Hugging Face's logo

OpenVINO
/
Qwen3-Reranker-0.6B-seq-cls-fp16-ov

Text Ranking
Transformers
OpenVINO
qwen3
text-classification
Model card Files
xet
Community
openvino-ci commited on
Commit
e3a68e3
·
verified ·
1 Parent(s): 7656994

Upload folder using huggingface_hub

Browse files
Files changed (15) hide show
  1. .gitattributes +1 -0
  2. README.md +105 -0
  3. added_tokens.json +28 -0
  4. config.json +67 -0
  5. merges.txt +0 -0
  6. openvino_detokenizer.bin +3 -0
  7. openvino_detokenizer.xml +219 -0
  8. openvino_model.bin +3 -0
  9. openvino_model.xml +0 -0
  10. openvino_tokenizer.bin +3 -0
  11. openvino_tokenizer.xml +722 -0
  12. special_tokens_map.json +31 -0
  13. tokenizer.json +3 -0
  14. tokenizer_config.json +239 -0
  15. vocab.json +0 -0
.gitattributes CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ tokenizer.json filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,105 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ base_model:
4
+ - tomaarsen/Qwen3-Reranker-0.6B-seq-cls
5
+ library_name: transformers
6
+ pipeline_tag: text-ranking
7
+ ---
8
+ # Qwen3-Reranker-0.6B-seq-cls-fp16-ov
9
+ * Model creator: [tomaarsen](https://huggingface.co/tomaarsen)
10
+ * Original model: [Qwen3-Reranker-0.6B-seq-cls](https://huggingface.co/tomaarsen/Qwen3-Reranker-0.6B-seq-cls)
11
+
12
+ ## Description
13
+ This is [Qwen3-Reranker-0.6B-seq-cls](https://huggingface.co/tomaarsen/Qwen3-Reranker-0.6B-seq-cls) model converted to the [OpenVINO™ IR](https://docs.openvino.ai/2025/documentation/openvino-ir-format.html) (Intermediate Representation) format with weights compressed to FP16.
14
+
15
+
16
+ ## Compatibility
17
+
18
+ The provided OpenVINO™ IR model is compatible with:
19
+
20
+ * OpenVINO version 2025.4.0 and higher
21
+ * Optimum Intel 1.26.0 and higher
22
+
23
+ ## Running Model Inference with [Optimum Intel](https://huggingface.co/docs/optimum/intel/index)
24
+
25
+ 1. Install packages required for using [Optimum Intel](https://huggingface.co/docs/optimum/intel/index) integration with the OpenVINO backend:
26
+
27
+ ```
28
+ pip install "git+https://github.com/huggingface/optimum-intel.git" "torch==2.8" --extra-index-url https://download.pytorch.org/whl/cpu
29
+
30
+ ```
31
+
32
+ 2. Run model inference:
33
+
34
+ ```
35
+ import torch
36
+ from transformers import AutoTokenizer
37
+ from optimum.intel import OVModelForCausalLM
38
+
39
+
40
+ def format_instruction(instruction, query, doc):
41
+ prefix = '<|im_start|>system\nJudge whether the Document meets the requirements based on the Query and the Instruct provided. Note that the answer can only be "yes" or "no".<|im_end|>\n<|im_start|>user\n'
42
+ suffix = "<|im_end|>\n<|im_start|>assistant\n<think>\n\n</think>\n\n"
43
+ if instruction is None:
44
+ instruction = (
45
+ "Given a web search query, retrieve relevant passages that answer the query"
46
+ )
47
+ output = f"{prefix}<Instruct>: {instruction}\n<Query>: {query}\n<Document>: {doc}{suffix}"
48
+ return output
49
+
50
+
51
+ model_id = "OpenVINO/Qwen3-Reranker-0.6B-seq-cls-fp16-ov"
52
+
53
+ model = OVModelForCausalLM.from_pretrained(model_id, use_cache=False, export=False)
54
+ tokenizer = AutoTokenizer.from_pretrained(model_id, padding_side="left")
55
+ # We recommend enabling flash_attention_2 for better acceleration and memory saving.
56
+ # model = AutoModelForSequenceClassification.from_pretrained("tomaarsen/Qwen3-Reranker-0.6B-seq-cls", torch_dtype=torch.float16, attn_implementation="flash_attention_2").cuda().eval()
57
+ max_length = 8192
58
+
59
+ task = "Given a web search query, retrieve relevant passages that answer the query"
60
+
61
+ queries = [
62
+ "Which planet is known as the Red Planet?",
63
+ "Which planet is known as the Red Planet?",
64
+ "Which planet is known as the Red Planet?",
65
+ "Which planet is known as the Red Planet?",
66
+ ]
67
+
68
+ documents = [
69
+ "Venus is often called Earth's twin because of its similar size and proximity.",
70
+ "Mars, known for its reddish appearance, is often referred to as the Red Planet.",
71
+ "Jupiter, the largest planet in our solar system, has a prominent red spot.",
72
+ "Saturn, famous for its rings, is sometimes mistaken for the Red Planet.",
73
+ ]
74
+
75
+ pairs = [format_instruction(task, query, doc) for query, doc in zip(queries, documents)]
76
+ inputs = tokenizer(
77
+ pairs,
78
+ padding=True,
79
+ truncation=True,
80
+ max_length=max_length,
81
+ return_tensors="pt",
82
+ )
83
+ logits = model(**inputs).logits.squeeze()
84
+ print(logits.tolist())
85
+ # [-3.109282970428467, 7.120373725891113, -0.37874650955200195, 3.5416228771209717]
86
+
87
+ scores = logits.sigmoid()
88
+ print(scores.tolist())
89
+ # [0.04272596165537834, 0.9991921782493591, 0.406429260969162, 0.9718491435050964]
90
+ ```
91
+
92
+ For more examples and possible optimizations, refer to the [Inference with Optimum Intel](https://docs.openvino.ai/2025/openvino-workflow-generative/inference-with-optimum-intel.html).
