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/932d49329217bf11_train_data.json
 model_type: AutoModelForCausalLM
@@ -93,7 +93,7 @@ wandb_name: 62684fea-5f30-41df-a276-fa59564c2fe2
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 62684fea-5f30-41df-a276-fa59564c2fe2
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -104,6 +104,8 @@ xformers_attention: null
 # 62684fea-5f30-41df-a276-fa59564c2fe2
 
 This model is a fine-tuned version of [NousResearch/Llama-3.2-1B](https://huggingface.co/NousResearch/Llama-3.2-1B) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.1227
 
 ## 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        | 0.0003 | 1    | 0.4408          |
+| 0.3812        | 0.0026 | 10   | 0.3603          |
+| 0.2045        | 0.0051 | 20   | 0.1829          |
+| 0.1247        | 0.0077 | 30   | 0.1344          |
+| 0.138         | 0.0102 | 40   | 0.1241          |
+| 0.1025        | 0.0128 | 50   | 0.1227          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:93196542b8ba5dde4018b5ef421e534e24d6a20ce9df07f068a239507c81cfa1
+oid sha256:312eec540b5758ba42cf495180d9b388b25cd04ee6c448a18e70e02b1dd4d867
 size 22624970