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: 8
 mlflow_experiment_name: /tmp/2155d29cc40d725b_train_data.json
 model_type: AutoModelForCausalLM
@@ -91,7 +91,7 @@ wandb_name: 72886a10-7367-4561-9e00-37b78522209f
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 72886a10-7367-4561-9e00-37b78522209f
-warmup_steps: 1
+warmup_steps: 2
 weight_decay: 0.0
 xformers_attention: null
 
@@ -102,6 +102,8 @@ xformers_attention: null
 # 72886a10-7367-4561-9e00-37b78522209f
 
 This model is a fine-tuned version of [NousResearch/Hermes-2-Pro-Llama-3-8B](https://huggingface.co/NousResearch/Hermes-2-Pro-Llama-3-8B) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.0352
 
 ## Model description
 
@@ -129,13 +131,18 @@ The following hyperparameters were used during training:
 - 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
+- training_steps: 50
 
 ### Training results
 
 | Training Loss | Epoch  | Step | Validation Loss |
 |:-------------:|:------:|:----:|:---------------:|
 | No log        | 0.0005 | 1    | 17.9249         |
+| 11.6608       | 0.0054 | 10   | 1.8808          |
+| 0.4836        | 0.0109 | 20   | 0.0724          |
+| 0.0432        | 0.0163 | 30   | 0.0418          |
+| 0.0262        | 0.0217 | 40   | 0.0374          |
+| 0.0435        | 0.0271 | 50   | 0.0352          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4881e3e763d8ee27a4867551f3e297bac4852950af7f9919724dff2f48f0923b
+oid sha256:27adad5be2d2eca50592184b7b9e069d4e82e57c4271edd5d4fb79ef93462f7e
 size 84047370