├── LICENSE ├── README.md ├── Step 0 Constructing the dataset.py ├── Step 1 Fine tuning using QLoRA.py └── Step 2 Fine tuning using LoRA.py /LICENSE: -------------------------------------------------------------------------------- 1 | Apache License 2 | Version 2.0, January 2004 3 | http://www.apache.org/licenses/ 4 | 5 | TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION 6 | 7 | 1. Definitions. 8 | 9 | "License" shall mean the terms and conditions for use, reproduction, 10 | and distribution as defined by Sections 1 through 9 of this document. 11 | 12 | "Licensor" shall mean the copyright owner or entity authorized by 13 | the copyright owner that is granting the License. 14 | 15 | "Legal Entity" shall mean the union of the acting entity and all 16 | other entities that control, are controlled by, or are under common 17 | control with that entity. For the purposes of this definition, 18 | "control" means (i) the power, direct or indirect, to cause the 19 | direction or management of such entity, whether by contract or 20 | otherwise, or (ii) ownership of fifty percent (50%) or more of the 21 | outstanding shares, or (iii) beneficial ownership of such entity. 22 | 23 | "You" (or "Your") shall mean an individual or Legal Entity 24 | exercising permissions granted by this License. 25 | 26 | "Source" form shall mean the preferred form for making modifications, 27 | including but not limited to software source code, documentation 28 | source, and configuration files. 29 | 30 | "Object" form shall mean any form resulting from mechanical 31 | transformation or translation of a Source form, including but 32 | not limited to compiled object code, generated documentation, 33 | and conversions to other media types. 34 | 35 | "Work" shall mean the work of authorship, whether in Source or 36 | Object form, made available under the License, as indicated by a 37 | copyright notice that is included in or attached to the work 38 | (an example is provided in the Appendix below). 39 | 40 | "Derivative Works" shall mean any work, whether in Source or Object 41 | form, that is based on (or derived from) the Work and for which the 42 | editorial revisions, annotations, elaborations, or other modifications 43 | represent, as a whole, an original work of authorship. For the purposes 44 | of this License, Derivative Works shall not include works that remain 45 | separable from, or merely link (or bind by name) to the interfaces of, 46 | the Work and Derivative Works thereof. 47 | 48 | "Contribution" shall mean any work of authorship, including 49 | the original version of the Work and any modifications or additions 50 | to that Work or Derivative Works thereof, that is intentionally 51 | submitted to Licensor for inclusion in the Work by the copyright owner 52 | or by an individual or Legal Entity authorized to submit on behalf of 53 | the copyright owner. For the purposes of this definition, "submitted" 54 | means any form of electronic, verbal, or written communication sent 55 | to the Licensor or its representatives, including but not limited to 56 | communication on electronic mailing lists, source code control systems, 57 | and issue tracking systems that are managed by, or on behalf of, the 58 | Licensor for the purpose of discussing and improving the Work, but 59 | excluding communication that is conspicuously marked or otherwise 60 | designated in writing by the copyright owner as "Not a Contribution." 