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/4af16b378df8fa36_train_data.json
 model_type: AutoModelForCausalLM
@@ -91,7 +91,7 @@ wandb_name: b744d6fe-cea2-4360-b845-cd79f911ce96
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: b744d6fe-cea2-4360-b845-cd79f911ce96
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -102,6 +102,8 @@ xformers_attention: null
 # b744d6fe-cea2-4360-b845-cd79f911ce96
 
 This model is a fine-tuned version of [Eurdem/Defne_llama3_2x8B](https://huggingface.co/Eurdem/Defne_llama3_2x8B) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 1.0423
 
 ## Model description
 
@@ -128,14 +130,19 @@ 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        | 0.0006 | 1    | 8.3884          |
+| 6.6528        | 0.0057 | 10   | 5.3022          |
+| 1.3588        | 0.0114 | 20   | 1.4998          |
+| 0.9964        | 0.0172 | 30   | 1.1371          |
+| 0.9933        | 0.0229 | 40   | 1.0566          |
+| 1.3119        | 0.0286 | 50   | 1.0423          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:31e376004ca2abbd8b670aecd66cb33082e91641ea1a6a3ef0acf0bf72e0066d
+oid sha256:87b6e1ea5f2b9fcfe527cb5beb5387e0847d35aafa0b20faa55d2dcf832159d3
 size 144965450