Qwen3-4B-INT8 / quantization_info.json
Redtash1's picture
Upload 12 files
4fa5a00 verified
Raw History Blame Contribute Delete
352 Bytes
{
"base_model": "Qwen/Qwen3-4B",
"format": "Optimum-Quanto persistent",
"weights": "qint8",
"activations": "BF16",
"linear_scope": "all Linear weights including lm_head",
"pad_token_id": 151643,
"eos_token_id": [
151645,
151643
],
"bos_token_id": 151643,
"torch": "2.12.0+cu130",
"transformers": "5.17.0"
}