Qwen3.8-27B-NVFP4 / generation_config.json
JustinTong's picture
Qwen3.8-27B NVFP4 (W4A4, group size 16) quantized with NVIDIA Model Optimizer
52d1adc
Raw History Blame Contribute Delete
214 Bytes
{
"bos_token_id": 248044,
"do_sample": true,
"eos_token_id": [
248046,
248044
],
"pad_token_id": 248044,
"temperature": 1.0,
"top_k": 20,
"top_p": 0.95,
"transformers_version": "5.12.1"
}