61 | 62 | "Contributor" shall mean Licensor and any individual or Legal Entity 63 | on behalf of whom a Contribution has been received by Licensor and 64 | subsequently incorporated within the Work. 65 | 66 | 2. Grant of Copyright License. Subject to the terms and conditions of 67 | this License, each Contributor hereby grants to You a perpetual, 68 | worldwide, non-exclusive, no-charge, royalty-free, irrevocable 69 | copyright license to reproduce, prepare Derivative Works of, 70 | publicly display, publicly perform, sublicense, and distribute the 71 | Work and such Derivative Works in Source or Object form. 72 | 73 | 3. Grant of Patent License. Subject to the terms and conditions of 74 | this License, each Contributor hereby grants to You a perpetual, 75 | worldwide, non-exclusive, no-charge, royalty-free, irrevocable 76 | (except as stated in this section) patent license to make, have made, 77 | use, offer to sell, sell, import, and otherwise transfer the Work, 78 | where such license applies only to those patent claims licensable 79 | by such Contributor that are necessarily infringed by their 80 | Contribution(s) alone or by combination of their Contribution(s) 81 | with the Work to which such Contribution(s) was submitted. If You 82 | institute patent litigation against any entity (including a 83 | cross-claim or counterclaim in a lawsuit) alleging that the Work 84 | or a Contribution incorporated within the Work constitutes direct 85 | or contributory patent infringement, then any patent licenses 86 | granted to You under this License for that Work shall terminate 87 | as of the date such litigation is filed. 88 | 89 | 4. Redistribution. You may reproduce and distribute copies of the 90 | Work or Derivative Works thereof in any medium, with or without 91 | modifications, and in Source or Object form, provided that You 92 | meet the following conditions: 93 | 94 | (a) You must give any other recipients of the Work or 95 | Derivative Works a copy of this License; and 96 | 97 | (b) You must cause any modified files to carry prominent notices 98 | stating that You changed the files; and 99 | 100 | (c) You must retain, in the Source form of any Derivative Works 101 | that You distribute, all copyright, patent, trademark, and 102 | attribution notices from the Source form of the Work, 103 | excluding those notices that do not pertain to any part of 104 | the Derivative Works; and 105 | 106 | (d) If the Work includes a "NOTICE" text file as part of its 107 | distribution, then any Derivative Works that You distribute must 108 | include a readable copy of the attribution notices contained 109 | within such NOTICE file, excluding those notices that do not 110 | pertain to any part of the Derivative Works, in at least one 111 | of the following places: within a NOTICE text file distributed 112 | as part of the Derivative Works; within the Source form or 113 | documentation, if provided along with the Derivative Works; or, 114 | within a display generated by the Derivative Works, if and 115 | wherever such third-party notices normally appear. The contents 116 | of the NOTICE file are for informational purposes only and 117 | do not modify the License. You may add Your own attribution 118 | notices within Derivative Works that You distribute, alongside 119 | or as an addendum to the NOTICE text from the Work, provided 120 | that such additional attribution notices cannot be construed 121 | as modifying the License. 122 | 123 | You may add Your own copyright statement to Your modifications and 124 | may provide additional or different license terms and conditions 125 | for use, reproduction, or distribution of Your modifications, or 126 | for any such Derivative Works as a whole, provided Your use, 127 | reproduction, and distribution of the Work otherwise complies with 128 | the conditions stated in this License. 129 | 130 | 5. Submission of Contributions. Unless You explicitly state otherwise, 131 | any Contribution intentionally submitted for inclusion in the Work 132 | by You to the Licensor shall be under the terms and conditions of 133 | this License, without any additional terms or conditions. 