93
+
94
+
95
+ ## Limitations
96
+
97
+ Check the original [model card](https://huggingface.co/tomaarsen/Qwen3-Reranker-0.6B-seq-cls) for limitations.
98
+
99
+ ## Legal information
100
+
101
+ The original model is distributed under [Apache License Version 2.0](https://huggingface.co/datasets/choosealicense/licenses/blob/main/markdown/apache-2.0.md) license. More details can be found in [Qwen3-Reranker-0.6B-seq-cls](https://huggingface.co/tomaarsen/Qwen3-Reranker-0.6B-seq-cls).
102
+
103
+ ## Disclaimer
104
+
105
+ Intel is committed to respecting human rights and avoiding causing or contributing to adverse impacts on human rights. See [Intel’s Global Human Rights Principles](https://www.intel.com/content/dam/www/central-libraries/us/en/documents/policy-human-rights.pdf). Intel’s products and software are intended only to be used in applications that do not cause or contribute to adverse impacts on human rights.
added_tokens.json ADDED
@@ -0,0 +1,28 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "</think>": 151668,
3
+ "</tool_call>": 151658,
4
+ "</tool_response>": 151666,
5
+ "<think>": 151667,
6
+ "<tool_call>": 151657,
7
+ "<tool_response>": 151665,
8
+ "<|box_end|>": 151649,
9
+ "<|box_start|>": 151648,
10
+ "<|endoftext|>": 151643,
11
+ "<|file_sep|>": 151664,
12
+ "<|fim_middle|>": 151660,
13
+ "<|fim_pad|>": 151662,
14
+ "<|fim_prefix|>": 151659,
15
+ "<|fim_suffix|>": 151661,
16
+ "<|im_end|>": 151645,
17
+ "<|im_start|>": 151644,
18
+ "<|image_pad|>": 151655,
19
+ "<|object_ref_end|>": 151647,
20
+ "<|object_ref_start|>": 151646,
21
+ "<|quad_end|>": 151651,
22
+ "<|quad_start|>": 151650,
23
+ "<|repo_name|>": 151663,
24
+ "<|video_pad|>": 151656,
25
+ "<|vision_end|>": 151653,
26
+ "<|vision_pad|>": 151654,
27
+ "<|vision_start|>": 151652
28
+ }
config.json ADDED
@@ -0,0 +1,67 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Qwen3ForSequenceClassification"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 151643,
8
+ "eos_token_id": 151645,
9
+ "head_dim": 128,
10
+ "hidden_act": "silu",
11
+ "hidden_size": 1024,
12
+ "id2label": {
13
+ "0": "LABEL_0"
14
+ },
15
+ "initializer_range": 0.02,
16
+ "intermediate_size": 3072,
17
+ "label2id": {
18
+ "LABEL_0": 0
19
+ },
20
+ "layer_types": [
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
+ ],
50
+ "max_position_embeddings": 40960,
51
+ "max_window_layers": 28,
52
+ "model_type": "qwen3",
53
+ "num_attention_heads": 16,
54
+ "num_hidden_layers": 28,
55
+ "num_key_value_heads": 8,
56
+ "pad_token_id": 151643,
57
+ "rms_norm_eps": 1e-06,
58
+ "rope_scaling": null,
59
+ "rope_theta": 1000000,
60
+ "sliding_window": null,
61
+ "tie_word_embeddings": true,
62
+ "torch_dtype": "float32",
63
+ "transformers_version": "4.55.4",
64
+ "use_cache": true,
65
+ "use_sliding_window": false,
66
+ "vocab_size": 151669
67
+ }
merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
openvino_detokenizer.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d73093f72c80ebcaffb76a192727170ecd3c6a06c4df52df0e76b3f70ad9d7f7
3
+ size 2189717
openvino_detokenizer.xml ADDED
@@ -0,0 +1,219 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0"?>
2
+ <net name="detokenizer" version="11">
3
+ <layers>
4
+ <layer id="0" name="Parameter_71968" type="Parameter" version="opset1">
5
+ <data shape="?,?" element_type="i64" />
6
+ <output>
7
+ <port id="0" precision="I64" names="Parameter_71968">
8
+ <dim>-1</dim>
9
+ <dim>-1</dim>
10
+ </port>
11
+ </output>
12
+ </layer>
13
+ <layer id="1" name="Convert_72167" type="Convert" version="opset1">
14
+ <data destination_type="i32" />
15
+ <input>
16
+ <port id="0" precision="I64">
17
+ <dim>-1</dim>
18
+ <dim>-1</dim>
19
+ </port>
20
+ </input>
21
+ <output>
22
+ <port id="1" precision="I32">
23
+ <dim>-1</dim>
24
+ <dim>-1</dim>
25
+ </port>
26
+ </output>
27
+ </layer>
28
+ <layer id="2" name="Constant_71970" type="Const" version="opset1">
