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/9109af28f3122844_train_data.json
 model_type: AutoModelForCausalLM
@@ -90,7 +90,7 @@ wandb_name: 8a8721ba-3114-488c-b05c-7a87ee48ed64
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 8a8721ba-3114-488c-b05c-7a87ee48ed64
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -101,6 +101,8 @@ xformers_attention: null
 # 8a8721ba-3114-488c-b05c-7a87ee48ed64
 
 This model is a fine-tuned version of [oopsung/llama2-7b-koNqa-test-v1](https://huggingface.co/oopsung/llama2-7b-koNqa-test-v1) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 1.0898
 
 ## 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        | 0.0012 | 1    | 3.1607          |
+| 3.0908        | 0.0124 | 10   | 2.5644          |
+| 1.7327        | 0.0248 | 20   | 1.4430          |
+| 1.2256        | 0.0372 | 30   | 1.1802          |
+| 1.0754        | 0.0496 | 40   | 1.1031          |
+| 1.0947        | 0.0620 | 50   | 1.0898          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:71605d283c9dc33c41caddb9073c23f7749f92040cb920630dabea29f387dfe3
+oid sha256:98cf9e1b052d51bba26fb8a19e88ef80df2499ec750ad2e6177cb499e6a3d4db
 size 80115210