End of training

README.md

@@ -66,7 +66,7 @@ lora_model_dir: null
 lora_r: 8
 lora_target_linear: true
 lr_scheduler: cosine
-max_steps: 1
+max_steps: 50
 micro_batch_size: 2
 mlflow_experiment_name: /tmp/c5f8dabade462765_train_data.json
 model_type: AutoModelForCausalLM
@@ -91,7 +91,7 @@ wandb_name: dfefc67f-0c65-4962-aa26-15e73510675b
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: dfefc67f-0c65-4962-aa26-15e73510675b
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -102,6 +102,8 @@ xformers_attention: null
 # 07e427c6-0889-4a3c-9572-49cf69139c40
 
 This model is a fine-tuned version of [Qwen/Qwen2.5-3B-Instruct](https://huggingface.co/Qwen/Qwen2.5-3B-Instruct) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 2.2930
 
 ## Model description
 
@@ -128,14 +130,19 @@ The following hyperparameters were used during training:
 - total_train_batch_size: 8
 - optimizer: Use OptimizerNames.ADAMW_BNB with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
 - lr_scheduler_type: cosine
-- lr_scheduler_warmup_steps: 2
-- training_steps: 1
+- lr_scheduler_warmup_steps: 10
+- training_steps: 50
 
 ### Training results
 
 | Training Loss | Epoch  | Step | Validation Loss |
 |:-------------:|:------:|:----:|:---------------:|
 | No log        | 0.0003 | 1    | 8.1168          |
+| 7.6616        | 0.0031 | 10   | 7.3319          |
+| 5.2856        | 0.0061 | 20   | 4.4848          |
+| 2.5862        | 0.0092 | 30   | 2.7336          |
+| 2.0909        | 0.0123 | 40   | 2.3122          |
+| 2.1409        | 0.0154 | 50   | 2.2930          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:59670834e118354991b53e46e89c82674cde01822fc5330a67b08a481892b281
+oid sha256:77d7d9d3346e6202840817c9c940621134cf0ae1939419eb54a924f6b207b560
 size 60048618