End of training

README.md

@@ -65,7 +65,7 @@ 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/d0ac1e57101ade22_train_data.json
 model_type: AutoModelForCausalLM
@@ -92,7 +92,7 @@ wandb_name: e18851da-cd5b-4125-afca-cf14d857ac9f
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: e18851da-cd5b-4125-afca-cf14d857ac9f
-warmup_steps: 1
+warmup_steps: 2
 weight_decay: 0.0
 xformers_attention: null
 
@@ -103,6 +103,8 @@ xformers_attention: null
 # e18851da-cd5b-4125-afca-cf14d857ac9f
 
 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: 0.1711
 
 ## Model description
 
@@ -130,13 +132,17 @@ 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: 29
 
 ### Training results
 
 | Training Loss | Epoch  | Step | Validation Loss |
 |:-------------:|:------:|:----:|:---------------:|
 | No log        | 0.0354 | 1    | 1.2406          |
+| No log        | 0.2124 | 6    | 0.9171          |
+| 0.9919        | 0.4248 | 12   | 0.4827          |
+| 0.9919        | 0.6372 | 18   | 0.2492          |
+| 0.3935        | 0.8496 | 24   | 0.1711          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6887af6cbe66858684cca40287dd683451b74578565c144870ac41b4de83d171
+oid sha256:d57d673db61a85c88324b25a6f0b7fa1e40edc6ac9d608008237fc2c880e919b
 size 22624970