29
+ <data element_type="i32" shape="151669" offset="0" size="606676" />
30
+ <output>
31
+ <port id="0" precision="I32">
32
+ <dim>151669</dim>
33
+ </port>
34
+ </output>
35
+ </layer>
36
+ <layer id="3" name="Constant_71972" type="Const" version="opset1">
37
+ <data element_type="i32" shape="151669" offset="606676" size="606676" />
38
+ <output>
39
+ <port id="0" precision="I32">
40
+ <dim>151669</dim>
41
+ </port>
42
+ </output>
43
+ </layer>
44
+ <layer id="4" name="Constant_71974" type="Const" version="opset1">
45
+ <data element_type="u8" shape="976309" offset="1213352" size="976309" />
46
+ <output>
47
+ <port id="0" precision="U8">
48
+ <dim>976309</dim>
49
+ </port>
50
+ </output>
51
+ </layer>
52
+ <layer id="5" name="Slice_71979" type="Const" version="opset1">
53
+ <data element_type="i32" shape="14" offset="2189661" size="56" />
54
+ <output>
55
+ <port id="0" precision="I32">
56
+ <dim>14</dim>
57
+ </port>
58
+ </output>
59
+ </layer>
60
+ <layer id="6" name="VocabDecoder_71981" type="VocabDecoder" version="extension">
61
+ <data skip_tokens="" />
62
+ <input>
63
+ <port id="0" precision="I32">
64
+ <dim>-1</dim>
65
+ <dim>-1</dim>
66
+ </port>
67
+ <port id="1" precision="I32">
68
+ <dim>151669</dim>
69
+ </port>
70
+ <port id="2" precision="I32">
71
+ <dim>151669</dim>
72
+ </port>
73
+ <port id="3" precision="U8">
74
+ <dim>976309</dim>
75
+ </port>
76
+ <port id="4" precision="I32">
77
+ <dim>14</dim>
78
+ </port>
79
+ </input>
80
+ <output>
81
+ <port id="5" precision="I32">
82
+ <dim>-1</dim>
83
+ </port>
84
+ <port id="6" precision="I32">
85
+ <dim>-1</dim>
86
+ </port>
87
+ <port id="7" precision="I32">
88
+ <dim>-1</dim>
89
+ </port>
90
+ <port id="8" precision="I32">
91
+ <dim>-1</dim>
92
+ </port>
93
+ <port id="9" precision="U8">
94
+ <dim>-1</dim>
95
+ </port>
96
+ </output>
97
+ </layer>
98
+ <layer id="7" name="FuzeRagged_71982" type="FuzeRagged" version="extension">
99
+ <input>
100
+ <port id="0" precision="I32">
101
+ <dim>-1</dim>
102
+ </port>
103
+ <port id="1" precision="I32">
104
+ <dim>-1</dim>
105
+ </port>
106
+ <port id="2" precision="I32">
107
+ <dim>-1</dim>
108
+ </port>
109
+ <port id="3" precision="I32">
110
+ <dim>-1</dim>
111
+ </port>
112
+ </input>
113
+ <output>
114
+ <port id="4" precision="I32">
115
+ <dim>-1</dim>
116
+ </port>
117
+ <port id="5" precision="I32">
118
+ <dim>-1</dim>
119
+ </port>
120
+ </output>
121
+ </layer>
122
+ <layer id="8" name="UTF8Validate_71983" type="UTF8Validate" version="extension">
123
+ <data replace_mode="true" />
124
+ <input>
125
+ <port id="0" precision="I32">
126
+ <dim>-1</dim>
127
+ </port>
128
+ <port id="1" precision="I32">
129
+ <dim>-1</dim>
130
+ </port>
131
+ <port id="2" precision="U8">
132
+ <dim>-1</dim>
133
+ </port>
134
+ </input>
135
+ <output>
136
+ <port id="3" precision="I32">
137
+ <dim>-1</dim>
138
+ </port>
139
+ <port id="4" precision="I32">
140
+ <dim>-1</dim>
141
+ </port>
142
+ <port id="5" precision="U8">
143
+ <dim>-1</dim>
144
+ </port>
145
+ </output>
146
+ </layer>
147
+ <layer id="9" name="StringTensorPack_71984" type="StringTensorPack" version="opset15">
148
+ <input>
149
+ <port id="0" precision="I32">
150
+ <dim>-1</dim>
151
+ </port>
152
+ <port id="1" precision="I32">
153
+ <dim>-1</dim>
154
+ </port>
155
+ <port id="2" precision="U8">
156
+ <dim>-1</dim>
157
+ </port>
158
+ </input>
159
+ <output>
160
+ <port id="3" precision="STRING" names="Result_71985,string_output">
161
+ <dim>-1</dim>
162
+ </port>
163
+ </output>
164
+ </layer>
165
+ <layer id="10" name="Result_71985" type="Result" version="opset1" output_names="Result_71985,string_output">
166
+ <input>
167
+ <port id="0" precision="STRING">
168
+ <dim>-1</dim>
169
+ </port>
170
+ </input>
171
+ </layer>
172
+ </layers>
173
+ <edges>
174
+ <edge from-layer="0" from-port="0" to-layer="1" to-port="0" />
175
