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/930476505defbe50_train_data.json
 model_type: AutoModelForCausalLM
@@ -91,7 +91,7 @@ wandb_name: 3d4af4ca-ec53-4915-abfb-fc3df4cc2f94
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 3d4af4ca-ec53-4915-abfb-fc3df4cc2f94
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -102,6 +102,8 @@ xformers_attention: null
 # d439cab1-c6e2-40fc-9fba-e2f7a4df826a
 
 This model is a fine-tuned version of [unsloth/Llama-3.2-3B](https://huggingface.co/unsloth/Llama-3.2-3B) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.0781
 
 ## 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    | 0.1393          |
+| 0.1181        | 0.0003 | 10   | 0.1334          |
+| 0.0994        | 0.0006 | 20   | 0.1048          |
+| 0.0904        | 0.0008 | 30   | 0.0881          |
+| 0.0634        | 0.0011 | 40   | 0.0796          |
+| 0.1125        | 0.0014 | 50   | 0.0781          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a31642337157fb0a0a5c55ac8c09130be4abff8c47d320e060ec0eb878593f08
+oid sha256:19e6232c75cffe050a825b7e8f8927df6e41d97645eb80340caca2174be5e88b
 size 48768810