Hugging Face's logo

Disty0
/
FLUX.2-klein-4B-SDNQ-4bit-dynamic

Text-to-Image
Diffusers
Safetensors
Flux2KleinPipeline
sdnq
flux
4-bit precision
Model card Files
xet
 Community
2
FLUX.2-klein-4B-SDNQ-4bit-dynamic
File size: 6,496 Bytes
4d9d38f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
{
  "add_skip_keys": false,
  "dequantize_fp32": false,
  "dynamic_loss_threshold": 0.01,
  "group_size": 0,
  "is_integer": true,
  "is_training": false,
  "modules_dtype_dict": {
    "int5": [
      "single_transformer_blocks.19.attn.to_qkv_mlp_proj.weight",
      "transformer_blocks.0.attn.to_k.weight"
    ],
    "uint4": [
      "single_transformer_blocks.0.attn.to_out.weight",
      "single_transformer_blocks.0.attn.to_qkv_mlp_proj.weight",
      "single_transformer_blocks.1.attn.to_out.weight",
      "single_transformer_blocks.1.attn.to_qkv_mlp_proj.weight",
      "single_transformer_blocks.10.attn.to_out.weight",
      "single_transformer_blocks.10.attn.to_qkv_mlp_proj.weight",
      "single_transformer_blocks.11.attn.to_out.weight",
      "single_transformer_blocks.11.attn.to_qkv_mlp_proj.weight",
      "single_transformer_blocks.12.attn.to_out.weight",
      "single_transformer_blocks.12.attn.to_qkv_mlp_proj.weight",
      "single_transformer_blocks.13.attn.to_out.weight",
      "single_transformer_blocks.13.attn.to_qkv_mlp_proj.weight",
      "single_transformer_blocks.14.attn.to_out.weight",
      "single_transformer_blocks.14.attn.to_qkv_mlp_proj.weight",
      "single_transformer_blocks.15.attn.to_out.weight",
      "single_transformer_blocks.15.attn.to_qkv_mlp_proj.weight",
      "single_transformer_blocks.16.attn.to_out.weight",
      "single_transformer_blocks.16.attn.to_qkv_mlp_proj.weight",
      "single_transformer_blocks.17.attn.to_out.weight",
      "single_transformer_blocks.17.attn.to_qkv_mlp_proj.weight",
      "single_transformer_blocks.18.attn.to_out.weight",
      "single_transformer_blocks.18.attn.to_qkv_mlp_proj.weight",
      "single_transformer_blocks.19.attn.to_out.weight",
      "single_transformer_blocks.2.attn.to_out.weight",
      "single_transformer_blocks.2.attn.to_qkv_mlp_proj.weight",
      "single_transformer_blocks.3.attn.to_out.weight",
      "single_transformer_blocks.3.attn.to_qkv_mlp_proj.weight",
      "single_transformer_blocks.4.attn.to_out.weight",
      "single_transformer_blocks.4.attn.to_qkv_mlp_proj.weight",
      "single_transformer_blocks.5.attn.to_out.weight",
      "single_transformer_blocks.5.attn.to_qkv_mlp_proj.weight",
      "single_transformer_blocks.6.attn.to_out.weight",
      "single_transformer_blocks.6.attn.to_qkv_mlp_proj.weight",
      "single_transformer_blocks.7.attn.to_out.weight",
      "single_transformer_blocks.7.attn.to_qkv_mlp_proj.weight",
      "single_transformer_blocks.8.attn.to_out.weight",
      "single_transformer_blocks.8.attn.to_qkv_mlp_proj.weight",
      "single_transformer_blocks.9.attn.to_out.weight",
      "single_transformer_blocks.9.attn.to_qkv_mlp_proj.weight",
      "transformer_blocks.0.attn.add_k_proj.weight",
      "transformer_blocks.0.attn.add_q_proj.weight",
      "transformer_blocks.0.attn.add_v_proj.weight",
      "transformer_blocks.0.attn.to_add_out.weight",
      "transformer_blocks.0.attn.to_out.0.weight",
      "transformer_blocks.0.attn.to_q.weight",
      "transformer_blocks.0.attn.to_v.weight",
      "transformer_blocks.0.ff.linear_in.weight",
