 1, 1, 1]], ["tile_y", "sp", [2, 1, 4, 7]], ["tile_x", "sp", [1, 1, 56, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 4126080}]}
26 | {"v": 0.1, "r": [[0.00020232675], 0, 2.3310959339141846, 1548584787.231282], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 64, 56, 56], "float32"], ["TENSOR", [128, 64, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 64, 56, 56, "float32"], [128, 64, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"e": [["tile_f", "sp", [2, 1, 8, 8]], ["tile_y", "sp", [28, 1, 2, 1]], ["tile_x", "sp", [1, 7, 8, 1]], ["tile_rc", "sp", [8, 8]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 0], ["unroll_explicit", "ot", 0]], "t": "direct", "c": null, "i": 2525857}]}
27 | {"v": 0.1, "r": [[0.000208358], 0, 2.127153158187866, 1548585621.7478325], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 64, 112, 112], "float32"], ["TENSOR", [64, 1, 3, 3], "float32"], [2, 2], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 64, 112, 112, "float32"], [64, 1, 3, 3, "float32"], [2, 2], [1, 1], "float32"], {"e": [["tile_f", "sp", [64, 1, 1, 1]], ["tile_y", "sp", [7, 2, 4, 1]], ["tile_x", "sp", [1, 1, 56, 1]], ["auto_unroll_max_step", "ot", 256], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 2346540}]}
28 | {"v": 0.1, "r": [[0.00021625475], 0, 2.464158296585083, 1548587361.9766982], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 32, 112, 112], "float32"], ["TENSOR", [64, 32, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 32, 112, 112, "float32"], [64, 32, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"e": [["tile_f", "sp", [2, 2, 8, 2]], ["tile_y", "sp", [28, 1, 1, 4]], ["tile_x", "sp", [7, 1, 16, 1]], ["tile_rc", "sp", [8, 4]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 0], ["unroll_explicit", "ot", 0]], "t": "direct", "c": null, "i": 3734264}]}
29 | {"v": 0.1, "r": [[0.00013695125], 0, 2.6825642585754395, 1548588199.2572803], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 32, 112, 112], "float32"], ["TENSOR", [32, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 32, 112, 112, "float32"], [32, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"e": [["tile_f", "sp", [32, 1, 1, 1]], ["tile_y", "sp", [4, 1, 4, 7]], ["tile_x", "sp", [2, 1, 56, 1]], ["auto_unroll_max_step", "ot", 256], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 4725392}]}
30 | {"v": 0.1, "r": [[0.000175575], 0, 1.2250316143035889, 1548589605.8968496], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 3, 224, 224], "float32"], ["TENSOR", [32, 3, 3, 3], "float32"], [2, 2], [1, 1], "NCHW", "float32"], {}, ["conv2d", [1, 3, 224, 224, "float32"], [32, 3, 3, 3, "float32"], [2, 2], [1, 1], "NCHW", "float32"], {"e": [["tile_f", "sp", [1, 2, 8, 2]], ["tile_y", "sp", [28, 4, 1, 1]], ["tile_x", "sp", [7, 1, 16, 1]], ["tile_rc", "sp", [3, 1]], ["tile_ry", "sp", [1, 3]], ["tile_rx", "sp", [1, 3]], ["auto_unroll_max_step", "ot", 0], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 33218226}]}
31 | 


--------------------------------------------------------------------------------
/tvm_compile/tuning/tx2-gpu.mobilenet-nnconv5dw.trials=2000.stop=600.log:
--------------------------------------------------------------------------------
 1 | {"v": 0.1, "r": [[0.00019565525], 0, 2.689757823944092, 1549429087.5617292], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 32, 224, 224], "float32"], ["TENSOR", [1, 32, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 32, 224, 224, "float32"], [1, 32, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"e": [["tile_f", "sp", [1, 1, 1, 1]], ["tile_y", "sp", [56, 4, 1, 1]], ["tile_x", "sp", [1, 1, 224, 1]], ["tile_rc", "sp", [32, 1]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 512], ["unroll_explicit", "ot", 0]], "t": "direct", "c": null, "i": 314946}]}
