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/f925ec5472bd6eac_train_data.json
 model_type: AutoModelForCausalLM
@@ -92,7 +92,7 @@ wandb_name: 9670d58f-4bf6-46f7-b76f-d44f3b5fb1f5
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 9670d58f-4bf6-46f7-b76f-d44f3b5fb1f5
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -103,6 +103,8 @@ xformers_attention: null
 # 90ef7d05-8c1c-d791-9ceb-32b890c9a01a
 
 This model is a fine-tuned version of [Korabbit/llama-2-ko-7b](https://huggingface.co/Korabbit/llama-2-ko-7b) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 1.6261
 
 ## Model description
 
@@ -129,14 +131,19 @@ 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        | 0.0002 | 1    | 1.8959          |
+| 1.8975        | 0.0021 | 10   | 1.7665          |
+| 1.6819        | 0.0042 | 20   | 1.6751          |
+| 1.6449        | 0.0063 | 30   | 1.6353          |
+| 1.6317        | 0.0084 | 40   | 1.6276          |
+| 1.6149        | 0.0105 | 50   | 1.6261          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b83dc66898ccc67fe27449a12b20724be4ed45ef20ebe46001bcf2bd2be9b5de
+oid sha256:9e6ed07f0b11b47e9db273025dccbd569a6b9df3fdb9ac7a7171d9e5d17bcf36
 size 80115210