Qwen2.5-3B / config.json
Geaming's picture
Update config.json to align with tokenizer_config.json and other Qwen2.5 models
2359650 verified
raw
history blame
685 Bytes
{
"architectures": [
"Qwen2ForCausalLM"
],
"attention_dropout": 0.0,
"bos_token_id": 151643,
"eos_token_id": 151643,
"hidden_act": "silu",
"hidden_size": 2048,
"initializer_range": 0.02,
"intermediate_size": 11008,
"max_position_embeddings": 131072,
"max_window_layers": 36,
"model_type": "qwen2",
"num_attention_heads": 16,
"num_hidden_layers": 36,
"num_key_value_heads": 2,
"rms_norm_eps": 1e-06,
"rope_theta": 1000000.0,
"sliding_window": 131072,
"tie_word_embeddings": true,
"torch_dtype": "bfloat16",
"transformers_version": "4.40.1",
"use_cache": true,
"use_mrope": false,
"use_sliding_window": false,
"vocab_size": 151936
}