CMB-Qwen3-32B-Eagle3 / config.json
CMBTech's picture
Upload 4 files
253b7cc verified
raw
history blame contribute delete
740 Bytes
{
"architectures": [
"LlamaForCausalLMEagle3"
],
"bos_token_id": 151643,
"draft_vocab_size": 38000,
"eos_token_id": 151645,
"head_dim": 128,
"hidden_act": "silu",
"hidden_size": 5120,
"initializer_range": 0.02,
"intermediate_size": 25600,
"model_type": "llama",
"max_position_embeddings": 40960,
"num_attention_heads": 64,
"num_hidden_layers": 1,
"num_key_value_heads": 8,
"pad_token_id": 151645,
"pretraining_tp": 1,
"rms_norm_eps": 1e-06,
"rope_scaling": null,
"target_model_name_or_path": "/model/Qwen3-32B/",
"template": "qwen3",
"tie_word_embeddings": false,
"torch_dtype": "bfloat16",
"transformers_version": "4.51.3",
"ttt_len": 7,
"use_cache": false,
"vocab_size": 151936
}