llama3-8b-instuct-finetuned-gptq-8bit / generation_config.json
belyakoff's picture
Upload LlamaForCausalLM
40e2d62 verified
raw
history blame
277 Bytes
{
"_from_model_config": true,
"bos_token_id": 128000,
"do_sample": true,
"eos_token_id": 128009,
"max_new_tokens": 1536,
"pad_token_id": 128000,
"repetition_penalty": 1.12,
"temperature": 0.2,
"top_k": 30,
"top_p": 0.9,
"transformers_version": "4.44.2"
}