Hugging Face's logo

Youssofal
/
Qwen3.5-4B-Optimized-MTPLX

Text Generation
MLX
Safetensors
qwen3_5
apple-silicon
speculative-decoding
qwen
qwen3
mtp
mtplx
local-ai
q8
8-bit precision
conversational
Model card Files
xet
 Community
Qwen3.5-4B-Optimized-MTPLX
File size: 4,336 Bytes
b8825d6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
{
  "architectures": [
    "Qwen3_5ForConditionalGeneration"
  ],
  "image_token_id": 248056,
  "library_name": "mlx",
  "mlx_lm_extra_tensors": {
    "mtp_file": "mtp.safetensors",
    "mtp_tensor_count": 29
  },
  "model_type": "qwen3_5",
  "mtplx_mtp_quantization": {
    "bits": 4,
    "description": "Official Qwen3.5-4B MTP head with layer-0 attention/MLP linears quantized to MLX affine INT4 group64; mtp.fc and norms stay BF16.",
    "group_size": 64,
    "mode": "affine",
    "policy": "cyankiwi",
    "prequantized": true
  },
  "mtplx_policy": {
    "description": "mlx-community/Qwen3.5-4B-MLX-8bit trunk plus official Qwen3.5-4B MTP sidecar stored as body-int4 for MTPLX speed testing.",
    "local_official_source": "/Users/youssof/.mtplx/sources/Qwen--Qwen3.5-4B",
    "local_trunk_source": "/Users/youssof/.mtplx/sources/mlx-community--Qwen3.5-4B-MLX-8bit",
    "mtp_quantization": {
      "bits": 4,
      "group_size": 64,
      "linears": [
        "mtp.layers.0.mlp.down_proj",
        "mtp.layers.0.mlp.gate_proj",
        "mtp.layers.0.mlp.up_proj",
        "mtp.layers.0.self_attn.k_proj",
        "mtp.layers.0.self_attn.o_proj",
        "mtp.layers.0.self_attn.q_proj",
        "mtp.layers.0.self_attn.v_proj"
      ],
      "mode": "affine",
      "norms": "bf16"
    },
    "name": "qwen35-4b-q8-official-mtp-body-int4-speed",
    "quantization_family": "mlx-affine-q8",
    "source": "Qwen/Qwen3.5-4B",
    "trunk_quantization": {
      "bits": 8,
      "group_size": 64,
      "mode": "affine"
    },
    "trunk_source": "mlx-community/Qwen3.5-4B-MLX-8bit"
  },
  "quantization": {
    "bits": 8,
    "group_size": 64,
    "mode": "affine"
  },
  "quantization_config": {
    "bits": 8,
    "group_size": 64,
    "mode": "affine"
  },
  "text_config": {
    "attention_bias": false,
    "attention_dropout": 0.0,
    "attn_output_gate": true,
    "dtype": "bfloat16",
    "eos_token_id": 248044,
    "full_attention_interval": 4,
    "head_dim": 256,
    "hidden_act": "silu",
    "hidden_size": 2560,
    "initializer_range": 0.02,
    "intermediate_size": 9216,
    "layer_types": [
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention"
    ],
    "linear_conv_kernel_dim": 4,
    "linear_key_head_dim": 128,
    "linear_num_key_heads": 16,
    "linear_num_value_heads": 32,
    "linear_value_head_dim": 128,
    "mamba_ssm_dtype": "float32",
    "max_position_embeddings": 262144,
    "mlp_only_layers": [],
    "model_type": "qwen3_5_text",
    "mtp_num_hidden_layers": 1,
    "mtp_use_dedicated_embeddings": false,
    "num_attention_heads": 16,
    "num_hidden_layers": 32,
    "num_key_value_heads": 4,
    "rms_norm_eps": 1e-06,
    "rope_parameters": {
      "mrope_interleaved": true,
      "mrope_section": [
        11,
        11,
        10
      ],
      "partial_rotary_factor": 0.25,
      "rope_theta": 10000000,
      "rope_type": "default"
    },
    "tie_word_embeddings": true,
    "use_cache": true,
    "vocab_size": 248320
  },
  "tie_word_embeddings": true,
  "transformers_version": "4.57.0.dev0",
  "video_token_id": 248057,
  "vision_config": {
    "deepstack_visual_indexes": [],
    "depth": 24,
    "hidden_act": "gelu_pytorch_tanh",
    "hidden_size": 1024,
    "in_channels": 3,
    "initializer_range": 0.02,
    "intermediate_size": 4096,
    "model_type": "qwen3_5",
    "num_heads": 16,
    "num_position_embeddings": 2304,
    "out_hidden_size": 2560,
    "patch_size": 16,
    "spatial_merge_size": 2,
    "temporal_patch_size": 2
  },
  "vision_end_token_id": 248054,
  "vision_start_token_id": 248053
}