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/53e70ad11f6482d2_train_data.json
 model_type: AutoModelForCausalLM
@@ -91,7 +91,7 @@ wandb_name: 99fd57f3-0a4a-48ec-aa5d-f09a845afa53
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 99fd57f3-0a4a-48ec-aa5d-f09a845afa53
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -102,6 +102,8 @@ xformers_attention: null
 # a497d1bb-89b8-4c50-c57a-16e9c2996730
 
 This model is a fine-tuned version of [unsloth/codegemma-7b-it](https://huggingface.co/unsloth/codegemma-7b-it) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 1.7038
 
 ## 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        | 0.0000 | 1    | 3.4402          |
+| 2.927         | 0.0001 | 10   | 2.6244          |
+| 1.9578        | 0.0003 | 20   | 1.9394          |
+| 1.7791        | 0.0004 | 30   | 1.7751          |
+| 1.5244        | 0.0006 | 40   | 1.7151          |
+| 1.4187        | 0.0007 | 50   | 1.7038          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a79b04d917ec0d33d1db5370380d7b23f1d054c7bbbd6b617ad88f63169d400e
+oid sha256:ef2d0ba7f75e06be11c1ea909e972f146cd0ba4a093b944808d6f27e42c8d995
 size 100149034