Hugging Face's logo

FlyCockpit
/
GLM-5.2-NVFP4-AQLM-Abliterated-Vision

Image-Text-to-Text
Safetensors
glm5v
glm
glm-5.2
abliterated
vision
nvfp4
aqlm
dgx-spark
conversational
8-bit precision
nvfp4_aqlm_hybrid
Model card Files
xet
 Community
GLM-5.2-NVFP4-AQLM-Abliterated-Vision
File size: 1,697 Bytes
fdc6474
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
{
  "comment": "GLM-5.2 hybrid NVFP4+GGUF quantization plan for 4x RTX PRO 6000 (SM120, 384GB). Layer -> expert precision assignment. Sources: lukealonso/GLM-5.2-NVFP4 (NVFP4 experts + all BF16 non-expert tensors), unsloth/GLM-5.2-GGUF UD-Q3_K_XL and UD-Q2_K_XL (routed expert donors, imatrix-calibrated).",
  "target": {
    "gpus": 4,
    "gpu_mem_mib": 97887,
    "sm": 120,
    "context_len": 1048576,
    "kv_cache_dtype": "fp8_ds_mla",
    "parallelism": "PP=4"
  },
  "nvfp4_layers": [
    78
  ],
  "aqlm_mixed_layers": [
    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
  ],
  "aqlm_cold_layers": [],
  "aqlm": {
    "nbits": 16,
    "entries": 65536,
    "group_size": 8,
    "mixed": {
      "w13_books": 1,
      "w2_books": 2
    },
    "cold": {
      "w13_books": 1,
      "w2_books": 1
    }
  },
  "sources": {
    "nvfp4": "https://huggingface.co/lukealonso/GLM-5.2-NVFP4/resolve/main"
  },
  "budget_gb": {
    "nonexpert_bf16": 38.0,
    "experts": 281.1,
    "kv_fp8_ds_mla_1m": 54.3,
    "indexer_cache_1m": 3.0,
    "total_need": 376.3,
    "total_have": 384.8
  },
  "per_expert_assignment": "/data/glm52-expert-assignment.json"
}