End of training

README.md

@@ -65,7 +65,7 @@ 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/50a3279bbe0d6c57_train_data.json
 model_type: AutoModelForCausalLM
@@ -90,7 +90,7 @@ wandb_name: b1558c1d-92f1-4acc-948a-bb5d77f97503
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: b1558c1d-92f1-4acc-948a-bb5d77f97503
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -101,6 +101,8 @@ xformers_attention: null
 # b674cfc8-d0a7-94a7-52df-8ce4be32eb9c
 
 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: 2.2525
 
 ## Model description
 
@@ -127,14 +129,19 @@ 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        | 0.0001 | 1    | 2.5663          |
+| 1.518         | 0.0013 | 10   | 2.4846          |
+| 1.6358        | 0.0025 | 20   | 2.3336          |
+| 1.8006        | 0.0038 | 30   | 2.2754          |
+| 1.5463        | 0.0051 | 40   | 2.2546          |
+| 1.9366        | 0.0064 | 50   | 2.2525          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6dd48165e4026e5ad75b692bb03a9812cfd8c038ea9238c991953284a4247500
+oid sha256:f055c1f133cd6879532029fc6fa9cff38faa35bf0d28ac7023dc5970a877cd8f
 size 48768810