Hugging Face's logo

echarlaix
/
SmolVLM2-500M-Video-Instruct-openvino-8bit-woq

Image-Text-to-Text
Transformers
OpenVINO
English
smolvlm
nncf
8-bit precision
conversational
Model card Files
xet
Community
echarlaix HF Staff commited on
Commit
ad203d4
·
verified ·
1 Parent(s): 6196c23

Upload openvino_text_embeddings_model.xml with huggingface_hub

Browse files
Files changed (1) hide show
  1. openvino_text_embeddings_model.xml +176 -0
openvino_text_embeddings_model.xml ADDED
@@ -0,0 +1,176 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0"?>
2
+ <net name="Model3" version="11">
3
+ <layers>
4
+ <layer id="0" name="input" type="Parameter" version="opset1">
5
+ <data shape="?,?" element_type="i64" />
6
+ <output>
7
+ <port id="0" precision="I64" names="input">
8
+ <dim>-1</dim>
9
+ <dim>-1</dim>
10
+ </port>
11
+ </output>
12
+ </layer>
13
+ <layer id="1" name="self.weight" type="Const" version="opset1">
14
+ <data element_type="i8" shape="49280, 960" offset="0" size="47308800" />
15
+ <output>
16
+ <port id="0" precision="I8">
17
+ <dim>49280</dim>
18
+ <dim>960</dim>
19
+ </port>
20
+ </output>
21
+ </layer>
22
+ <layer id="2" name="Convert_1106758" type="Convert" version="opset1">
23
+ <data destination_type="f16" />
24
+ <input>
25
+ <port id="0" precision="I8">
26
+ <dim>49280</dim>
27
+ <dim>960</dim>
28
+ </port>
29
+ </input>
30
+ <output>
31
+ <port id="1" precision="FP16">
32
+ <dim>49280</dim>
33
+ <dim>960</dim>
34
+ </port>
35
+ </output>
36
+ </layer>
37
+ <layer id="3" name="self.weight/scale" type="Const" version="opset1">
38
+ <data element_type="f16" shape="49280, 1" offset="47308800" size="98560" />
39
+ <output>
40
+ <port id="0" precision="FP16">
41
+ <dim>49280</dim>
42
+ <dim>1</dim>
43
+ </port>
44
+ </output>
45
+ </layer>
46
+ <layer id="4" name="self.weight/fq_weights_0" type="Multiply" version="opset1">
47
+ <data auto_broadcast="numpy" />
48
+ <input>
49
+ <port id="0" precision="FP16">
50
+ <dim>49280</dim>
51
+ <dim>960</dim>
52
+ </port>
53
+ <port id="1" precision="FP16">
54
+ <dim>49280</dim>
55
+ <dim>1</dim>
56
+ </port>
57
+ </input>
58
+ <output>
59
+ <port id="2" precision="FP16">
60
+ <dim>49280</dim>
61
+ <dim>960</dim>
62
+ </port>
63
+ </output>
64
+ </layer>
65
+ <layer id="5" name="self.weight/fq_weights_0/convert" type="Convert" version="opset1">
66
+ <data destination_type="f32" />
67
+ <input>
68
+ <port id="0" precision="FP16">
69
+ <dim>49280</dim>
70
+ <dim>960</dim>
71
+ </port>
72
+ </input>
73
+ <output>
74
+ <port id="1" precision="FP32">
75
+ <dim>49280</dim>
76
+ <dim>960</dim>
77
+ </port>
78
+ </output>
79
+ </layer>
80
+ <layer id="6" name="aten::embedding/Convert" type="Convert" version="opset1">
81
+ <data destination_type="i32" />
82
+ <input>
83
+ <port id="0" precision="I64">
84
+ <dim>-1</dim>
85
+ <dim>-1</dim>
86
+ </port>
87
+ </input>
88
+ <output>
89
+ <port id="1" precision="I32">
90
+ <dim>-1</dim>
91
+ <dim>-1</dim>
92
+ </port>
93
+ </output>
94
+ </layer>
95
+ <layer id="7" name="aten::embedding/Constant" type="Const" version="opset1">
96
+ <data element_type="i32" shape="" offset="47407360" size="4" />
97
+ <output>
98
+ <port id="0" precision="I32" />
99
+ </output>
