Update: Link for data prepare, fine tuning and inference

README.md

@@ -11,6 +11,9 @@ tags:
 This chatbot model was built via Parameter-Efficient Fine-Tuning of [GPT-J-6B](https://huggingface.co/EleutherAI/gpt-j-6b) on all 16.3k rows of Medical Data. Finetuning was executed on a single A100 (40 GB) for roughly 1 day 7 hours.
 
 * Model license: GPT-J Community License Agreement
+* Data Prepare: [data_prepapre code](https://github.com/ghimiresunil/LLM-PowerHouse-A-Curated-Guide-for-Large-Language-Models-with-Custom-Training-and-Inferencing/blob/main/src/data_generate_prepare/data_prepare.py)
+* Finetuning: [finetune code](https://github.com/ghimiresunil/LLM-PowerHouse-A-Curated-Guide-for-Large-Language-Models-with-Custom-Training-and-Inferencing/blob/main/src/train_inference_int_peft/trainer_int_peft_lora.py)
+* Inference: [inference code](https://github.com/ghimiresunil/LLM-PowerHouse-A-Curated-Guide-for-Large-Language-Models-with-Custom-Training-and-Inferencing/blob/main/src/train_inference_int_peft/inference_int_peft_lora.py)
 
 ### Example prompts and responses
 
@@ -82,7 +85,16 @@ This model was trained on a single A100 (40 GB) for about 1 Day 7 hours.
 
 Run: July 23, 2023
 * args: {'lr': 0.001, 'num_epochs': 10, 'seed': 42}
-* logs: {'train_runtime': 118063.0495, 'train_samples_per_second': 1.107, 'train_steps_per_second': 0.069, 'train_loss': 0.7715376593637642, 'epoch': 10.0}
+* log_of_epoch_01:{'eval_loss': 0.9936667084693909, 'eval_runtime': 450.8767, 'eval_samples_per_second': 7.246, 'eval_steps_per_second': 0.455, 'epoch': 1.0}
+* log_of_epoch_02:{'eval_loss': 0.9738781452178955, 'eval_runtime': 447.3755, 'eval_samples_per_second': 7.303, 'eval_steps_per_second': 0.458, 'epoch': 2.0}
+* log_of_epoch_03:{'eval_loss': 0.9600604176521301, 'eval_runtime': 441.2023, 'eval_samples_per_second': 7.405, 'eval_steps_per_second': 0.465, 'epoch': 3.0}
+* log_of_epoch_04:{'eval_loss': 0.9634631872177124, 'eval_runtime': 441.53, 'eval_samples_per_second': 7.399, 'eval_steps_per_second': 0.464, 'epoch': 4.0}
+* log_of_epoch_05:{'eval_loss': 0.961345374584198, 'eval_runtime': 441.3189, 'eval_samples_per_second': 7.403, 'eval_steps_per_second': 0.465, 'epoch': 5.0}
+* log_of_epoch_06:{'eval_loss': 0.9655225872993469, 'eval_runtime': 441.9449, 'eval_samples_per_second': 7.392, 'eval_steps_per_second': 0.464, 'epoch': 6.0}
+* log_of_epoch_07:{'eval_loss': 0.9740663766860962, 'eval_runtime': 441.7603, 'eval_samples_per_second': 7.395, 'eval_steps_per_second': 0.464, 'epoch': 7.0}
+* log_of_epoch_08:{'eval_loss': 0.9907786846160889, 'eval_runtime': 441.6064, 'eval_samples_per_second': 7.398, 'eval_steps_per_second': 0.464, 'epoch': 8.0}
+* log_of_epoch_09:{'eval_loss': 1.0046937465667725, 'eval_runtime': 441.9242, 'eval_samples_per_second': 7.393, 'eval_steps_per_second': 0.464, 'epoch': 9.0}
+* log_of_epoch_10:{'train_runtime': 118063.0495, 'train_samples_per_second': 1.107, 'train_steps_per_second': 0.069, 'train_loss': 0.7715376593637642, 'epoch': 10.0}
 
 ## PreTraining Data
 For more details on the pretraining process, see [GPT-J-6B](https://huggingface.co/EleutherAI/gpt-j-6b).