 2 | {"v": 0.1, "r": [[0.0003639425], 0, 4.725115060806274, 1549430548.7858527], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 64, 112, 112], "float32"], ["TENSOR", [32, 64, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 64, 112, 112, "float32"], [32, 64, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"e": [["tile_f", "sp", [2, 4, 1, 4]], ["tile_y", "sp", [112, 1, 1, 1]], ["tile_x", "sp", [1, 1, 112, 1]], ["tile_rc", "sp", [32, 2]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 0]], "t": "direct", "c": null, "i": 16808998}]}
 3 | {"v": 0.1, "r": [[0.00033395075], 0, 2.3271985054016113, 1549431980.0935845], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 64, 112, 112], "float32"], ["TENSOR", [64, 1, 5, 5], "float32"], [1, 1], [2, 2], "float32"], {}, ["depthwise_conv2d_nchw", [1, 64, 112, 112, "float32"], [64, 1, 5, 5, "float32"], [1, 1], [2, 2], "float32"], {"e": [["tile_f", "sp", [64, 1, 1, 1]], ["tile_y", "sp", [2, 1, 14, 4]], ["tile_x", "sp", [7, 1, 16, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 8674680}]}
 4 | {"v": 0.1, "r": [[0.0002864375], 0, 2.3175532817840576, 1549433087.935274], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 128, 56, 56], "float32"], ["TENSOR", [64, 128, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 128, 56, 56, "float32"], [64, 128, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"e": [["tile_f", "sp", [2, 32, 1, 1]], ["tile_y", "sp", [56, 1, 1, 1]], ["tile_x", "sp", [1, 2, 28, 1]], ["tile_rc", "sp", [32, 4]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 0]], "t": "direct", "c": null, "i": 9864965}]}
 5 | {"v": 0.1, "r": [[0.00019396625], 0, 1.462557315826416, 1549433838.8822808], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 128, 56, 56], "float32"], ["TENSOR", [128, 1, 5, 5], "float32"], [1, 1], [2, 2], "float32"], {}, ["depthwise_conv2d_nchw", [1, 128, 56, 56, "float32"], [128, 1, 5, 5, "float32"], [1, 1], [2, 2], "float32"], {"e": [["tile_f", "sp", [128, 1, 1, 1]], ["tile_y", "sp", [2, 2, 7, 2]], ["tile_x", "sp", [1, 1, 28, 2]], ["auto_unroll_max_step", "ot", 256], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 3528360}]}
 6 | {"v": 0.1, "r": [[0.000215031], 0, 1.3755156993865967, 1549435352.1497386], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 256, 28, 28], "float32"], ["TENSOR", [128, 256, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 256, 28, 28, "float32"], [128, 256, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"e": [["tile_f", "sp", [2, 4, 16, 1]], ["tile_y", "sp", [14, 1, 1, 2]], ["tile_x", "sp", [1, 2, 14, 1]], ["tile_rc", "sp", [16, 16]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 0], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 6030988}]}
 7 | {"v": 0.1, "r": [[0.0001163515], 0, 3.0618081092834473, 1549436102.657577], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 256, 28, 28], "float32"], ["TENSOR", [256, 1, 5, 5], "float32"], [1, 1], [2, 2], "float32"], {}, ["depthwise_conv2d_nchw", [1, 256, 28, 28, "float32"], [256, 1, 5, 5, "float32"], [1, 1], [2, 2], "float32"], {"e": [["tile_f", "sp", [256, 1, 1, 1]], ["tile_y", "sp", [2, 1, 2, 7]], ["tile_x", "sp", [1, 1, 28, 1]], ["auto_unroll_max_step", "ot", 256], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 1173645}]}
 8 | {"v": 0.1, "r": [[0.00017022325], 0, 3.1339035034179688, 1549437395.9139726], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 512, 14, 14], "float32"], ["TENSOR", [256, 512, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 512, 14, 14, "float32"], [256, 512, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"e": [["tile_f", "sp", [4, 4, 16, 1]], ["tile_y", "sp", [2, 7, 1, 1]], ["tile_x", "sp", [1, 1, 14, 1]], ["tile_rc", "sp", [32, 16]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 512], ["unroll_explicit", "ot", 0]], "t": "direct", "c": null, "i": 612842}]}
