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: 50
+max_steps: 1
 micro_batch_size: 2
 mlflow_experiment_name: /tmp/f8b665cd3d4206f3_train_data.json
 model_type: AutoModelForCausalLM
@@ -90,7 +90,7 @@ wandb_name: 68c4aa27-69a3-4514-89b3-e4b0c68c9653
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 68c4aa27-69a3-4514-89b3-e4b0c68c9653
-warmup_steps: 10
+warmup_steps: 1
 weight_decay: 0.0
 xformers_attention: null
 
@@ -101,8 +101,6 @@ xformers_attention: null
 # 8def7c39-e3ea-e983-2346-efa85e89c2a5
 
 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: 0.9816
 
 ## Model description
 
@@ -129,19 +127,14 @@ 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: 10
-- training_steps: 50
+- lr_scheduler_warmup_steps: 2
+- training_steps: 1
 
 ### Training results
 
 | Training Loss | Epoch  | Step | Validation Loss |
 |:-------------:|:------:|:----:|:---------------:|
 | No log        | 0.0002 | 1    | 1.5773          |
-| 1.5554        | 0.0021 | 10   | 1.4865          |
-| 1.2781        | 0.0043 | 20   | 1.1905          |
-| 1.0853        | 0.0064 | 30   | 1.0309          |
-| 1.0531        | 0.0085 | 40   | 0.9869          |
-| 1.037         | 0.0107 | 50   | 0.9816          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:dbf09fdf98cc60e75213aca60cfd1c82fc1a4e8465d02e71f396c7519637d701
+oid sha256:7ef01d5362c33adbcad460588b47576d920959de8f0e6ace078fbac0970f58d1
 size 80115210