multi-dataset-model / generation_config.json
jkgl's picture
Training loss: 0.0000
78ed855 verified
{
"bos_token_id": 203,
"cache_implementation": "hybrid",
"do_sample": true,
"eos_token_id": [
212,
1,
106
],
"pad_token_id": 262143,
"top_k": 64,
"top_p": 0.95,
"transformers_version": "4.56.2"
}