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/63d20f606a6e3ac1_train_data.json
 model_type: AutoModelForCausalLM
@@ -89,7 +89,7 @@ wandb_name: 0eb9c165-0d99-4436-b87a-c5018d237306
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 0eb9c165-0d99-4436-b87a-c5018d237306
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -100,6 +100,8 @@ xformers_attention: null
 # 0eb9c165-0d99-4436-b87a-c5018d237306
 
 This model is a fine-tuned version of [MNC-Jihun/Mistral-7B-AO-u0.5-b2-ver0.4](https://huggingface.co/MNC-Jihun/Mistral-7B-AO-u0.5-b2-ver0.4) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 1.3793
 
 ## Model description
 
@@ -126,14 +128,18 @@ 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: 9
 
 ### Training results
 
 | Training Loss | Epoch  | Step | Validation Loss |
 |:-------------:|:------:|:----:|:---------------:|
-| No log        | 0.1111 | 1    | 1.6721          |
+| No log        | 0.1111 | 1    | 1.5672          |
+| No log        | 0.2222 | 2    | 1.5624          |
+| No log        | 0.4444 | 4    | 1.5175          |
+| 1.3554        | 0.6667 | 6    | 1.4444          |
+| 1.3554        | 0.8889 | 8    | 1.3793          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:cc944d34c889951d58c1b6123b85966cbfb1f3ee530725301d03df662d7a3a57
+oid sha256:dabad84e4bbc544f2d0c39eae5db5255e45fbbb4ae02eb0a26004f1ba0700f1e
 size 608352466