starling-lora

This model is a fine-tuned version of Nexusflow/Starling-LM-7B-beta on the generator dataset. It achieves the following results on the evaluation set:

• Loss: 0.1048

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0003
• train_batch_size: 8
• eval_batch_size: 8
• seed: 1399
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• optimizer: Use adamw_torch 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: 100
• training_steps: 500

Training results

Training Loss	Epoch	Step	Validation Loss
6.416	2.7586	20	1.1396
2.3556	5.5172	40	0.1913
0.28	8.2759	60	0.0993
0.1179	11.0345	80	0.1048

Framework versions

• PEFT 0.13.2
• Transformers 4.46.1
• Pytorch 2.5.0+cu121
• Datasets 3.1.0
• Tokenizers 0.20.1