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/6df7e1fa95dbebf2_train_data.json
 model_type: AutoModelForCausalLM
@@ -93,7 +93,7 @@ wandb_name: 4b0fd424-af59-4f0e-9e92-7a6ec6b9ded8
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 4b0fd424-af59-4f0e-9e92-7a6ec6b9ded8
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -104,6 +104,8 @@ xformers_attention: null
 # 4eb29a4f-1942-4d31-a40c-f3f2ef1059f8
 
 This model is a fine-tuned version of [defog/llama-3-sqlcoder-8b](https://huggingface.co/defog/llama-3-sqlcoder-8b) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.0755
 
 ## Model description
 
@@ -130,14 +132,19 @@ 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    | 3.4576          |
+| 2.7027        | 0.0002 | 10   | 1.6180          |
+| 0.2079        | 0.0004 | 20   | 0.1040          |
+| 0.1313        | 0.0006 | 30   | 0.1009          |
+| 0.0817        | 0.0009 | 40   | 0.0756          |
+| 0.0739        | 0.0011 | 50   | 0.0755          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e3651eb81d0956a09a9f4c9d76b4781932954dce486d36ef826996e8576a0356
+oid sha256:47a3adcfc38ac6465a43a2f8a5579d2413148fd78e66e26972a69d0c2c4ef1d2
 size 84047370