+ <edge from-layer="1" from-port="1" to-layer="6" to-port="0" />
176
+ <edge from-layer="2" from-port="0" to-layer="6" to-port="1" />
177
+ <edge from-layer="3" from-port="0" to-layer="6" to-port="2" />
178
+ <edge from-layer="4" from-port="0" to-layer="6" to-port="3" />
179
+ <edge from-layer="5" from-port="0" to-layer="6" to-port="4" />
180
+ <edge from-layer="6" from-port="7" to-layer="7" to-port="2" />
181
+ <edge from-layer="6" from-port="9" to-layer="8" to-port="2" />
182
+ <edge from-layer="6" from-port="8" to-layer="7" to-port="3" />
183
+ <edge from-layer="6" from-port="6" to-layer="7" to-port="1" />
184
+ <edge from-layer="6" from-port="5" to-layer="7" to-port="0" />
185
+ <edge from-layer="7" from-port="4" to-layer="8" to-port="0" />
186
+ <edge from-layer="7" from-port="5" to-layer="8" to-port="1" />
187
+ <edge from-layer="8" from-port="3" to-layer="9" to-port="0" />
188
+ <edge from-layer="8" from-port="4" to-layer="9" to-port="1" />
189
+ <edge from-layer="8" from-port="5" to-layer="9" to-port="2" />
190
+ <edge from-layer="9" from-port="3" to-layer="10" to-port="0" />
191
+ </edges>
192
+ <rt_info>
193
+ <add_attention_mask value="True" />
194
+ <add_prefix_space />
195
+ <add_special_tokens value="True" />
196
+ <clean_up_tokenization_spaces />
197
+ <detokenizer_input_type value="i64" />
198
+ <eos_token_id value="151645" />
199
+ <handle_special_tokens_with_re />
200
+ <max_length />
201
+ <number_of_inputs value="1" />
202
+ <openvino_tokenizers_version value="2025.4.0.0-617-b2d52c80e23" />
203
+ <openvino_version value="2025.4.0-20128-67fddb40e31" />
204
+ <original_post_processor_template value="{&quot;type&quot;: &quot;ByteLevel&quot;, &quot;add_prefix_space&quot;: false, &quot;trim_offsets&quot;: false, &quot;use_regex&quot;: false}" />
205
+ <original_tokenizer_class value="&lt;class 'transformers.models.qwen2.tokenization_qwen2_fast.Qwen2TokenizerFast'>" />
206
+ <pad_token_id value="151643" />
207
+ <sentencepiece_version value="0.2.1" />
208
+ <skip_special_tokens value="True" />
209
+ <streaming_detokenizer value="False" />
210
+ <tiktoken_version value="0.12.0" />
211
+ <tokenizer_output_type value="i64" />
212
+ <tokenizers_version value="0.21.4" />
213
+ <transformers_version value="4.55.4" />
214
+ <use_max_padding value="False" />
215
+ <use_sentencepiece_backend value="False" />
216
+ <utf8_replace_mode value="replace" />
217
+ <with_detokenizer value="True" />
218
+ </rt_info>
219
+ </net>
openvino_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cbcdc7e5d7b70d730b8100addb61480dcc4109ebd0977a24aac0253e039bf2fd
3
+ size 1191555461
openvino_model.xml ADDED
The diff for this file is too large to render. See raw diff
 
openvino_tokenizer.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:563c00e8a3a260601f9200179a23bb9aa6a184fc1dbe0ed0ec5320e26a888849
3
+ size 5588836
openvino_tokenizer.xml ADDED
@@ -0,0 +1,722 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0"?>
2
+ <net name="tokenizer" version="11">
3
+ <layers>
4
+ <layer id="0" name="Parameter_71845" type="Parameter" version="opset1">
5
+ <data shape="?" element_type="string" />
6
+ <output>
7
+ <port id="0" precision="STRING" names="Parameter_71845">
8
+ <dim>-1</dim>
9
+ </port>
10
+ </output>
11
+ </layer>
12
+ <layer id="1" name="Constant_71851" type="Const" version="opset1">
13
+ <data element_type="i64" shape="" offset="0" size="8" />
14
+ <output>
15
+ <port id="0" precision="I64" />
16
+ </output>
17
+ </layer>
18
+ <layer id="2" name="StringTensorUnpack_71846" type="StringTensorUnpack" version="opset15">
19
+ <input>
20
+ <port id="0" precision="STRING">
21
+ <dim>-1</dim>
22
+ </port>
23
+ </input>
24
+ <output>
25
+ <port id="1" precision="I32">
26
+ <dim>-1</dim>
27
+ </port>
28
+ <port id="2" precision="I32">
29
+ <dim>-1</dim>
30
+ </port>
31
+ <port id="3" precision="U8">
32
+ <dim>-1</dim>
33
+ </port>
34
+ </output>
35
+ </layer>
36