 9 | {"v": 0.1, "r": [[7.806325e-05], 0, 1.9228358268737793, 1549438115.9921758], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 512, 14, 14], "float32"], ["TENSOR", [512, 1, 5, 5], "float32"], [1, 1], [2, 2], "float32"], {}, ["depthwise_conv2d_nchw", [1, 512, 14, 14, "float32"], [512, 1, 5, 5, "float32"], [1, 1], [2, 2], "float32"], {"e": [["tile_f", "sp", [512, 1, 1, 1]], ["tile_y", "sp", [1, 2, 7, 1]], ["tile_x", "sp", [1, 1, 7, 2]], ["auto_unroll_max_step", "ot", 256], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 265540}]}
10 | {"v": 0.1, "r": [[0.00027059775], 0, 1.8419475555419922, 1549439012.3243341], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 1024, 7, 7], "float32"], ["TENSOR", [512, 1024, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 1024, 7, 7, "float32"], [512, 1024, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"e": [["tile_f", "sp", [8, 4, 16, 1]], ["tile_y", "sp", [1, 1, 1, 7]], ["tile_x", "sp", [1, 1, 7, 1]], ["tile_rc", "sp", [64, 16]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 210136}]}
11 | {"v": 0.1, "r": [[7.395175e-05], 0, 1.3589377403259277, 1549439569.3904214], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 1024, 7, 7], "float32"], ["TENSOR", [1024, 1, 5, 5], "float32"], [1, 1], [2, 2], "float32"], {}, ["depthwise_conv2d_nchw", [1, 1024, 7, 7, "float32"], [1024, 1, 5, 5, "float32"], [1, 1], [2, 2], "float32"], {"e": [["tile_f", "sp", [128, 1, 4, 2]], ["tile_y", "sp", [1, 1, 7, 1]], ["tile_x", "sp", [1, 1, 7, 1]], ["auto_unroll_max_step", "ot", 256], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 21249}]}
12 | {"v": 0.1, "r": [[0.000567654], 0, 2.060800075531006, 1549440270.1668124], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 1024, 7, 7], "float32"], ["TENSOR", [1024, 1024, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 1024, 7, 7, "float32"], [1024, 1024, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"e": [["tile_f", "sp", [16, 2, 32, 1]], ["tile_y", "sp", [1, 1, 1, 7]], ["tile_x", "sp", [1, 1, 7, 1]], ["tile_rc", "sp", [64, 16]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 0], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 172504}]}
13 | {"v": 0.1, "r": [[4.21195e-05], 0, 1.4836952686309814, 1549440881.346037], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 1024, 7, 7], "float32"], ["TENSOR", [1024, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 1024, 7, 7, "float32"], [1024, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"e": [["tile_f", "sp", [256, 1, 4, 1]], ["tile_y", "sp", [1, 7, 1, 1]], ["tile_x", "sp", [1, 1, 7, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 25475}]}
14 | {"v": 0.1, "r": [[0.00029675725], 0, 2.273451805114746, 1549441441.2646382], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 512, 7, 7], "float32"], ["TENSOR", [1024, 512, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 512, 7, 7, "float32"], [1024, 512, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"e": [["tile_f", "sp", [32, 2, 8, 2]], ["tile_y", "sp", [1, 1, 1, 7]], ["tile_x", "sp", [1, 1, 7, 1]], ["tile_rc", "sp", [64, 8]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 0], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 154248}]}
15 | {"v": 0.1, "r": [[5.194375e-05], 0, 1.7700073719024658, 1549442009.8341005], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 512, 14, 14], "float32"], ["TENSOR", [512, 1, 3, 3], "float32"], [2, 2], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 512, 14, 14, "float32"], [512, 1, 3, 3, "float32"], [2, 2], [1, 1], "float32"], {"e": [["tile_f", "sp", [32, 1, 8, 2]], ["tile_y", "sp", [1, 1, 7, 1]], ["tile_x", "sp", [1, 1, 7, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 19879}]}
16 | {"v": 0.1, "r": [[0.00031262125], 0, 1.2954254150390625, 1549442532.348055], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 512, 14, 14], "float32"], ["TENSOR", [512, 512, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 512, 14, 14, "float32"], [512, 512, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"e": [["tile_f", "sp", [8, 4, 8, 2]], ["tile_y", "sp", [2, 7, 1, 1]], ["tile_x", "sp", [1, 1, 14, 1]], ["tile_rc", "sp", [64, 8]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 0]], "t": "direct", "c": null, "i": 1324041}]}
