gemma3-1b-moe / generation_config.json
Efe2898's picture
Dense → MoE dönüşümü (4 uzman, top-1, uzmanlar dense MLP'den init)
a0b4a25 verified
raw
history blame contribute delete
168 Bytes
{
"_from_model_config": true,
"bos_token_id": 2,
"cache_implementation": "hybrid",
"eos_token_id": 1,
"pad_token_id": 0,
"transformers_version": "4.57.1"
}