Teddysum/bllossom-Llama-2-13b-chat-hf-lima-ko-adapter-4bit

test

This model is a fine-tuned version of meta-llama/Llama-2-13b-chat-hf on LIMA dataset(translated in Koean)

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0002
• train_batch_size: 4
• eval_batch_size: 1
• seed: 0
• gradient_accumulation_steps: 16
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: constant
• lr_scheduler_warmup_ratio: 0.03
• training_steps: 101(6.34 epoch)

Framework versions

• Transformers 4.31.0.dev0
• Pytorch 1.13.1
• Datasets 2.13.1
• Tokenizers 0.13.3