End of training

README.md

@@ -34,8 +34,8 @@ datasets:
 dataset_prepared_path: AFinUpTagsNoTestNoExNewCodeLlama
 val_set_size: 0
 output_dir: models/Acodellama70bL4
-# lora_model_dir: models/codellamaTest1/checkpoint-80
-# auto_resume_from_checkpoints: true
+lora_model_dir: models/Acodellama70bL4/checkpoint-44
+auto_resume_from_checkpoints: true
 sequence_len: 4096
 sample_packing: true
 pad_to_sequence_len: true
@@ -54,12 +54,12 @@ wandb_project: 'codellamaFeed'
 wandb_entity:
 wandb_watch:
 wandb_run_id:
-wandb_name: 'A70bL4'                                     
+wandb_name: 'A70bL4'                                      
 wandb_log_model:
 
 gradient_accumulation_steps: 4
 micro_batch_size: 1
-num_epochs: 4
+num_epochs: 8
 optimizer: adamw_torch
 lr_scheduler: cosine
 learning_rate: 0.0002
@@ -98,7 +98,7 @@ special_tokens:
 
 </details><br>
 
-[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="200" height="32"/>](https://wandb.ai/afrias5/codellamaFeed/runs/pb22442t)
+[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="200" height="32"/>](https://wandb.ai/afrias5/codellamaFeed/runs/5vpimzij)
 # Acodellama70b
 
 This model is a fine-tuned version of [meta-llama/CodeLlama-70b-Python-hf](https://huggingface.co/meta-llama/CodeLlama-70b-Python-hf) on the None dataset.
@@ -132,7 +132,7 @@ The following hyperparameters were used during training:
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: cosine
 - lr_scheduler_warmup_steps: 10
-- num_epochs: 4
+- num_epochs: 8
 
 ### Training results