Upload config.json to GDN-distill
Browse files- config.json +167 -0
config.json
ADDED
|
@@ -0,0 +1,167 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"Qwen3NextForCausalLM"
|
| 4 |
+
],
|
| 5 |
+
"attention_bias": true,
|
| 6 |
+
"attention_dropout": 0.0,
|
| 7 |
+
"attn_output_gate": false,
|
| 8 |
+
"auto_map": {
|
| 9 |
+
"AutoConfig": "configuration_qwen3_next.Qwen3NextConfig",
|
| 10 |
+
"AutoModel": "modeling_qwen3_next.Qwen3NextForCausalLM",
|
| 11 |
+
"AutoModelForCausalLM": "modeling_qwen3_next.Qwen3NextForCausalLM"
|
| 12 |
+
},
|
| 13 |
+
"bos_token_id": 1,
|
| 14 |
+
"decoder_sparse_step": 1,
|
| 15 |
+
"dtype": "float32",
|
| 16 |
+
"enable_qk_norm": false,
|
| 17 |
+
"eos_token_id": 2,
|
| 18 |
+
"full_attention_interval": 0,
|
| 19 |
+
"head_dim": 64,
|
| 20 |
+
"hidden_act": "silu",
|
| 21 |
+
"hidden_size": 1920,
|
| 22 |
+
"initializer_range": 0.02,
|
| 23 |
+
"intermediate_size": 4800,
|
| 24 |
+
"layer_types": [
|
| 25 |
+
"full_attention",
|
| 26 |
+
"linear_attention",
|
| 27 |
+
"linear_attention",
|
| 28 |
+
"linear_attention",
|
| 29 |
+
"linear_attention",
|
| 30 |
+
"linear_attention",
|
| 31 |
+
"linear_attention",
|
| 32 |
+
"linear_attention",
|
| 33 |
+
"full_attention",
|
| 34 |
+
"linear_attention",
|
| 35 |
+
"linear_attention",
|
| 36 |
+
"linear_attention",
|
| 37 |
+
"linear_attention",
|
| 38 |
+
"linear_attention",
|
| 39 |
+
"linear_attention",
|
| 40 |
+
"linear_attention",
|
| 41 |
+
"full_attention",
|
| 42 |
+
"linear_attention",
|
| 43 |
+
"linear_attention",
|
| 44 |
+
"linear_attention",
|
| 45 |
+
"linear_attention",
|
| 46 |
+
"linear_attention",
|
| 47 |
+
"linear_attention",
|
| 48 |
+
"linear_attention",
|
| 49 |
+
"full_attention",
|
| 50 |
+
"linear_attention",
|
| 51 |
+
"linear_attention",
|
| 52 |
+
"linear_attention",
|
| 53 |
+
"linear_attention",
|
| 54 |
+
"linear_attention",
|
| 55 |
+
"linear_attention",
|
| 56 |
+
"linear_attention",
|
| 57 |
+
"full_attention",
|
| 58 |
+
"linear_attention",
|
| 59 |
+
"linear_attention",
|
| 60 |
+
"linear_attention",
|
| 61 |
+
"linear_attention",
|
| 62 |
+
"linear_attention",
|
| 63 |
+
"linear_attention",
|
| 64 |
+
"linear_attention",
|
| 65 |
+
"full_attention",
|
| 66 |
+
"linear_attention",
|
| 67 |
+
"linear_attention",
|
| 68 |
+
"linear_attention",
|
| 69 |
+
"linear_attention",
|
| 70 |
+
"linear_attention",
|
| 71 |
+
"linear_attention",
|
| 72 |
+
"linear_attention",
|
| 73 |
+
"full_attention",
|
| 74 |
+
"linear_attention",
|
| 75 |
+
"linear_attention",
|
| 76 |
+
"linear_attention",
|
| 77 |
+
"linear_attention",
|
| 78 |
+
"linear_attention",
|
| 79 |
+
"linear_attention",
|
| 80 |
+
"linear_attention"
|
| 81 |
+
],
|
| 82 |
+
"linear_conv_kernel_dim": 4,
|
| 83 |
+
"linear_key_head_dim": 64,
|
| 84 |
+
"linear_num_key_heads": 8,
|
| 85 |
+
"linear_num_value_heads": 32,
|
| 86 |
+
"linear_value_head_dim": 64,
|
| 87 |
+
"max_position_embeddings": 32768,
|
| 88 |
+
"mlp_only_layers": [
|
| 89 |
+
0,
|
| 90 |
+
1,
|
| 91 |
+
2,
|
| 92 |
+
3,
|
| 93 |
+
4,
|
| 94 |
+
5,
|
| 95 |
+
6,
|
| 96 |
+
7,
|
| 97 |
+
8,
|
| 98 |
+
9,
|
| 99 |
+
10,
|
| 100 |
+
11,
|
| 101 |
+
12,
|
| 102 |
+
13,
|
| 103 |
+
14,
|
| 104 |
+
15,
|
| 105 |
+
16,
|
| 106 |
+
17,
|
| 107 |
+
18,
|
| 108 |
+
19,
|
| 109 |
+
20,
|
| 110 |
+
21,
|
| 111 |
+
22,
|
| 112 |
+
23,
|
| 113 |
+
24,
|
| 114 |
+
25,
|
| 115 |
+
26,
|
| 116 |
+
27,
|
| 117 |
+
28,
|
| 118 |
+
29,
|
| 119 |
+
30,
|
| 120 |
+
31,
|
| 121 |
+
32,
|
| 122 |
+
33,
|
| 123 |
+
34,
|
| 124 |
+
35,
|
| 125 |
+
36,
|
| 126 |
+
37,
|
| 127 |
+
38,
|
| 128 |
+
39,
|
| 129 |
+
40,
|
| 130 |
+
41,
|
| 131 |
+
42,
|
| 132 |
+
43,
|
| 133 |
+
44,
|
| 134 |
+
45,
|
| 135 |
+
46,
|
| 136 |
+
47,
|
| 137 |
+
48,
|
| 138 |
+
49,
|
| 139 |
+
50,
|
| 140 |
+
51,
|
| 141 |
+
52,
|
| 142 |
+
53,
|
| 143 |
+
54,
|
| 144 |
+
55
|
| 145 |
+
],
|
| 146 |
+
"model_type": "qwen3_next",
|
| 147 |
+
"moe_intermediate_size": 0,
|
| 148 |
+
"norm_topk_prob": true,
|
| 149 |
+
"num_attention_heads": 30,
|
| 150 |
+
"num_experts": 0,
|
| 151 |
+
"num_experts_per_tok": 0,
|
| 152 |
+
"num_hidden_layers": 56,
|
| 153 |
+
"num_key_value_heads": 6,
|
| 154 |
+
"output_router_logits": false,
|
| 155 |
+
"partial_rotary_factor": 1.0,
|
| 156 |
+
"rms_norm_eps": 1e-05,
|
| 157 |
+
"rope_scaling": null,
|
| 158 |
+
"rope_theta": 490000.0,
|
| 159 |
+
"router_aux_loss_coef": 0.001,
|
| 160 |
+
"router_bias": false,
|
| 161 |
+
"shared_expert_intermediate_size": 0,
|
| 162 |
+
"tie_word_embeddings": false,
|
| 163 |
+
"transformers_version": "4.57.1",
|
| 164 |
+
"use_cache": true,
|
| 165 |
+
"use_sliding_window": false,
|
| 166 |
+
"vocab_size": 99000
|
| 167 |
+
}
|