DeepSeek-V4-Flash-W4A16-FP8 / generation_config.json
pastapaul's picture
Phase 3b: AWQ-W4A16 quantization (FP8_BLOCK attn + W4A16 routed experts)
2e7ef6a verified
raw
history blame contribute delete
174 Bytes
{
"_from_model_config": true,
"bos_token_id": 0,
"do_sample": true,
"eos_token_id": 1,
"temperature": 1.0,
"top_p": 1.0,
"transformers_version": "5.8.0.dev0"
}