134 | Notwithstanding the above, nothing herein shall supersede or modify 135 | the terms of any separate license agreement you may have executed 136 | with Licensor regarding such Contributions. 137 | 138 | 6. Trademarks. This License does not grant permission to use the trade 139 | names, trademarks, service marks, or product names of the Licensor, 140 | except as required for reasonable and customary use in describing the 141 | origin of the Work and reproducing the content of the NOTICE file. 142 | 143 | 7. Disclaimer of Warranty. Unless required by applicable law or 144 | agreed to in writing, Licensor provides the Work (and each 145 | Contributor provides its Contributions) on an "AS IS" BASIS, 146 | WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or 147 | implied, including, without limitation, any warranties or conditions 148 | of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A 149 | PARTICULAR PURPOSE. You are solely responsible for determining the 150 | appropriateness of using or redistributing the Work and assume any 151 | risks associated with Your exercise of permissions under this License. 152 | 153 | 8. Limitation of Liability. In no event and under no legal theory, 154 | whether in tort (including negligence), contract, or otherwise, 155 | unless required by applicable law (such as deliberate and grossly 156 | negligent acts) or agreed to in writing, shall any Contributor be 157 | liable to You for damages, including any direct, indirect, special, 158 | incidental, or consequential damages of any character arising as a 159 | result of this License or out of the use or inability to use the 160 | Work (including but not limited to damages for loss of goodwill, 161 | work stoppage, computer failure or malfunction, or any and all 162 | other commercial damages or losses), even if such Contributor 163 | has been advised of the possibility of such damages. 164 | 165 | 9. Accepting Warranty or Additional Liability. While redistributing 166 | the Work or Derivative Works thereof, You may choose to offer, 167 | and charge a fee for, acceptance of support, warranty, indemnity, 168 | or other liability obligations and/or rights consistent with this 169 | License. However, in accepting such obligations, You may act only 170 | on Your own behalf and on Your sole responsibility, not on behalf 171 | of any other Contributor, and only if You agree to indemnify, 172 | defend, and hold each Contributor harmless for any liability 173 | incurred by, or claims asserted against, such Contributor by reason 174 | of your accepting any such warranty or additional liability. 175 | 176 | END OF TERMS AND CONDITIONS 177 | 178 | APPENDIX: How to apply the Apache License to your work. 179 | 180 | To apply the Apache License to your work, attach the following 181 | boilerplate notice, with the fields enclosed by brackets "[]" 182 | replaced with your own identifying information. (Don't include 183 | the brackets!) The text should be enclosed in the appropriate 184 | comment syntax for the file format. We also recommend that a 185 | file or class name and description of purpose be included on the 186 | same "printed page" as the copyright notice for easier 187 | identification within third-party archives. 188 | 189 | Copyright [yyyy] [name of copyright owner] 190 | 191 | Licensed under the Apache License, Version 2.0 (the "License"); 192 | you may not use this file except in compliance with the License. 193 | You may obtain a copy of the License at 194 | 195 | http://www.apache.org/licenses/LICENSE-2.0 196 | 197 | Unless required by applicable law or agreed to in writing, software 198 | distributed under the License is distributed on an "AS IS" BASIS, 199 | WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 200 | See the License for the specific language governing permissions and 201 | limitations under the License. 