+ <layer id="3" name="ShapeOf_71847" type="ShapeOf" version="opset3">
37
+ <data output_type="i64" />
38
+ <input>
39
+ <port id="0" precision="I32">
40
+ <dim>-1</dim>
41
+ </port>
42
+ </input>
43
+ <output>
44
+ <port id="1" precision="I64">
45
+ <dim>1</dim>
46
+ </port>
47
+ </output>
48
+ </layer>
49
+ <layer id="4" name="Constant_71848" type="Const" version="opset1">
50
+ <data element_type="i64" shape="" offset="0" size="8" />
51
+ <output>
52
+ <port id="0" precision="I64" />
53
+ </output>
54
+ </layer>
55
+ <layer id="5" name="Constant_71849" type="Const" version="opset1">
56
+ <data element_type="i64" shape="" offset="0" size="8" />
57
+ <output>
58
+ <port id="0" precision="I64" />
59
+ </output>
60
+ </layer>
61
+ <layer id="6" name="Gather_71850" type="Gather" version="opset8">
62
+ <data batch_dims="0" />
63
+ <input>
64
+ <port id="0" precision="I64">
65
+ <dim>1</dim>
66
+ </port>
67
+ <port id="1" precision="I64" />
68
+ <port id="2" precision="I64" />
69
+ </input>
70
+ <output>
71
+ <port id="3" precision="I64" />
72
+ </output>
73
+ </layer>
74
+ <layer id="7" name="Constant_71852" type="Const" version="opset1">
75
+ <data element_type="i64" shape="" offset="8" size="8" />
76
+ <output>
77
+ <port id="0" precision="I64" />
78
+ </output>
79
+ </layer>
80
+ <layer id="8" name="Range_71853" type="Range" version="opset4">
81
+ <data output_type="i32" />
82
+ <input>
83
+ <port id="0" precision="I64" />
84
+ <port id="1" precision="I64" />
85
+ <port id="2" precision="I64" />
86
+ </input>
87
+ <output>
88
+ <port id="3" precision="I32">
89
+ <dim>-1</dim>
90
+ </port>
91
+ </output>
92
+ </layer>
93
+ <layer id="9" name="Constant_71854" type="Const" version="opset1">
94
+ <data element_type="i64" shape="" offset="8" size="8" />
95
+ <output>
96
+ <port id="0" precision="I64" />
97
+ </output>
98
+ </layer>
99
+ <layer id="10" name="Constant_71855" type="Const" version="opset1">
100
+ <data element_type="i64" shape="" offset="8" size="8" />
101
+ <output>
102
+ <port id="0" precision="I64" />
103
+ </output>
104
+ </layer>
105
+ <layer id="11" name="Add_71856" type="Add" version="opset1">
106
+ <data auto_broadcast="numpy" />
107
+ <input>
108
+ <port id="0" precision="I64" />
109
+ <port id="1" precision="I64" />
110
+ </input>
111
+ <output>
112
+ <port id="2" precision="I64" />
113
+ </output>
114
+ </layer>
115
+ <layer id="12" name="Constant_71857" type="Const" version="opset1">
116
+ <data element_type="i64" shape="" offset="8" size="8" />
117
+ <output>
118
+ <port id="0" precision="I64" />
119
+ </output>
120
+ </layer>
121
+ <layer id="13" name="Range_71858" type="Range" version="opset4">
122
+ <data output_type="i32" />
123
+ <input>
124
+ <port id="0" precision="I64" />
125
+ <port id="1" precision="I64" />
126
+ <port id="2" precision="I64" />
127
+ </input>
128
+ <output>
129
+ <port id="3" precision="I32">
130
+ <dim>-1</dim>
131
+ </port>
132
+ </output>
133
+ </layer>
134
+ <layer id="14" name="Constant_71922" type="Const" version="opset1">
135
+ <data element_type="u8" shape="462" offset="16" size="462" />
136
+ <output>
137
+ <port id="0" precision="U8">
138
+ <dim>462</dim>
139
+ </port>
140
+ </output>
141
+ </layer>
142
+ <layer id="15" name="SpecialTokensSplit_71923" type="SpecialTokensSplit" version="extension">
143
+ <input>
144
+ <port id="0" precision="I32">
145
+ <dim>-1</dim>
146
+ </port>
147
+ <port id="1" precision="I32">
148
+ <dim>-1</dim>
149
+ </port>
150
+ <port id="2" precision="I32">
151
+ <dim>-1</dim>
152
+ </port>
153
+ <port id="3" precision="I32">
154
+ <dim>-1</dim>
155
+ </port>
156
+ <port id="4" precision="U8">
157
+ <dim>-1</dim>
158
+ </port>
159
+ <port id="5" precision="U8">
160
+ <dim>462</dim>
161
+ </port>
162
+ </input>
163
+ <output>
164
+ <port id="6" precision="I32">
165
+ <dim>-1</dim>
166
+ </port>
167
+ <port id="7" precision="I32">
168
