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/2b51ee441a5edfeb_train_data.json
 model_type: AutoModelForCausalLM
@@ -91,7 +91,7 @@ wandb_name: 9cba1a2a-82d3-4f99-8411-8b9422aa8c46
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 9cba1a2a-82d3-4f99-8411-8b9422aa8c46
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -102,6 +102,8 @@ xformers_attention: null
 # 9cba1a2a-82d3-4f99-8411-8b9422aa8c46
 
 This model is a fine-tuned version of [unsloth/codegemma-2b](https://huggingface.co/unsloth/codegemma-2b) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 3.1557
 
 ## 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        | 0.0014 | 1    | 3.4073          |
+| 3.2007        | 0.0135 | 10   | 3.2740          |
+| 3.2742        | 0.0271 | 20   | 3.2074          |
+| 3.1028        | 0.0406 | 30   | 3.1700          |
+| 3.1737        | 0.0541 | 40   | 3.1572          |
+| 3.1068        | 0.0677 | 50   | 3.1557          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:68d0d6dc3638de9ee16dabffbdebe7d352ae730d655a238a0f821b8fe4c4a262
+oid sha256:d57ce0d6a2b4ea55c4472554507aabfe903475c1efcb683db3ca33e9ce173082
 size 39313978