17 | {"v": 0.1, "r": [[5.823175e-05], 0, 1.1607718467712402, 1549443333.8850622], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 512, 14, 14], "float32"], ["TENSOR", [512, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 512, 14, 14, "float32"], [512, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"e": [["tile_f", "sp", [512, 1, 1, 1]], ["tile_y", "sp", [1, 2, 7, 1]], ["tile_x", "sp", [1, 1, 7, 2]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 321860}]}
18 | {"v": 0.1, "r": [[0.0001771745], 0, 2.1588521003723145, 1549443916.2324407], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 256, 14, 14], "float32"], ["TENSOR", [512, 256, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 256, 14, 14, "float32"], [512, 256, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"e": [["tile_f", "sp", [8, 8, 8, 1]], ["tile_y", "sp", [2, 1, 1, 7]], ["tile_x", "sp", [1, 1, 14, 1]], ["tile_rc", "sp", [32, 8]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 512], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 2227310}]}
19 | {"v": 0.1, "r": [[7.740725e-05], 0, 2.187483310699463, 1549444777.0748699], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 256, 28, 28], "float32"], ["TENSOR", [256, 1, 3, 3], "float32"], [2, 2], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 256, 28, 28, "float32"], [256, 1, 3, 3, "float32"], [2, 2], [1, 1], "float32"], {"e": [["tile_f", "sp", [64, 1, 4, 1]], ["tile_y", "sp", [1, 1, 7, 2]], ["tile_x", "sp", [2, 1, 7, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 228872}]}
20 | {"v": 0.1, "r": [[0.00036090275], 0, 1.9614224433898926, 1549445574.4298754], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 256, 28, 28], "float32"], ["TENSOR", [256, 256, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 256, 28, 28, "float32"], [256, 256, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"e": [["tile_f", "sp", [4, 8, 8, 1]], ["tile_y", "sp", [7, 1, 1, 4]], ["tile_x", "sp", [1, 2, 14, 1]], ["tile_rc", "sp", [32, 8]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 12782082}]}
21 | {"v": 0.1, "r": [[7.871125e-05], 0, 1.6131527423858643, 1549446317.7387452], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 256, 28, 28], "float32"], ["TENSOR", [256, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 256, 28, 28, "float32"], [256, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"e": [["tile_f", "sp", [256, 1, 1, 1]], ["tile_y", "sp", [1, 1, 4, 7]], ["tile_x", "sp", [1, 1, 28, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 1437975}]}
22 | {"v": 0.1, "r": [[0.00019607125], 0, 1.8411102294921875, 1549446989.5312948], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 128, 28, 28], "float32"], ["TENSOR", [256, 128, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 128, 28, 28, "float32"], [256, 128, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"e": [["tile_f", "sp", [4, 4, 8, 2]], ["tile_y", "sp", [7, 1, 1, 4]], ["tile_x", "sp", [1, 1, 28, 1]], ["tile_rc", "sp", [16, 8]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 0], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 7244723}]}
23 | {"v": 0.1, "r": [[0.000116695], 0, 1.8163161277770996, 1549447789.34965], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 128, 56, 56], "float32"], ["TENSOR", [128, 1, 3, 3], "float32"], [2, 2], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 128, 56, 56, "float32"], [128, 1, 3, 3, "float32"], [2, 2], [1, 1], "float32"], {"e": [["tile_f", "sp", [128, 1, 1, 1]], ["tile_y", "sp", [2, 1, 7, 2]], ["tile_x", "sp", [1, 1, 28, 1]], ["auto_unroll_max_step", "ot", 256], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 852480}]}