+ <dim>-1</dim>
169
+ </port>
170
+ <port id="8" precision="I32">
171
+ <dim>-1</dim>
172
+ </port>
173
+ <port id="9" precision="I32">
174
+ <dim>-1</dim>
175
+ </port>
176
+ <port id="10" precision="U8">
177
+ <dim>-1</dim>
178
+ </port>
179
+ <port id="11" precision="BOOL">
180
+ <dim>-1</dim>
181
+ </port>
182
+ </output>
183
+ </layer>
184
+ <layer id="16" name="CharsMapNormalization_71924" type="CharsMapNormalization" version="extension">
185
+ <data add_dummy_prefix="false" remove_extra_whitespaces="false" escape_whitespaces="false" normalization_form="nfc" case_fold="false" nmt="false" />
186
+ <input>
187
+ <port id="0" precision="I32">
188
+ <dim>-1</dim>
189
+ </port>
190
+ <port id="1" precision="I32">
191
+ <dim>-1</dim>
192
+ </port>
193
+ <port id="2" precision="U8">
194
+ <dim>-1</dim>
195
+ </port>
196
+ <port id="3" precision="BOOL">
197
+ <dim>-1</dim>
198
+ </port>
199
+ </input>
200
+ <output>
201
+ <port id="4" precision="I32">
202
+ <dim>-1</dim>
203
+ </port>
204
+ <port id="5" precision="I32">
205
+ <dim>-1</dim>
206
+ </port>
207
+ <port id="6" precision="U8">
208
+ <dim>-1</dim>
209
+ </port>
210
+ <port id="7" precision="BOOL">
211
+ <dim>-1</dim>
212
+ </port>
213
+ </output>
214
+ </layer>
215
+ <layer id="17" name="Constant_71926" type="Const" version="opset1">
216
+ <data element_type="u8" shape="110" offset="478" size="110" />
217
+ <output>
218
+ <port id="0" precision="U8">
219
+ <dim>110</dim>
220
+ </port>
221
+ </output>
222
+ </layer>
223
+ <layer id="18" name="RegexSplit_71927" type="RegexSplit" version="extension">
224
+ <data behaviour="isolate" invert="false" max_splits="-1" />
225
+ <input>
226
+ <port id="0" precision="I32">
227
+ <dim>-1</dim>
228
+ </port>
229
+ <port id="1" precision="I32">
230
+ <dim>-1</dim>
231
+ </port>
232
+ <port id="2" precision="I32">
233
+ <dim>-1</dim>
234
+ </port>
235
+ <port id="3" precision="I32">
236
+ <dim>-1</dim>
237
+ </port>
238
+ <port id="4" precision="U8">
239
+ <dim>-1</dim>
240
+ </port>
241
+ <port id="5" precision="BOOL">
242
+ <dim>-1</dim>
243
+ </port>
244
+ <port id="6" precision="U8">
245
+ <dim>110</dim>
246
+ </port>
247
+ </input>
248
+ <output>
249
+ <port id="7" precision="I32">
250
+ <dim>-1</dim>
251
+ </port>
252
+ <port id="8" precision="I32">
253
+ <dim>-1</dim>
254
+ </port>
255
+ <port id="9" precision="I32">
256
+ <dim>-1</dim>
257
+ </port>
258
+ <port id="10" precision="I32">
259
+ <dim>-1</dim>
260
+ </port>
261
+ <port id="11" precision="U8">
262
+ <dim>-1</dim>
263
+ </port>
264
+ <port id="12" precision="BOOL">
265
+ <dim>-1</dim>
266
+ </port>
267
+ </output>
268
+ </layer>
269
+ <layer id="19" name="Constant_71929" type="Const" version="opset1">
270
+ <data element_type="i32" shape="151669" offset="588" size="606676" />
271
+ <output>
272
+ <port id="0" precision="I32">
273
+ <dim>151669</dim>
274
+ </port>
275
+ </output>
276
+ </layer>
277
+ <layer id="20" name="Constant_71931" type="Const" version="opset1">
278
+ <data element_type="i32" shape="151669" offset="607264" size="606676" />
279
+ <output>
280
+ <port id="0" precision="I32">
281
+ <dim>151669</dim>
282
+ </port>
283
+ </output>
284
+ </layer>
285
+ <layer id="21" name="Constant_71933" type="Const" version="opset1">
286
+ <data element_type="u8" shape="976309" offset="1213940" size="976309" />
287
+ <output>
288
+ <port id="0" precision="U8">
289
+ <dim>976309</dim>
290
+ </port>
291
+ </output>
292
+ </layer>
293
+ <layer id="22" name="Constant_71941" type="Const" version="opset1">
294
+ <data element_type="i32" shape="151387" offset="2190249" size="605548" />
295
+ <output>
296
+ <port id="0" precision="I32">
297
+ <dim>151387</dim>
298
+ </port>
299
+ </output>
300
+ </layer>
301
+ <layer id="23" name="Constant_71943" type="Const" version="opset1">
302
+ <data element_type="i32" shape="151387" offset="2795797" size="605548" />
303
+ <output>
304