100
+ </layer>
101
+ <layer id="8" name="aten::embedding/Gather" type="Gather" version="opset8">
102
+ <data batch_dims="0" />
103
+ <input>
104
+ <port id="0" precision="FP32">
105
+ <dim>49280</dim>
106
+ <dim>960</dim>
107
+ </port>
108
+ <port id="1" precision="I32">
109
+ <dim>-1</dim>
110
+ <dim>-1</dim>
111
+ </port>
112
+ <port id="2" precision="I32" />
113
+ </input>
114
+ <output>
115
+ <port id="3" precision="FP32" names="inputs_embeds">
116
+ <dim>-1</dim>
117
+ <dim>-1</dim>
118
+ <dim>960</dim>
119
+ </port>
120
+ </output>
121
+ </layer>
122
+ <layer id="9" name="Result_11454" type="Result" version="opset1" output_names="inputs_embeds">
123
+ <input>
124
+ <port id="0" precision="FP32">
125
+ <dim>-1</dim>
126
+ <dim>-1</dim>
127
+ <dim>960</dim>
128
+ </port>
129
+ </input>
130
+ </layer>
131
+ </layers>
132
+ <edges>
133
+ <edge from-layer="0" from-port="0" to-layer="6" to-port="0" />
134
+ <edge from-layer="1" from-port="0" to-layer="2" to-port="0" />
135
+ <edge from-layer="2" from-port="1" to-layer="4" to-port="0" />
136
+ <edge from-layer="3" from-port="0" to-layer="4" to-port="1" />
137
+ <edge from-layer="4" from-port="2" to-layer="5" to-port="0" />
138
+ <edge from-layer="5" from-port="1" to-layer="8" to-port="0" />
139
+ <edge from-layer="6" from-port="1" to-layer="8" to-port="1" />
140
+ <edge from-layer="7" from-port="0" to-layer="8" to-port="2" />
141
+ <edge from-layer="8" from-port="3" to-layer="9" to-port="0" />
142
+ </edges>
143
+ <rt_info>
144
+ <Runtime_version value="2025.2.0-19140-c01cd93e24d-releases/2025/2" />
145
+ <conversion_parameters>
146
+ <framework value="pytorch" />
147
+ <is_python_object value="True" />
148
+ </conversion_parameters>
149
+ <nncf>
150
+ <friendly_names_were_updated value="True" />
151
+ <version value="2.17.0" />
152
+ <weight_compression>
153
+ <advanced_parameters value="{'statistics_path': None, 'awq_params': {'subset_size': 32, 'percent_to_apply': 0.002, 'alpha_min': 0.0, 'alpha_max': 1.0, 'steps': 100, 'prefer_data_aware_scaling': True}, 'scale_estimation_params': {'subset_size': 64, 'initial_steps': 5, 'scale_steps': 5, 'weight_penalty': -1.0}, 'gptq_params': {'damp_percent': 0.1, 'block_size': 128, 'subset_size': 128}, 'lora_correction_params': {'adapter_rank': 8, 'num_iterations': 3, 'apply_regularization': True, 'subset_size': 128, 'use_int8_adapters': True}, 'lora_adapter_rank': 256, 'backend_params': {}}" />
154
+ <all_layers value="False" />
155
+ <awq value="False" />
156
+ <backup_mode value="int8_asym" />
157
+ <compression_format value="dequantize" />
158
+ <gptq value="False" />
159
+ <group_size value="-1" />
160
+ <ignored_scope value="[]" />
161
+ <lora_correction value="False" />
162
+ <mode value="int8_sym" />
163
+ <ratio value="1.0" />
164
+ <scale_estimation value="False" />
165
+ <sensitivity_metric value="weight_quantization_error" />
166
+ </weight_compression>
167
+ </nncf>
168
+ <optimum>
169
+ <nncf_version value="2.17.0" />
170
+ <optimum_intel_version value="1.24.0" />
171
+ <optimum_version value="1.26.1" />
172
+ <pytorch_version value="2.7.1" />
173
+ <transformers_version value="4.52.4" />
174
+ </optimum>
175
+ </rt_info>
176
+ </net>