Files
kanana-2-1.3b-instruct/sglang/config.json

74 lines
1.7 KiB
JSON
Raw Normal View History

{
"architectures": [
"Qwen3ForCausalLM"
],
"attention_bias": false,
"attention_dropout": 0.0,
"dtype": "bfloat16",
"head_dim": 128,
"hidden_act": "silu",
"hidden_size": 1280,
"initializer_range": 0.02,
"intermediate_size": 5760,
"layer_types": [
"sliding_attention",
"sliding_attention",
"sliding_attention",
"full_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"full_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"full_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"full_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"full_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"full_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"full_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"full_attention"
],
"max_position_embeddings": 32768,
"max_window_layers": 32,
"model_type": "qwen3",
"num_attention_heads": 32,
"num_hidden_layers": 32,
"num_key_value_heads": 8,
"rms_norm_eps": 1e-06,
"rope_theta": null,
"rope_parameters": {
"full_attention": {
"rope_type": "yarn",
"rope_theta": 10000,
"factor": 40.0,
"original_max_position_embeddings": 4096
},
"sliding_attention": {
"rope_type": "default",
"rope_theta": 10000.0
}
},
"sliding_window": 1024,
"tie_word_embeddings": true,
"transformers_version": "4.57.6",
"use_cache": true,
"use_sliding_window": true,
"vocab_size": 128256
}