+ <port id="0" precision="I32">
305
+ <dim>151387</dim>
306
+ </port>
307
+ </output>
308
+ </layer>
309
+ <layer id="24" name="Constant_71945" type="Const" version="opset1">
310
+ <data element_type="u8" shape="491359" offset="3401345" size="491359" />
311
+ <output>
312
+ <port id="0" precision="U8">
313
+ <dim>491359</dim>
314
+ </port>
315
+ </output>
316
+ </layer>
317
+ <layer id="25" name="Constant_71947" type="Const" version="opset1">
318
+ <data element_type="i32" shape="151387" offset="3892704" size="605548" />
319
+ <output>
320
+ <port id="0" precision="I32">
321
+ <dim>151387</dim>
322
+ </port>
323
+ </output>
324
+ </layer>
325
+ <layer id="26" name="Constant_71949" type="Const" version="opset1">
326
+ <data element_type="i32" shape="151387" offset="4498252" size="605548" />
327
+ <output>
328
+ <port id="0" precision="I32">
329
+ <dim>151387</dim>
330
+ </port>
331
+ </output>
332
+ </layer>
333
+ <layer id="27" name="Constant_71951" type="Const" version="opset1">
334
+ <data element_type="u8" shape="484354" offset="5103800" size="484354" />
335
+ <output>
336
+ <port id="0" precision="U8">
337
+ <dim>484354</dim>
338
+ </port>
339
+ </output>
340
+ </layer>
341
+ <layer id="28" name="Constant_71935" type="Const" version="opset1">
342
+ <data element_type="i32" shape="26" offset="5588154" size="104" />
343
+ <output>
344
+ <port id="0" precision="I32">
345
+ <dim>26</dim>
346
+ </port>
347
+ </output>
348
+ </layer>
349
+ <layer id="29" name="Constant_71937" type="Const" version="opset1">
350
+ <data element_type="i32" shape="26" offset="5588258" size="104" />
351
+ <output>
352
+ <port id="0" precision="I32">
353
+ <dim>26</dim>
354
+ </port>
355
+ </output>
356
+ </layer>
357
+ <layer id="30" name="Constant_71939" type="Const" version="opset1">
358
+ <data element_type="u8" shape="340" offset="5588362" size="340" />
359
+ <output>
360
+ <port id="0" precision="U8">
361
+ <dim>340</dim>
362
+ </port>
363
+ </output>
364
+ </layer>
365
+ <layer id="31" name="Constant_71952" type="Const" version="opset1">
366
+ <data element_type="i32" shape="26" offset="5588702" size="104" />
367
+ <output>
368
+ <port id="0" precision="I32">
369
+ <dim>26</dim>
370
+ </port>
371
+ </output>
372
+ </layer>
373
+ <layer id="32" name="BPETokenizer_71953" type="BPETokenizer" version="extension">
374
+ <data unk_token="" fuse_unk="false" suffix_indicator="" end_suffix="" byte_fallback="false" cache_capacity="30328" />
375
+ <input>
376
+ <port id="0" precision="I32">
377
+ <dim>-1</dim>
378
+ </port>
379
+ <port id="1" precision="I32">
380
+ <dim>-1</dim>
381
+ </port>
382
+ <port id="2" precision="I32">
383
+ <dim>-1</dim>
384
+ </port>
385
+ <port id="3" precision="I32">
386
+ <dim>-1</dim>
387
+ </port>
388
+ <port id="4" precision="U8">
389
+ <dim>-1</dim>
390
+ </port>
391
+ <port id="5" precision="I32">
392
+ <dim>151669</dim>
393
+ </port>
394
+ <port id="6" precision="I32">
395
+ <dim>151669</dim>
396
+ </port>
397
+ <port id="7" precision="U8">
398
+ <dim>976309</dim>
399
+ </port>
400
+ <port id="8" precision="I32">
401
+ <dim>151387</dim>
402
+ </port>
403
+ <port id="9" precision="I32">
404
+ <dim>151387</dim>
405
+ </port>
406
+ <port id="10" precision="U8">
407
+ <dim>491359</dim>
408
+ </port>
409
+ <port id="11" precision="I32">
410
+ <dim>151387</dim>
411
+ </port>
412
+ <port id="12" precision="I32">
413
+ <dim>151387</dim>
414
+ </port>
415
+ <port id="13" precision="U8">
416
+ <dim>484354</dim>
417
+ </port>
418
+ <port id="14" precision="I32">
419
+ <dim>26</dim>
420
+ </port>
421
+ <port id="15" precision="I32">
422
+ <dim>26</dim>
423
+ </port>
424
+ <port id="16" precision="U8">
425
+ <dim>340</dim>
426
+ </port>
427
+ <port id="17" precision="I32">
428
+ <dim>26</dim>
429
+ </port>
430
+ </input>
431
+ <output>
432
+ <port id="18" precision="I32">
433
+ <dim>-1</dim>
434
+ </port>
435
+ <port id="19" precision="I32">