202 | -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 | # Tuning-the-Finetuning 2 | Tuning the Finetuning: An exploration of achieving success with QLoRA 3 | 4 | This repo is associated with the following Databricks blog: https://www.databricks.com/blog/efficient-fine-tuning-lora-guide-llms 5 | -------------------------------------------------------------------------------- /Step 0 Constructing the dataset.py: -------------------------------------------------------------------------------- 1 | # Databricks notebook source 2 | from datasets import load_dataset , Dataset, concatenate_datasets 3 | import numpy as np 4 | import pandas as pd 5 | import random 6 | 7 | # COMMAND ---------- 8 | 9 | rd_ds = load_dataset("xiyuez/red-dot-design-award-product-description") 10 | rd_df = pd.DataFrame(rd_ds['train']) 11 | display(rd_df) 12 | 13 | # COMMAND ---------- 14 | 15 | rd_df['instruction'] = 'Create a detailed description for the following product: '+ rd_df['product']+', belonging to category: '+ rd_df['category'] 16 | rd_df = rd_df[['instruction', 'description']] 17 | display(rd_df) 18 | 19 | # COMMAND ---------- 20 | 21 | rd_df_sample = rd_df.sample(n=5000, random_state=42) 22 | display(rd_df_sample) 23 | 24 | # COMMAND ---------- 25 | 26 | template = """Below is an instruction that describes a task. Write a response that appropriately completes the request. 27 | 28 | ### Instruction: 29 | {} 30 | 31 | ### Response:\n""" 32 | 33 | # COMMAND ---------- 34 | 35 | rd_df_sample['prompt'] = rd_df_sample["instruction"].apply(lambda x: template.format(x)) 36 | 37 | # COMMAND ---------- 38 | 39 | rd_df_sample.rename(columns={'description': 'response'}, inplace=True) 40 | 41 | # COMMAND ---------- 42 | 43 | rd_df_sample['response'] = rd_df_sample['response'] + "\n### End" 44 | rd_df_sample = rd_df_sample[['prompt', 'response']] 45 | display(rd_df_sample) 46 | 47 | 48 | # COMMAND ---------- 49 | 50 | # MAGIC 51 | # MAGIC %sql 52 | # MAGIC CREATE DATABASE IF NOT EXISTS description_generator; 53 | # MAGIC USE description_generator; 54 | # MAGIC 55 | 56 | # COMMAND ---------- 57 | 58 | spark.createDataFrame(rd_df_sample).write.saveAsTable('product_name_to_description') 59 | 60 | -------------------------------------------------------------------------------- /Step 1 Fine tuning using QLoRA.py: -------------------------------------------------------------------------------- 1 | # Databricks notebook source 2 | # MAGIC %pip install transformers==4.31.0 datasets==2.13.0 peft==0.4.0 accelerate==0.21.0 bitsandbytes==0.40.2 trl==0.4.7 3 | 4 | # COMMAND ---------- 5 | 6 | from peft import get_peft_config, PeftModel, PeftConfig, get_peft_model, LoraConfig, TaskType 7 | from transformers import AutoModelForCausalLM 8 | from transformers import LlamaTokenizer, LlamaForCausalLM 9 | import torch 10 | from transformers.trainer_callback import TrainerCallback 11 | import os 12 | from transformers import BitsAndBytesConfig 13 | from trl import SFTTrainer 14 | import mlflow 15 | 16 | # COMMAND ---------- 17 | 18 | # MAGIC %sql 19 | # MAGIC USE description_generator; 20 | 21 | # COMMAND ---------- 22 | 23 | df = spark.sql("SELECT * FROM product_name_to_description").toPandas() 24 | df['text'] = df["prompt"]+df["response"] 25 | df.drop(columns=['prompt', 'response'], inplace=True) 26 | display(df), df.shape 27 | 28 | # COMMAND ---------- 29 | 30 | from datasets import load_dataset 31 | from datasets import Dataset 32 | dataset = Dataset.from_pandas(df).train_test_split(test_size=0.05, seed=42) 33 | 34 | # COMMAND ---------- 35 | 36 | target_modules = ['q_proj','k_proj','v_proj','o_proj','gate_proj','down_proj','up_proj','lm_head'] 37 | #or if only tageting attention blocks 38 | # target_modules = ['q_proj','v_proj'] 39 | 40 | lora_config = LoraConfig( 41 | r=8,#or r=16 42 | lora_alpha=8, 43 | lora_dropout=0.05, 44 | bias="none", 45 | target_modules = target_modules, 46 | task_type="CAUSAL_LM", 47 | ) 48 | 49 | base_dir = "" 50 | 51 | per_device_train_batch_size = 4 52 | gradient_accumulation_steps = 4 53 | optim = 'adamw_hf' 54 | learning_rate = 1e-5 55 | max_grad_norm = 0.3 56 | warmup_ratio = 0.03 57 | lr_scheduler_type = "linear" 58 | 59 | # COMMAND ---------- 60 | 61 | from transformers import TrainingArguments 62 | training_args = TrainingArguments( 63 | output_dir=base_dir, 64 | save_strategy="epoch", 65 | evaluation_strategy="epoch", 66 | num_train_epochs = 3.0, 67 | per_device_train_batch_size=per_device_train_batch_size, 68 | gradient_accumulation_steps=gradient_accumulation_steps, 69 | optim=optim, 70 | learning_rate=learning_rate, 71 | fp16=True, 72 | max_grad_norm=max_grad_norm, 73 | warmup_ratio=warmup_ratio, 74 | group_by_length=True, 75 | lr_scheduler_type=lr_scheduler_type, 76 | ) 77 | 78 | 79 | # COMMAND ---------- 80 | 81 | nf4_config = BitsAndBytesConfig( 82 | load_in_4bit=True, 83 | bnb_4bit_quant_type="nf4", 84 | bnb_4bit_use_double_quant=True, 85 | bnb_4bit_compute_dtype=torch.bfloat16 86 | ) 87 | 88 | # COMMAND ---------- 89 | 90 | model_path = 'openlm-research/open_llama_3b_v2' 91 | 92 | # COMMAND ---------- 93 | 94 | tokenizer = LlamaTokenizer.from_pretrained(model_path) 95 | tokenizer.add_special_tokens({'pad_token': '[PAD]'}) 96 | 97 | # COMMAND ---------- 98 | 99 | model = LlamaForCausalLM.from_pretrained( 100 | model_path, device_map='auto', quantization_config=nf4_config, 101 | ) 102 | 103 | # COMMAND ---------- 104 | 105 | model = get_peft_model(model, lora_config) 106 | model.print_trainable_parameters() 107 | 108 | # COMMAND ---------- 109 | 110 | trainer = SFTTrainer( 111 | model, 112 | train_dataset=dataset['train'], 113 | eval_dataset = dataset['test'], 114 | dataset_text_field="text", 115 | max_seq_length=256, 116 | args=training_args, 117 | ) 118 | #Upcast layer norms to float 32 for stability 119 | for name, module in trainer.model.named_modules(): 120 | if "norm" in name: 121 | module = module.to(torch.float32) 122 | 123 | # COMMAND ---------- 124 | # Initiate the training process 125 | with mlflow.start_run(run_name='run_name_of_choice'): 126 | trainer.train() 127 | 128 | # COMMAND ---------- 129 | 130 | # #https://github.com/NVIDIA/apex/issues/965 131 | # for param in model.parameters(): 132 | # # Check if parameter dtype is Half (float16) 133 | # if param.dtype == torch.float16: 134 | # param.data = param.data.to(torch.float32) 135 | 136 | # COMMAND ---------- 137 | 138 | # MAGIC %md 139 | # MAGIC ### If loading from saved adapter 140 | 141 | # COMMAND ---------- 142 | 143 | dbutils.fs.ls('') 144 | 145 | # COMMAND ---------- 146 | 147 | model_path = 'openlm-research/open_llama_3b_v2' 148 | 149 | # COMMAND ---------- 150 | 151 | tokenizer = LlamaTokenizer.from_pretrained(model_path) 152 | tokenizer.add_special_tokens({'pad_token': '[PAD]'}) 153 | 154 | # COMMAND ---------- 155 | 156 | model = LlamaForCausalLM.from_pretrained( 157 | model_path, load_in_8bit=True, device_map='auto', 158 | ) 159 | 160 | # COMMAND ---------- 161 | 162 | peft_model_id = '' 163 | 164 | # COMMAND ---------- 165 | 166 | peft_model = PeftModel.from_pretrained(model, peft_model_id) 167 | 168 | # COMMAND ---------- 169 | 170 | test_strings = ["Create a detailed description for the following product: Corelogic Smooth Mouse, belonging to category: Optical Mouse", 171 | "Create a detailed description for the following product: Hoover Lightspeed, belonging to category: Cordless Vacuum Cleaner", 172 | "Create a detailed description for the following product: Flattronic Cinematron, belonging to category: High Definition Flatscreen TV"] 173 | 174 | # COMMAND ---------- 175 | 176 | predictions = [] 177 | for test in test_strings: 178 | prompt = """Below is an instruction that describes a task. Write a response that appropriately completes the request. 