Hugging Face's logo

callgg
/
glm-encoder

Diffusers
Safetensors
glm4v
4-bit precision
bitsandbytes
Model card Files
xet
Community
chatpig commited on
Commit
6422738
·
verified ·
1 Parent(s): c74d4ee

Upload config.json with huggingface_hub

Browse files
Files changed (1) hide show
  1. config.json +121 -0
config.json ADDED
@@ -0,0 +1,121 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Glm4vForConditionalGeneration"
4
+ ],
5
+ "attention_bias": true,
6
+ "attention_dropout": 0.0,
7
+ "eos_token_id": [
8
+ 151329,
9
+ 151336,
10
+ 151338,
11
+ 151348
12
+ ],
13
+ "hidden_act": "silu",
14
+ "hidden_size": 4096,
15
+ "image_end_token_id": 151340,
16
+ "image_start_token_id": 151339,
17
+ "image_token_id": 151343,
18
+ "initializer_range": 0.02,
19
+ "intermediate_size": 13696,
20
+ "max_position_embeddings": 32768,
21
+ "model_type": "glm4v",
22
+ "num_attention_heads": 32,
23
+ "num_hidden_layers": 40,
24
+ "num_key_value_heads": 2,
25
+ "pad_token_id": 151329,
26
+ "partial_rotary_factor": 0.5,
27
+ "quantization_config": {
28
+ "_load_in_4bit": true,
29
+ "_load_in_8bit": false,
30
+ "bnb_4bit_compute_dtype": "bfloat16",
31
+ "bnb_4bit_quant_storage": "uint8",
32
+ "bnb_4bit_quant_type": "nf4",
33
+ "bnb_4bit_use_double_quant": true,
34
+ "llm_int8_enable_fp32_cpu_offload": false,
35
+ "llm_int8_has_fp16_weight": false,
36
+ "llm_int8_skip_modules": null,
37
+ "llm_int8_threshold": 6.0,
38
+ "load_in_4bit": true,
39
+ "load_in_8bit": false,
40
+ "quant_method": "bitsandbytes"
41
+ },
42
+ "rms_norm_eps": 1e-05,
43
+ "rope_scaling": {
44
+ "mrope_section": [
45
+ 8,
46
+ 12,
47
+ 12
48
+ ],
49
+ "rope_type": "default",
50
+ "type": "default"
51
+ },
52
+ "rope_theta": 10000.0,
53
+ "text_config": {
54
+ "architectures": [
55
+ "Glm4vForConditionalGeneration"
56
+ ],
57
+ "attention_bias": true,
58
+ "attention_dropout": 0.0,
59
+ "eos_token_id": [
60
+ 151329,
61
+ 151336,
62
+ 151338,
63
+ 151348
64
+ ],
65
+ "hidden_act": "silu",
66
+ "hidden_size": 4096,
67
+ "image_token_id": null,
68
+ "initializer_range": 0.02,
69
+ "intermediate_size": 13696,
70
+ "max_position_embeddings": 32768,
71
+ "model_type": "glm4v_text",
72
+ "num_attention_heads": 32,
73
+ "num_hidden_layers": 40,
74
+ "num_key_value_heads": 2,
75
+ "pad_token_id": 151329,
76
+ "partial_rotary_factor": 0.5,
77
+ "rms_norm_eps": 1e-05,
78
+ "rope_scaling": {
79
+ "mrope_section": [
80
+ 8,
81
+ 12,
82
+ 12
83
+ ],
84
+ "rope_type": "default",
85
+ "type": "default"
86
+ },
87
+ "rope_theta": 10000.0,
88
+ "torch_dtype": "float16",
89
+ "use_cache": true,
90
+ "video_token_id": null,
91
+ "vocab_size": 151552
92
+ },
93
+ "tie_word_embeddings": false,
94
+ "torch_dtype": "float16",
95
+ "transformers_version": "4.54.0.dev0",
96
+ "use_cache": true,
97
+ "video_end_token_id": 151342,
98
+ "video_start_token_id": 151341,
99
+ "video_token_id": 151344,
100
+ "vision_config": {
101
+ "attention_bias": false,
102
+ "attention_dropout": 0.0,
103
+ "depth": 24,
104
+ "hidden_act": "silu",
105
+ "hidden_dropout_prob": 0.0,
106
+ "hidden_size": 1536,
107
+ "image_size": 336,
108
+ "in_channels": 3,
109
+ "initializer_range": 0.02,
110
+ "intermediate_size": 13696,
111
+ "model_type": "glm4v",
112
+ "num_heads": 12,
113
+ "out_hidden_size": 4096,
114
+ "patch_size": 14,
115
+ "rms_norm_eps": 1e-05,
116
+ "spatial_merge_size": 2,
117
+ "temporal_patch_size": 2,
118
+ "torch_dtype": "float16"
119
+ },
120
+ "vocab_size": 151552
121
+ }