End of training

README.md

@@ -66,7 +66,7 @@ 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/bbf0c4108ff67641_train_data.json
 model_type: AutoModelForCausalLM
@@ -93,7 +93,7 @@ wandb_name: 32437e7b-da19-44d7-8b1b-a999a85ad06f
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 32437e7b-da19-44d7-8b1b-a999a85ad06f
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -104,6 +104,8 @@ xformers_attention: null
 # 73f3fabd-9900-7f2f-a011-788b4a8a4e7a
 
 This model is a fine-tuned version of [NousResearch/Yarn-Mistral-7b-128k](https://huggingface.co/NousResearch/Yarn-Mistral-7b-128k) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.7039
 
 ## Model description
 
@@ -130,14 +132,19 @@ 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        | 0.0000 | 1    | 0.9515          |
+| 3.4405        | 0.0002 | 10   | 0.8283          |
+| 2.8482        | 0.0005 | 20   | 0.7552          |
+| 3.1534        | 0.0007 | 30   | 0.7181          |
+| 2.943         | 0.0010 | 40   | 0.7067          |
+| 2.7567        | 0.0012 | 50   | 0.7039          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:dda37f87dce8bcb688dfb91717cbbcf2ec1db196ca46782bb8b5ce175512c8f6
+oid sha256:027df755e29009bfa84a12faa1f1bbbc41b4d96fb404c3313691af1e5a10ddff
 size 84047370