24 | {"v": 0.1, "r": [[0.00031092525], 0, 3.726764440536499, 1549448612.6387038], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 128, 56, 56], "float32"], ["TENSOR", [128, 128, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 128, 56, 56, "float32"], [128, 128, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"e": [["tile_f", "sp", [2, 4, 8, 2]], ["tile_y", "sp", [28, 1, 2, 1]], ["tile_x", "sp", [1, 7, 8, 1]], ["tile_rc", "sp", [16, 8]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 33245816}]}
25 | {"v": 0.1, "r": [[0.0001369275], 0, 2.400529623031616, 1549449447.2495077], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 128, 56, 56], "float32"], ["TENSOR", [128, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 128, 56, 56, "float32"], [128, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"e": [["tile_f", "sp", [128, 1, 1, 1]], ["tile_y", "sp", [1, 1, 8, 7]], ["tile_x", "sp", [1, 1, 56, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 4126320}]}
26 | {"v": 0.1, "r": [[0.00020027125], 0, 2.450047016143799, 1549450572.8107345], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 64, 56, 56], "float32"], ["TENSOR", [128, 64, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 64, 56, 56, "float32"], [128, 64, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"e": [["tile_f", "sp", [2, 2, 8, 4]], ["tile_y", "sp", [28, 1, 2, 1]], ["tile_x", "sp", [1, 7, 8, 1]], ["tile_rc", "sp", [8, 8]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 0], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 18653840}]}
27 | {"v": 0.1, "r": [[0.000202495], 0, 1.4411323070526123, 1549451405.1190436], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 64, 112, 112], "float32"], ["TENSOR", [64, 1, 3, 3], "float32"], [2, 2], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 64, 112, 112, "float32"], [64, 1, 3, 3, "float32"], [2, 2], [1, 1], "float32"], {"e": [["tile_f", "sp", [64, 1, 1, 1]], ["tile_y", "sp", [7, 2, 4, 1]], ["tile_x", "sp", [1, 1, 56, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 2884140}]}
28 | {"v": 0.1, "r": [[0.000216318], 0, 1.7042787075042725, 1549452490.2196207], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 32, 112, 112], "float32"], ["TENSOR", [64, 32, 1, 1], "float32"], [1, 1], [0, 0], "NCHW", "float32"], {}, ["conv2d", [1, 32, 112, 112, "float32"], [64, 32, 1, 1, "float32"], [1, 1], [0, 0], "NCHW", "float32"], {"e": [["tile_f", "sp", [2, 2, 8, 2]], ["tile_y", "sp", [28, 2, 1, 2]], ["tile_x", "sp", [7, 1, 16, 1]], ["tile_rc", "sp", [8, 4]], ["tile_ry", "sp", [1, 1]], ["tile_rx", "sp", [1, 1]], ["auto_unroll_max_step", "ot", 0], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 33367028}]}
29 | {"v": 0.1, "r": [[0.0001368875], 0, 1.4162554740905762, 1549453299.762188], "i": ["cuda", "topi_nn_depthwise_conv2d_nchw", [["TENSOR", [1, 32, 112, 112], "float32"], ["TENSOR", [32, 1, 3, 3], "float32"], [1, 1], [1, 1], "float32"], {}, ["depthwise_conv2d_nchw", [1, 32, 112, 112, "float32"], [32, 1, 3, 3, "float32"], [1, 1], [1, 1], "float32"], {"e": [["tile_f", "sp", [32, 1, 1, 1]], ["tile_y", "sp", [4, 1, 4, 7]], ["tile_x", "sp", [2, 1, 56, 1]], ["auto_unroll_max_step", "ot", 1500], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 5822992}]}
30 | {"v": 0.1, "r": [[0.0001747745], 0, 1.469916820526123, 1549454076.4118538], "i": ["cuda", "topi_nn_conv2d", [["TENSOR", [1, 3, 224, 224], "float32"], ["TENSOR", [32, 3, 3, 3], "float32"], [2, 2], [1, 1], "NCHW", "float32"], {}, ["conv2d", [1, 3, 224, 224, "float32"], [32, 3, 3, 3, "float32"], [2, 2], [1, 1], "NCHW", "float32"], {"e": [["tile_f", "sp", [1, 2, 8, 2]], ["tile_y", "sp", [56, 2, 1, 1]], ["tile_x", "sp", [7, 1, 16, 1]], ["tile_rc", "sp", [3, 1]], ["tile_ry", "sp", [1, 3]], ["tile_rx", "sp", [1, 3]], ["auto_unroll_max_step", "ot", 0], ["unroll_explicit", "ot", 1]], "t": "direct", "c": null, "i": 33218170}]}
31 | 


