End of training

README.md

@@ -64,7 +64,7 @@ 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/859d718bfd5c530f_train_data.json
 model_type: AutoModelForCausalLM
@@ -89,7 +89,7 @@ wandb_name: 6b9d7473-77ed-43d6-bd32-41f610ee3c67
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 6b9d7473-77ed-43d6-bd32-41f610ee3c67
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -100,6 +100,8 @@ xformers_attention: null
 # 6b9d7473-77ed-43d6-bd32-41f610ee3c67
 
 This model is a fine-tuned version of [The-matt/llama2_ko-7b_distinctive-snowflake-182_1060](https://huggingface.co/The-matt/llama2_ko-7b_distinctive-snowflake-182_1060) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 1.3546
 
 ## Model description
 
@@ -126,14 +128,19 @@ 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        | 0.0010 | 1    | 1.4713          |
+| 1.4978        | 0.0099 | 10   | 1.4501          |
+| 1.4762        | 0.0199 | 20   | 1.3945          |
+| 1.4188        | 0.0298 | 30   | 1.3662          |
+| 1.3236        | 0.0398 | 40   | 1.3566          |
+| 1.3179        | 0.0497 | 50   | 1.3546          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:49a8d7f3416bfaee68b08ead854c8c99ecdd129e931641ceda63d3212113be7a
+oid sha256:f02fd57e822333fd1973dc77936c307bde599618663c513276daa07527953d62
 size 80115210