Model save

README.md

@@ -1,11 +1,9 @@
 ---
 library_name: transformers
-license: mit
-base_model: microsoft/phi-4
+license: apache-2.0
+base_model: Qwen/Qwen2.5-32B
 tags:
 - llama-factory
-- full
-- lora
 - generated_from_trainer
 model-index:
 - name: pretrain
@@ -17,7 +15,9 @@ should probably proofread and complete it, then remove this comment. -->
 
 # pretrain
 
-This model is a fine-tuned version of [microsoft/phi-4](https://huggingface.co/microsoft/phi-4) on the openalex_small dataset.
+This model is a fine-tuned version of [Qwen/Qwen2.5-32B](https://huggingface.co/Qwen/Qwen2.5-32B) on an unknown dataset.
+It achieves the following results on the evaluation set:
+- Loss: 1.1995
 
 ## Model description
 
@@ -38,13 +38,13 @@ More information needed
 The following hyperparameters were used during training:
 - learning_rate: 0.0001
 - train_batch_size: 1
-- eval_batch_size: 8
+- eval_batch_size: 1
 - seed: 42
 - distributed_type: multi-GPU
 - num_devices: 8
 - gradient_accumulation_steps: 8
 - total_train_batch_size: 64
-- total_eval_batch_size: 64
+- total_eval_batch_size: 8
 - 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
@@ -52,6 +52,9 @@ The following hyperparameters were used during training:
 
 ### Training results
 
+| Training Loss | Epoch  | Step | Validation Loss |
+|:-------------:|:------:|:----:|:---------------:|
+| 1.168         | 0.7273 | 10   | 1.1995          |
 
 
 ### Framework versions

generation_config.json

@@ -1,9 +1,6 @@
 {
-  "_from_model_config": true,
-  "bos_token_id": 100257,
-  "eos_token_id": [
-    100257,
-    100265
-  ],
+  "bos_token_id": 151643,
+  "eos_token_id": 151643,
+  "max_new_tokens": 2048,
   "transformers_version": "4.48.1"
 }