End of training

README.md

@@ -64,7 +64,7 @@ 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/ce18f8956f8ec3e2_train_data.json
 model_type: AutoModelForCausalLM
@@ -91,7 +91,7 @@ wandb_name: 58629422-fa9c-4ade-b34e-61b776fa7977
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 58629422-fa9c-4ade-b34e-61b776fa7977
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -102,6 +102,8 @@ xformers_attention: null
 # 58629422-fa9c-4ade-b34e-61b776fa7977
 
 This model is a fine-tuned version of [fxmarty/tiny-llama-fast-tokenizer](https://huggingface.co/fxmarty/tiny-llama-fast-tokenizer) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 10.3825
 
 ## 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        | 0.0009 | 1    | 10.3929         |
+| 10.3839       | 0.0086 | 10   | 10.3913         |
+| 10.3906       | 0.0172 | 20   | 10.3875         |
+| 10.3864       | 0.0258 | 30   | 10.3844         |
+| 10.3852       | 0.0344 | 40   | 10.3828         |
+| 10.3877       | 0.0430 | 50   | 10.3825         |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8d219ad721595fff5b6e0c4173384ffc39d8f588817186c4650717e96e4aadaa
+oid sha256:a949f59b8e68204dc2ca0738b52b2d73729edceb9e9a71f7b5745baba9e935e7
 size 33666