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/eb79f09081be9070_train_data.json
 model_type: AutoModelForCausalLM
@@ -93,7 +93,7 @@ wandb_name: 3d83e1fb-02b6-4f8d-9b0d-75029a42d1b3
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 3d83e1fb-02b6-4f8d-9b0d-75029a42d1b3
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -104,6 +104,8 @@ xformers_attention: null
 # 892dec20-6c13-490d-a512-3f1542747041
 
 This model is a fine-tuned version of [echarlaix/tiny-random-PhiForCausalLM](https://huggingface.co/echarlaix/tiny-random-PhiForCausalLM) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 6.9233
 
 ## 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        | 0.0014 | 1    | 6.9305          |
+| 6.9288        | 0.0142 | 10   | 6.9293          |
+| 6.9273        | 0.0283 | 20   | 6.9268          |
+| 6.9239        | 0.0425 | 30   | 6.9247          |
+| 6.9217        | 0.0566 | 40   | 6.9235          |
+| 6.9204        | 0.0708 | 50   | 6.9233          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:def40d9042cec45563d434c9c1a950291a4f0c22bc07bd75402613c845ce400d
+oid sha256:750f198f8bd8233aa4a7403f07a133c406cbf64e0ccff9827dc7a0bad6989295
 size 168063