qwen-short-thinking / config.json
volcanos's picture
Upload model
e7b9498 verified
raw
history blame contribute delete
856 Bytes
{
"_name_or_path": "/mnt/llm-train/users/wucanhui.volcano/output/DeepSeek-R1-Distill-Qwen-7B/v25-20250515-162009/checkpoint-14288",
"architectures": [
"Qwen2Model"
],
"attention_dropout": 0.0,
"bos_token_id": 151643,
"eos_token_id": 151643,
"hidden_act": "silu",
"hidden_size": 3584,
"initializer_range": 0.02,
"intermediate_size": 18944,
"max_position_embeddings": 131072,
"max_window_layers": 28,
"model_type": "qwen2",
"num_attention_heads": 28,
"num_hidden_layers": 28,
"num_key_value_heads": 4,
"pad_token_id": 151643,
"rms_norm_eps": 1e-06,
"rope_scaling": null,
"rope_theta": 10000,
"sliding_window": null,
"tie_word_embeddings": false,
"torch_dtype": "float32",
"transformers_version": "4.48.0",
"use_cache": false,
"use_mrope": false,
"use_sliding_window": false,
"vocab_size": 152064
}