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/0810c6b2551460e9_train_data.json
 model_type: AutoModelForCausalLM
@@ -93,7 +93,7 @@ wandb_name: 18ed554f-aed2-448a-bb28-bdaaf2ad3ab5
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 18ed554f-aed2-448a-bb28-bdaaf2ad3ab5
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -104,6 +104,8 @@ xformers_attention: null
 # 18ed554f-aed2-448a-bb28-bdaaf2ad3ab5
 
 This model is a fine-tuned version of [TinyLlama/TinyLlama_v1.1](https://huggingface.co/TinyLlama/TinyLlama_v1.1) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.1328
 
 ## 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        | 0.0004 | 1    | 0.2146          |
+| 0.1748        | 0.0038 | 10   | 0.1967          |
+| 0.2802        | 0.0075 | 20   | 0.1544          |
+| 0.1406        | 0.0113 | 30   | 0.1396          |
+| 0.2164        | 0.0150 | 40   | 0.1337          |
+| 0.125         | 0.0188 | 50   | 0.1328          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e292a300aea3c9edba2ca44cd0cc75c004466ec053c97e7b4e4892db88e80bf0
+oid sha256:8919aa6943187d7a4a935f341758c9f6322b13dedf2a642504636dbe64400073
 size 25342042