Hugging Face's logo

pytorch
/
SmolLM3-3B-INT8-INT4

Text Generation
PyTorch
Transformers
optimum-executorch
smollm3
executorch
smollm
conversational
torchao
Model card Files
xet
Community
1
metascroy commited on
Commit
d83ed77
·
verified ·
1 Parent(s): 1c25305

Upload config.json with huggingface_hub

Browse files
Files changed (1) hide show
  1. config.json +192 -0
config.json ADDED
@@ -0,0 +1,192 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "SmolLM3ForCausalLM"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 128000,
8
+ "dtype": "bfloat16",
9
+ "eos_token_id": 128012,
10
+ "hidden_act": "silu",
11
+ "hidden_size": 2048,
12
+ "initializer_range": 0.02,
13
+ "intermediate_size": 11008,
14
+ "layer_types": [
15
+ "full_attention",
16
+ "full_attention",
17
+ "full_attention",
18
+ "full_attention",
19
+ "full_attention",
20
+ "full_attention",
21
+ "full_attention",
22
+ "full_attention",
23
+ "full_attention",
24
+ "full_attention",
25
+ "full_attention",
26
+ "full_attention",
27
+ "full_attention",
28
+ "full_attention",
29
+ "full_attention",
30
+ "full_attention",
31
+ "full_attention",
32
+ "full_attention",
33
+ "full_attention",
34
+ "full_attention",
35
+ "full_attention",
36
+ "full_attention",
37
+ "full_attention",
38
+ "full_attention",
39
+ "full_attention",
40
+ "full_attention",
41
+ "full_attention",
42
+ "full_attention",
43
+ "full_attention",
44
+ "full_attention",
45
+ "full_attention",
46
+ "full_attention",
47
+ "full_attention",
48
+ "full_attention",
49
+ "full_attention",
50
+ "full_attention"
51
+ ],
52
+ "max_position_embeddings": 65536,
53
+ "max_window_layers": 28,
54
+ "mlp_bias": false,
55
+ "model_type": "smollm3",
56
+ "no_rope_layer_interval": 4,
57
+ "no_rope_layers": [
58
+ 1,
59
+ 1,
60
+ 1,
61
+ 0,
62
+ 1,
63
+ 1,
64
+ 1,
65
+ 0,
66
+ 1,
67
+ 1,
68
+ 1,
69
+ 0,
70
+ 1,
71
+ 1,
72
+ 1,
73
+ 0,
74
+ 1,
75
+ 1,
76
+ 1,
77
+ 0,
78
+ 1,
79
+ 1,
80
+ 1,
81
+ 0,
82
+ 1,
83
+ 1,
84
+ 1,
85
+ 0,
86
+ 1,
87
+ 1,
88
+ 1,
89
+ 0,
90
+ 1,
91
+ 1,
92
+ 1,
93
+ 0
94
+ ],
95
+ "num_attention_heads": 16,
96
+ "num_hidden_layers": 36,
97
+ "num_key_value_heads": 4,
98
+ "pad_token_id": 128004,
99
+ "pretraining_tp": 2,
100
+ "quantization_config": {
101
+ "include_input_output_embeddings": true,
102
+ "modules_to_not_convert": [],
103
+ "quant_method": "torchao",
104
+ "quant_type": {
105
+ "default": {
106
+ "_data": {
107
+ "module_fqn_to_config": {
108
+ "_default": {
109
+ "_data": {
110
+ "act_mapping_type": {
111
+ "_data": "ASYMMETRIC",
112
+ "_type": "MappingType"
113
+ },
114
+ "intx_packing_format": {
115
+ "_data": "UNPACKED_TO_INT8",
116
+ "_type": "IntxPackingFormat"
117
+ },
118
+ "layout": {
119
+ "_data": {},
120
+ "_type": "QDQLayout",
121
+ "_version": 1
122
+ },
123
+ "weight_dtype": {
124
+ "_data": "int4",
125
+ "_type": "torch.dtype"
126
+ },
127
+ "weight_granularity": {
128
+ "_data": {
129
+ "group_size": 32
130
+ },
131
+ "_type": "PerGroup",
132
+ "_version": 1
133
+ },
134
+ "weight_mapping_type": {
135
+ "_data": "SYMMETRIC",
136
+ "_type": "MappingType"
137
+ },
138
+ "weight_scale_dtype": null
139
+ },
140
+ "_type": "Int8DynamicActivationIntxWeightConfig",
141
+ "_version": 2
142
+ },
143
+ "model.embed_tokens": {
144
+ "_data": {
145
+ "granularity": {
146
+ "_data": {
147
+ "axis": 0
148
+ },
149
+ "_type": "PerAxis",
150
+ "_version": 1
151
+ },
152
+ "intx_packing_format": {
153
+ "_data": "UNPACKED_TO_INT8",
154
+ "_type": "IntxPackingFormat"
155
+ },
156
+ "layout": {
157
+ "_data": {},
158
+ "_type": "QDQLayout",
159
+ "_version": 1
160
+ },
161
+ "mapping_type": {
162
+ "_data": "SYMMETRIC",
163
+ "_type": "MappingType"
164
+ },
165
+ "scale_dtype": null,
166
+ "weight_dtype": {
167
+ "_data": "int8",
168
+ "_type": "torch.dtype"
169
+ }
170
+ },
171
+ "_type": "IntxWeightOnlyConfig",
172
+ "_version": 2
173
+ }
174
+ }
175
+ },
176
+ "_type": "ModuleFqnToConfig",
177
+ "_version": 1
178
+ }
179
+ },
180
+ "quant_type_kwargs": {},
181
+ "untie_embedding_weights": false
182
+ },
183
+ "rms_norm_eps": 1e-06,
184
+ "rope_scaling": null,
185
+ "rope_theta": 5000000.0,
186
+ "sliding_window": null,
187
+ "tie_word_embeddings": false,
188
+ "transformers_version": "4.57.0.dev0",
189
+ "use_cache": false,
190
+ "use_sliding_window": false,
191
+ "vocab_size": 128256
192
+ }