Model save

README.md

@@ -4,7 +4,6 @@ license: apache-2.0
 base_model: Qwen/Qwen2.5-32B
 tags:
 - llama-factory
-- full
 - generated_from_trainer
 model-index:
 - name: pretrain
@@ -16,9 +15,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # pretrain
 
-This model is a fine-tuned version of [Qwen/Qwen2.5-32B](https://huggingface.co/Qwen/Qwen2.5-32B) on the openalex_small dataset.
-It achieves the following results on the evaluation set:
-- Loss: 1.1904
+This model is a fine-tuned version of [Qwen/Qwen2.5-32B](https://huggingface.co/Qwen/Qwen2.5-32B) on an unknown dataset.
 
 ## Model description
 
@@ -39,13 +36,13 @@ More information needed
 The following hyperparameters were used during training:
 - learning_rate: 0.0001
 - train_batch_size: 1
-- eval_batch_size: 1
+- eval_batch_size: 8
 - seed: 42
 - distributed_type: multi-GPU
 - num_devices: 8
-- gradient_accumulation_steps: 8
-- total_train_batch_size: 64
-- total_eval_batch_size: 8
+- gradient_accumulation_steps: 16
+- total_train_batch_size: 128
+- total_eval_batch_size: 64
 - optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
 - lr_scheduler_type: cosine
 - lr_scheduler_warmup_ratio: 0.1
@@ -53,9 +50,6 @@ The following hyperparameters were used during training:
 
 ### Training results
 
-| Training Loss | Epoch  | Step | Validation Loss |
-|:-------------:|:------:|:----:|:---------------:|
-| 1.168         | 0.7273 | 10   | 1.1995          |
 
 
 ### Framework versions