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/2a8257b8d6322416_train_data.json
 model_type: AutoModelForCausalLM
@@ -93,7 +93,7 @@ wandb_name: 90080c80-be06-4907-93f7-786a9d69a3ab
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 90080c80-be06-4907-93f7-786a9d69a3ab
-warmup_steps: 1
+warmup_steps: 2
 weight_decay: 0.0
 xformers_attention: null
 
@@ -104,6 +104,8 @@ xformers_attention: null
 # 90080c80-be06-4907-93f7-786a9d69a3ab
 
 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: 0.8674
 
 ## Model description
 
@@ -131,13 +133,18 @@ 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        | 0.0062 | 1    | 1.4452          |
+| 1.2914        | 0.0621 | 10   | 1.1288          |
+| 0.9854        | 0.1242 | 20   | 0.9413          |
+| 0.9005        | 0.1863 | 30   | 0.8873          |
+| 0.8699        | 0.2484 | 40   | 0.8705          |
+| 0.8423        | 0.3104 | 50   | 0.8674          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c67871b12112711d44f88d437ef754601ed3fa74c4a3ce06ab5290ef94c10f81
+oid sha256:a6e89c01861d10ff07f851f633fed151b8d92d35eee7becacb2547633c1e9882
 size 22624970