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/ce7fcd2d05dffaef_train_data.json
 model_type: AutoModelForCausalLM
@@ -91,7 +91,7 @@ wandb_name: 54f52b18-f019-41eb-b70b-23aa1dcdada5
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 54f52b18-f019-41eb-b70b-23aa1dcdada5
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -102,6 +102,8 @@ xformers_attention: null
 # f1ba572f-11bb-45e3-b8f0-29989c2f106d
 
 This model is a fine-tuned version of [Qwen/Qwen2.5-0.5B-Instruct](https://huggingface.co/Qwen/Qwen2.5-0.5B-Instruct) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.0382
 
 ## 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        | 0.0032 | 1    | 1.3815          |
+| 0.7314        | 0.0323 | 10   | 0.7887          |
+| 0.2222        | 0.0647 | 20   | 0.1924          |
+| 0.0518        | 0.0970 | 30   | 0.0539          |
+| 0.035         | 0.1293 | 40   | 0.0396          |
+| 0.039         | 0.1617 | 50   | 0.0382          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:cfcb06d307edfc4c6d4b2c1b1e252368f5789e0ab132eecf82440c0c8ab9057d
+oid sha256:4ba8fac6c9fc3e4a079c2b09d7f1195efe7c1b2e2c8874f173ad86fe901efd23
 size 17717130