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: 50
+max_steps: 1
 micro_batch_size: 2
 mlflow_experiment_name: /tmp/4eba0df0a656dd91_train_data.json
 model_type: AutoModelForCausalLM
@@ -93,7 +93,7 @@ wandb_name: 4e5efe2f-5960-4644-9e88-198303f5d2db
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 4e5efe2f-5960-4644-9e88-198303f5d2db
-warmup_steps: 10
+warmup_steps: 1
 weight_decay: 0.0
 xformers_attention: null
 
@@ -104,8 +104,6 @@ xformers_attention: null
 # 4e5efe2f-5960-4644-9e88-198303f5d2db
 
 This model is a fine-tuned version of [elyza/Llama-3-ELYZA-JP-8B](https://huggingface.co/elyza/Llama-3-ELYZA-JP-8B) on the None dataset.
-It achieves the following results on the evaluation set:
-- Loss: 0.0094
 
 ## Model description
 
@@ -132,19 +130,14 @@ 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: 10
-- training_steps: 50
+- lr_scheduler_warmup_steps: 2
+- training_steps: 1
 
 ### Training results
 
 | Training Loss | Epoch  | Step | Validation Loss |
 |:-------------:|:------:|:----:|:---------------:|
 | No log        | 0.0006 | 1    | 0.7707          |
-| 0.5867        | 0.0064 | 10   | 0.3562          |
-| 0.1011        | 0.0127 | 20   | 0.0508          |
-| 0.02          | 0.0191 | 30   | 0.0161          |
-| 0.0121        | 0.0254 | 40   | 0.0101          |
-| 0.0093        | 0.0318 | 50   | 0.0094          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2ee03bb78c5e295f7d4f8a40dc8d9ee53d28ea200a5fa116a4994f30d76ec8c7
+oid sha256:cf78820d10097bbd51117b3f2f36e705a68ac71fbce6e4f007acbb7a83023374
 size 84047370