      "transformer_blocks.0.ff.linear_out.weight",
      "transformer_blocks.0.ff_context.linear_in.weight",
      "transformer_blocks.0.ff_context.linear_out.weight",
      "transformer_blocks.1.attn.add_k_proj.weight",
      "transformer_blocks.1.attn.add_q_proj.weight",
      "transformer_blocks.1.attn.add_v_proj.weight",
      "transformer_blocks.1.attn.to_add_out.weight",
      "transformer_blocks.1.attn.to_k.weight",
      "transformer_blocks.1.attn.to_out.0.weight",
      "transformer_blocks.1.attn.to_q.weight",
      "transformer_blocks.1.attn.to_v.weight",
      "transformer_blocks.1.ff.linear_in.weight",
      "transformer_blocks.1.ff.linear_out.weight",
      "transformer_blocks.1.ff_context.linear_in.weight",
      "transformer_blocks.1.ff_context.linear_out.weight",
      "transformer_blocks.2.attn.add_k_proj.weight",
      "transformer_blocks.2.attn.add_q_proj.weight",
      "transformer_blocks.2.attn.add_v_proj.weight",
      "transformer_blocks.2.attn.to_add_out.weight",
      "transformer_blocks.2.attn.to_k.weight",
      "transformer_blocks.2.attn.to_out.0.weight",
      "transformer_blocks.2.attn.to_q.weight",
      "transformer_blocks.2.attn.to_v.weight",
      "transformer_blocks.2.ff.linear_in.weight",
      "transformer_blocks.2.ff.linear_out.weight",
      "transformer_blocks.2.ff_context.linear_in.weight",
      "transformer_blocks.2.ff_context.linear_out.weight",
      "transformer_blocks.3.attn.add_k_proj.weight",
      "transformer_blocks.3.attn.add_q_proj.weight",
      "transformer_blocks.3.attn.add_v_proj.weight",
      "transformer_blocks.3.attn.to_add_out.weight",
      "transformer_blocks.3.attn.to_k.weight",
      "transformer_blocks.3.attn.to_out.0.weight",
      "transformer_blocks.3.attn.to_q.weight",
      "transformer_blocks.3.attn.to_v.weight",
      "transformer_blocks.3.ff.linear_in.weight",
      "transformer_blocks.3.ff.linear_out.weight",
      "transformer_blocks.3.ff_context.linear_in.weight",
      "transformer_blocks.3.ff_context.linear_out.weight",
      "transformer_blocks.4.attn.add_k_proj.weight",
      "transformer_blocks.4.attn.add_q_proj.weight",
      "transformer_blocks.4.attn.add_v_proj.weight",
      "transformer_blocks.4.attn.to_add_out.weight",
      "transformer_blocks.4.attn.to_k.weight",
      "transformer_blocks.4.attn.to_out.0.weight",
      "transformer_blocks.4.attn.to_q.weight",
      "transformer_blocks.4.attn.to_v.weight",
      "transformer_blocks.4.ff.linear_in.weight",
      "transformer_blocks.4.ff.linear_out.weight",
      "transformer_blocks.4.ff_context.linear_in.weight",
      "transformer_blocks.4.ff_context.linear_out.weight"
    ]
  },
  "modules_to_not_convert": [
    "double_stream_modulation_img",
    "time_guidance_embed",
    ".proj_out",
    "norm_out",
    "x_embedder",
    "double_stream_modulation_txt",
    "single_stream_modulation",
    "context_embedder"
  ],
  "non_blocking": false,
  "quant_conv": false,
  "quant_method": "sdnq",
  "quantization_device": null,
  "quantized_matmul_dtype": null,
  "return_device": null,
  "sdnq_version": "0.1.4",
  "svd_rank": 32,
  "svd_steps": 32,
  "use_dynamic_quantization": true,
  "use_grad_ckpt": true,
  "use_quantized_matmul": false,
  "use_quantized_matmul_conv": false,
  "use_static_quantization": true,
  "use_stochastic_rounding": false,
  "use_svd": false,
  "weights_dtype": "uint4"
}