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/1b7bca7b022b0b6c_train_data.json
 model_type: AutoModelForCausalLM
@@ -93,7 +93,7 @@ wandb_name: 51da049f-53af-4e0d-b40d-ba77cac13359
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 51da049f-53af-4e0d-b40d-ba77cac13359
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -104,6 +104,8 @@ xformers_attention: null
 # 51da049f-53af-4e0d-b40d-ba77cac13359
 
 This model is a fine-tuned version of [princeton-nlp/Sheared-LLaMA-1.3B](https://huggingface.co/princeton-nlp/Sheared-LLaMA-1.3B) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.1783
 
 ## 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        | 0.0005 | 1    | 2.7945          |
+| 2.508         | 0.0053 | 10   | 2.3117          |
+| 1.1545        | 0.0106 | 20   | 0.6815          |
+| 0.387         | 0.0159 | 30   | 0.2443          |
+| 0.173         | 0.0213 | 40   | 0.1826          |
+| 0.1005        | 0.0266 | 50   | 0.1783          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4e960b775f551fec311147dead4c107e7697e7b4744d2fd613c92ec17ef0c044
+oid sha256:5e46cfc45963b1442e227b6f996fab5cf1728ab67232505ac6962cf7ac73b1f4
 size 30103498