--------------------------------------------------------------------------------
/utils.py:
--------------------------------------------------------------------------------
 1 | import os
 2 | import torch
 3 | import shutil
 4 | import numpy as np
 5 | import matplotlib.pyplot as plt
 6 | from PIL import Image
 7 | import math
 8 | 
 9 | cmap = plt.cm.viridis
10 | 
11 | 
12 | def parse_command():
13 |     data_names = ['nyudepthv2']
14 | 
15 |     from dataloaders.dataloader import MyDataloader
16 |     modality_names = MyDataloader.modality_names
17 | 
18 |     import argparse
19 |     parser = argparse.ArgumentParser(description='FastDepth')
20 |     parser.add_argument('--data', metavar='DATA', default='nyudepthv2',
21 |                         choices=data_names,
22 |                         help='dataset: ' + ' | '.join(data_names) + ' (default: nyudepthv2)')
23 |     parser.add_argument('--modality', '-m', metavar='MODALITY', default='rgb', choices=modality_names,
24 |                         help='modality: ' + ' | '.join(modality_names) + ' (default: rgb)')
25 |     parser.add_argument('-j', '--workers', default=16, type=int, metavar='N',
26 |                         help='number of data loading workers (default: 16)')
27 |     parser.add_argument('--print-freq', '-p', default=50, type=int,
28 |                         metavar='N', help='print frequency (default: 50)')
29 |     parser.add_argument('-e', '--evaluate', default='', type=str, metavar='PATH',)
30 |     parser.add_argument('--gpu', default='0', type=str, metavar='N', help="gpu id")
31 |     parser.set_defaults(cuda=True)
32 | 
33 |     args = parser.parse_args()
34 |     return args
35 | 
36 | 
37 | def colored_depthmap(depth, d_min=None, d_max=None):
38 |     if d_min is None:
39 |         d_min = np.min(depth)
40 |     if d_max is None:
41 |         d_max = np.max(depth)
42 |     depth_relative = (depth - d_min) / (d_max - d_min)
43 |     return 255 * cmap(depth_relative)[:,:,:3] # H, W, C
44 | 
45 | 
46 | def merge_into_row(input, depth_target, depth_pred):
47 |     rgb = 255 * np.transpose(np.squeeze(input.cpu().numpy()), (1,2,0)) # H, W, C
48 |     depth_target_cpu = np.squeeze(depth_target.cpu().numpy())
49 |     depth_pred_cpu = np.squeeze(depth_pred.data.cpu().numpy())
50 | 
51 |     d_min = min(np.min(depth_target_cpu), np.min(depth_pred_cpu))
52 |     d_max = max(np.max(depth_target_cpu), np.max(depth_pred_cpu))
53 |     depth_target_col = colored_depthmap(depth_target_cpu, d_min, d_max)
54 |     depth_pred_col = colored_depthmap(depth_pred_cpu, d_min, d_max)
55 |     img_merge = np.hstack([rgb, depth_target_col, depth_pred_col])
56 |     
57 |     return img_merge
58 | 
59 | 
60 | def merge_into_row_with_gt(input, depth_input, depth_target, depth_pred):
61 |     rgb = 255 * np.transpose(np.squeeze(input.cpu().numpy()), (1,2,0)) # H, W, C
62 |     depth_input_cpu = np.squeeze(depth_input.cpu().numpy())
63 |     depth_target_cpu = np.squeeze(depth_target.cpu().numpy())
64 |     depth_pred_cpu = np.squeeze(depth_pred.data.cpu().numpy())
65 | 
66 |     d_min = min(np.min(depth_input_cpu), np.min(depth_target_cpu), np.min(depth_pred_cpu))
67 |     d_max = max(np.max(depth_input_cpu), np.max(depth_target_cpu), np.max(depth_pred_cpu))
68 |     depth_input_col = colored_depthmap(depth_input_cpu, d_min, d_max)
69 |     depth_target_col = colored_depthmap(depth_target_cpu, d_min, d_max)
70 |     depth_pred_col = colored_depthmap(depth_pred_cpu, d_min, d_max)
71 | 
72 |     img_merge = np.hstack([rgb, depth_input_col, depth_target_col, depth_pred_col])
73 | 
74 |     return img_merge
75 | 
76 | 
77 | def add_row(img_merge, row):
78 |     return np.vstack([img_merge, row])
79 | 
80 | 
81 | def save_image(img_merge, filename):
82 |     img_merge = Image.fromarray(img_merge.astype('uint8'))
83 |     img_merge.save(filename)
84 | 


--------------------------------------------------------------------------------