generation_config.json

{
 "bos_token_id": 1,
 "eos_token_id": 2,
 "max_length": 4096,
 "max_memory": {
 "cpu": 129503662080
 },
 "no_split_module_classes": [
 "LlamaDecoderLayer"
 ],
 "pad_token_id": 0,
 "special_dtypes": {},
 "temperature": 0.9,
 "top_p": 0.6,
 "transformers_version": "4.31.0"
}