End of training

README.md

@@ -6,7 +6,7 @@ tags:
 - axolotl
 - generated_from_trainer
 model-index:
-- name: proof-reading-SeaLLM3-7B-Chat-3090-v9
+- name: proof-reading-SeaLLM3-7B-Chat-3090-v10
   results: []
 ---
 
@@ -26,8 +26,9 @@ load_in_4bit: true
 strict: false
 
 datasets:
-  - path: Tippawan/p9-seallm
+  - path: Tippawan/pr-10-wiki-seallm
     type: sharegpt
+    split: 'train[:100000]'
     conversation: chatml
     field_messages: messages
 chat_template: chatml
@@ -41,7 +42,7 @@ eval_sample_packing: false
 pad_to_sequence_len: false
 
 push_to_hub: true
-hub_model_id: Tippawan/proof-reading-SeaLLM3-7B-Chat-3090-v9  # Replace with your Hugging Face repo ID
+hub_model_id: Tippawan/proof-reading-SeaLLM3-7B-Chat-3090-v10  # Replace with your Hugging Face repo ID
 use_auth_token: true  # Ensure you have set your Hugging Face API token in the environment
 hub_private_repo: true  # Set to true if you want the repository to be private
 hub_strategy: all_checkpoints
@@ -49,22 +50,22 @@ save_total_limit: 3
 load_best_model_at_end: true
 
 adapter: lora
-lora_model_dir: Tippawan/proof-reading-SeaLLM3-7B-Chat-3090-v8
+lora_model_dir: Tippawan/proof-reading-SeaLLM3-7B-Chat-3090-v9
 lora_r: 16
 lora_alpha: 32
 lora_dropout: 0.05
 lora_target_linear: true
 lora_fan_in_fan_out:
 
-wandb_project: proof-reading-SeaLLM3-7B-Chat-3090-v9
+wandb_project: proof-reading-SeaLLM3-7B-Chat-3090-v10
 wandb_entity: 
 wandb_watch:
 wandb_name:
 wandb_log_model:
 
 gradient_accumulation_steps: 4
-micro_batch_size: 8
-num_epochs: 3 #editted 3
+micro_batch_size: 2
+num_epochs: 1 #editted 3
 optimizer: adamw_torch
 lr_scheduler: cosine
 learning_rate: 0.0002
@@ -96,7 +97,7 @@ special_tokens:
 
 </details><br>
 
-# proof-reading-SeaLLM3-7B-Chat-3090-v9
+# proof-reading-SeaLLM3-7B-Chat-3090-v10
 
 This model is a fine-tuned version of [SeaLLMs/SeaLLM3-7B-Chat](https://huggingface.co/SeaLLMs/SeaLLM3-7B-Chat) on the None dataset.
 
@@ -118,15 +119,15 @@ More information needed
 
 The following hyperparameters were used during training:
 - learning_rate: 0.0002
-- train_batch_size: 8
-- eval_batch_size: 8
+- train_batch_size: 2
+- eval_batch_size: 2
 - seed: 42
 - gradient_accumulation_steps: 4
-- total_train_batch_size: 32
+- total_train_batch_size: 8
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: cosine
 - lr_scheduler_warmup_steps: 10
-- num_epochs: 3
+- num_epochs: 1
 
 ### Training results
 

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5e8cba0fdd3c3de5054de7a5ea485dc2cb50911a2fef6c7be0f2d46d95e025c6
+oid sha256:510bb9fcb5e688917e13ab4eb4ad4b47014c4f16f157be766a5439ece5fe30b1
 size 161621802