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/09d582947cc2cea8_train_data.json
 model_type: AutoModelForCausalLM
@@ -93,7 +93,7 @@ wandb_name: e1e5bed9-0914-413d-b70e-d6822e7d6307
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: e1e5bed9-0914-413d-b70e-d6822e7d6307
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -104,6 +104,8 @@ xformers_attention: null
 # 42c9bfbe-db59-40fb-b4ad-a2f20c79748f
 
 This model is a fine-tuned version of [heegyu/WizardVicuna-open-llama-3b-v2](https://huggingface.co/heegyu/WizardVicuna-open-llama-3b-v2) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.0825
 
 ## Model description
 
@@ -130,14 +132,19 @@ 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        | 0.0009 | 1    | 0.8622          |
+| 0.4659        | 0.0086 | 10   | 0.6808          |
+| 0.1644        | 0.0172 | 20   | 0.2434          |
+| 0.0903        | 0.0258 | 30   | 0.1078          |
+| 0.0618        | 0.0344 | 40   | 0.0859          |
+| 0.1332        | 0.0430 | 50   | 0.0825          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:90d2d41c7e5899a85a9ced98963619f34fef982a1dd6ca956c12613901348e14
+oid sha256:cba33559e0f7c550532be02ee51b84f6803d26635a1cace30ba86633f5eebd6e
 size 50982842