179 | 180 | ### Instruction: 181 | {} 182 | 183 | ### Response:""".format(test) 184 | input_ids = tokenizer(prompt, return_tensors="pt").input_ids.to('cuda') 185 | 186 | generation_output = model.generate( 187 | input_ids=input_ids, max_new_tokens=156 188 | ) 189 | predictions.append(tokenizer.decode(generation_output[0])) 190 | 191 | # COMMAND ---------- 192 | 193 | def extract_response_text(input_string): 194 | start_marker = '### Response:' 195 | end_marker = '###' 196 | 197 | start_index = input_string.find(start_marker) 198 | if start_index == -1: 199 | return None 200 | 201 | start_index += len(start_marker) 202 | 203 | end_index = input_string.find(end_marker, start_index) 204 | if end_index == -1: 205 | return input_string[start_index:] 206 | 207 | return input_string[start_index:end_index].strip() 208 | 209 | # COMMAND ---------- 210 | 211 | # predictions[2] 212 | 213 | # COMMAND ---------- 214 | 215 | for i in range(3): 216 | pred = predictions[i] 217 | text = test_strings[i] 218 | print(text+'\n') 219 | print(extract_response_text(pred)) 220 | print('--------') 221 | 222 | # COMMAND ---------- 223 | 224 | 225 | -------------------------------------------------------------------------------- /Step 2 Fine tuning using LoRA.py: -------------------------------------------------------------------------------- 1 | # Databricks notebook source 2 | # MAGIC %pip install transformers==4.31.0 datasets==2.13.0 peft==0.4.0 accelerate==0.21.0 bitsandbytes==0.40.2 trl==0.4.7 3 | 4 | # COMMAND ---------- 5 | 6 | from peft import get_peft_config, PeftModel, PeftConfig, get_peft_model, LoraConfig, TaskType 7 | from transformers import AutoModelForCausalLM 8 | from transformers import LlamaTokenizer, LlamaForCausalLM 9 | import torch 10 | from transformers.trainer_callback import TrainerCallback 11 | import os 12 | from transformers import BitsAndBytesConfig 13 | from trl import SFTTrainer 14 | import mlflow 15 | 16 | # COMMAND ---------- 17 | 18 | # MAGIC %sql 19 | # MAGIC USE description_generator; 20 | 21 | # COMMAND ---------- 22 | 23 | df = spark.sql("SELECT * FROM product_name_to_description").toPandas() 24 | df['text'] = df["prompt"]+df["response"] 25 | df.drop(columns=['prompt', 'response'], inplace=True) 26 | display(df), df.shape 27 | 28 | # COMMAND ---------- 29 | 30 | from datasets import load_dataset 31 | from datasets import Dataset 32 | dataset = Dataset.from_pandas(df).train_test_split(test_size=0.05, seed=42) 33 | 34 | # COMMAND ---------- 35 | 36 | target_modules = ['q_proj','k_proj','v_proj','o_proj','gate_proj','down_proj','up_proj','lm_head'] 37 | #or 38 | # target_modules = ['q_proj','v_proj'] 39 | 40 | lora_config = LoraConfig( 41 | r=8,#or r=16 42 | lora_alpha=8, 43 | lora_dropout=0.05, 44 | bias="none", 45 | target_modules = target_modules, 46 | task_type="CAUSAL_LM", 47 | ) 48 | 49 | base_dir = "" 50 | 51 | per_device_train_batch_size = 4 52 | gradient_accumulation_steps = 4 53 | optim = 'adamw_hf' 54 | learning_rate = 1e-5 55 | max_grad_norm = 0.3 56 | warmup_ratio = 0.03 57 | lr_scheduler_type = "linear" 58 | 59 | # COMMAND ---------- 60 | 61 | from transformers import TrainingArguments 62 | training_args = TrainingArguments( 63 | output_dir=base_dir, 64 | save_strategy="epoch", 65 | evaluation_strategy="epoch", 66 | num_train_epochs = 3.0, 67 | per_device_train_batch_size=per_device_train_batch_size, 68 | gradient_accumulation_steps=gradient_accumulation_steps, 69 | optim=optim, 70 | learning_rate=learning_rate, 71 | fp16=True, 72 | max_grad_norm=max_grad_norm, 73 | warmup_ratio=warmup_ratio, 74 | group_by_length=True, 75 | lr_scheduler_type=lr_scheduler_type, 76 | ) 77 | 78 | 79 | # COMMAND ---------- 80 | 81 | model_path = 'openlm-research/open_llama_3b_v2' 82 | 83 | # COMMAND ---------- 84 | 85 | tokenizer = LlamaTokenizer.from_pretrained(model_path) 86 | tokenizer.add_special_tokens({'pad_token': '[PAD]'}) 87 | 88 | # COMMAND ---------- 89 | 90 | model = LlamaForCausalLM.from_pretrained( 91 | model_path, device_map='auto', load_in_8bit=True, 92 | ) 93 | 94 | # COMMAND ---------- 95 | 96 | model = get_peft_model(model, lora_config) 97 | model.print_trainable_parameters() 98 | 99 | # COMMAND ---------- 100 | 101 | trainer = SFTTrainer( 102 | model, 103 | train_dataset=dataset['train'], 104 | eval_dataset = dataset['test'], 105 | dataset_text_field="text", 106 | max_seq_length=256, 107 | args=training_args, 108 | ) 109 | #Upcast layer norms to float 32 for stability 110 | for name, module in trainer.model.named_modules(): 111 | if "norm" in name: 112 | module = module.to(torch.float32) 113 | 114 | # COMMAND ---------- 115 | 116 | # Initiate the training process 117 | with mlflow.start_run(run_name='run_name_of_choice'): 118 | trainer.train() 119 | 120 | # COMMAND ---------- 121 | 122 | # #https://github.com/NVIDIA/apex/issues/965 123 | # for param in model.parameters(): 124 | # # Check if parameter dtype is Half (float16) 125 | # if param.dtype == torch.float16: 126 | # param.data = param.data.to(torch.float32) 127 | 128 | # COMMAND ---------- 129 | 130 | # MAGIC %md 131 | # MAGIC ### If loading from saved adapter 132 | 133 | # COMMAND ---------- 134 | 135 | dbutils.fs.ls('') 136 | 137 | # COMMAND ---------- 138 | 139 | model_path = 'openlm-research/open_llama_3b_v2' 140 | 141 | # COMMAND ---------- 142 | 143 | tokenizer = LlamaTokenizer.from_pretrained(model_path) 144 | tokenizer.add_special_tokens({'pad_token': '[PAD]'}) 145 | 146 | # COMMAND ---------- 147 | 148 | model = LlamaForCausalLM.from_pretrained( 149 | model_path, load_in_8bit=True, device_map='auto', 150 | ) 151 | 152 | # COMMAND ---------- 153 | 154 | peft_model_id = '' 155 | 156 | # COMMAND ---------- 157 | 158 | peft_model = PeftModel.from_pretrained(model, peft_model_id) 159 | 160 | # COMMAND ---------- 161 | 162 | test_strings = ["Create a detailed description for the following product: Corelogic Smooth Mouse, belonging to category: Optical Mouse", 163 | "Create a detailed description for the following product: Hoover Lightspeed, belonging to category: Cordless Vacuum Cleaner", 164 | "Create a detailed description for the following product: Flattronic Cinematron, belonging to category: High Definition Flatscreen TV"] 165 | 166 | # COMMAND ---------- 167 | 168 | predictions = [] 169 | for test in test_strings: 170 | prompt = """Below is an instruction that describes a task. Write a response that appropriately completes the request. 171 | 172 | ### Instruction: 173 | {} 174 | 175 | ### Response:""".format(test) 176 | input_ids = tokenizer(prompt, return_tensors="pt").input_ids.to('cuda') 177 | 178 | generation_output = model.generate( 179 | input_ids=input_ids, max_new_tokens=156 180 | ) 181 | predictions.append(tokenizer.decode(generation_output[0])) 182 | 183 | # COMMAND ---------- 184 | 185 | def extract_response_text(input_string): 186 | start_marker = '### Response:' 187 | end_marker = '###' 188 | 189 | start_index = input_string.find(start_marker) 190 | if start_index == -1: 191 | return None 192 | 193 | start_index += len(start_marker) 194 | 195 | end_index = input_string.find(end_marker, start_index) 196 | if end_index == -1: 197 | return input_string[start_index:] 198 | 199 | return input_string[start_index:end_index].strip() 200 | 201 | # COMMAND ---------- 202 | 203 | # predictions[2] 204 | 205 | # COMMAND ---------- 206 | 207 | for i in range(3): 208 | pred = predictions[i] 209 | text = test_strings[i] 210 | print(text+'\n') 211 | print(extract_response_text(pred)) 212 | print('--------') 213 | 214 | # COMMAND ---------- 215 | 216 | 217 | --------------------------------------------------------------------------------