436
+ <dim>-1</dim>
437
+ </port>
438
+ <port id="20" precision="I32">
439
+ <dim>-1</dim>
440
+ </port>
441
+ </output>
442
+ </layer>
443
+ <layer id="33" name="Constant_71954" type="Const" version="opset1">
444
+ <data element_type="i32" shape="" offset="5588806" size="4" />
445
+ <output>
446
+ <port id="0" precision="I32" />
447
+ </output>
448
+ </layer>
449
+ <layer id="34" name="Constant_71956" type="Const" version="opset1">
450
+ <data element_type="u8" shape="5" offset="5588810" size="5" />
451
+ <output>
452
+ <port id="0" precision="U8">
453
+ <dim>5</dim>
454
+ </port>
455
+ </output>
456
+ </layer>
457
+ <layer id="35" name="Constant_71958" type="Const" version="opset1">
458
+ <data element_type="u8" shape="13" offset="5588815" size="13" />
459
+ <output>
460
+ <port id="0" precision="U8">
461
+ <dim>13</dim>
462
+ </port>
463
+ </output>
464
+ </layer>
465
+ <layer id="36" name="Truncate_71959" type="Truncate" version="extension">
466
+ <data m_num_inputs="1" />
467
+ <input>
468
+ <port id="0" precision="I32">
469
+ <dim>-1</dim>
470
+ </port>
471
+ <port id="1" precision="I32">
472
+ <dim>-1</dim>
473
+ </port>
474
+ <port id="2" precision="I32">
475
+ <dim>-1</dim>
476
+ </port>
477
+ <port id="3" precision="I32" />
478
+ <port id="4" precision="U8">
479
+ <dim>5</dim>
480
+ </port>
481
+ <port id="5" precision="U8">
482
+ <dim>13</dim>
483
+ </port>
484
+ </input>
485
+ <output>
486
+ <port id="6" precision="I32">
487
+ <dim>-1</dim>
488
+ </port>
489
+ <port id="7" precision="I32">
490
+ <dim>-1</dim>
491
+ </port>
492
+ <port id="8" precision="I32">
493
+ <dim>-1</dim>
494
+ </port>
495
+ </output>
496
+ </layer>
497
+ <layer id="37" name="Subtract_71960" type="Subtract" version="opset1">
498
+ <data auto_broadcast="numpy" />
499
+ <input>
500
+ <port id="0" precision="I32">
501
+ <dim>-1</dim>
502
+ </port>
503
+ <port id="1" precision="I32">
504
+ <dim>-1</dim>
505
+ </port>
506
+ </input>
507
+ <output>
508
+ <port id="2" precision="I32">
509
+ <dim>-1</dim>
510
+ </port>
511
+ </output>
512
+ </layer>
513
+ <layer id="38" name="Constant_71961" type="Const" version="opset1">
514
+ <data element_type="i32" shape="" offset="5588828" size="4" />
515
+ <output>
516
+ <port id="0" precision="I32" />
517
+ </output>
518
+ </layer>
519
+ <layer id="39" name="ReduceMax_71962" type="ReduceMax" version="opset1">
520
+ <data keep_dims="false" />
521
+ <input>
522
+ <port id="0" precision="I32">
523
+ <dim>-1</dim>
524
+ </port>
525
+ <port id="1" precision="I32" />
526
+ </input>
527
+ <output>
528
+ <port id="2" precision="I32" />
529
+ </output>
530
+ </layer>
531
+ <layer id="40" name="Constant_71963" type="Const" version="opset1">
532
+ <data element_type="i32" shape="" offset="5588832" size="4" />
533
+ <output>
534
+ <port id="0" precision="I32" />
535
+ </output>
536
+ </layer>
537
+ <layer id="41" name="RaggedToDense_71964" type="RaggedToDense" version="extension">
538
+ <data pad_right="true" m_pad_max_length="false" />
539
+ <input>
540
+ <port id="0" precision="I32">
541
+ <dim>-1</dim>
542
+ </port>
543
+ <port id="1" precision="I32">
544
+ <dim>-1</dim>
545
+ </port>
546
+ <port id="2" precision="I32">
547
+ <dim>-1</dim>
548
+ </port>
549
+ <port id="3" precision="I32" />
550
+ <port id="4" precision="I32" />
551
+ </input>
552
+ <output>
553
+ <port id="5" precision="I32">
554
+ <dim>-1</dim>
555
+ <dim>-1</dim>
556
+ </port>
557
+ <port id="6" precision="BOOL">
558
+ <dim>-1</dim>
559
+ <dim>-1</dim>
560
+ </port>
561
+ </output>
562
+ </layer>
563
+ <layer id="42" name="Convert_71965" type="Convert" version="opset1">
564
+ <data destination_type="i32" />
565
+ <input>
566
+ <port id="0" precision="BOOL">
567
+ <dim>-1</dim>
568
+ <dim>-1</dim>
569
+ </port>
570
+ </input>
571
+ <output>
572
+ <port id="1" precision="I32">
573
+ <dim>-1</dim>
574
+ <dim>-1</dim>
575
+ </port>
576
+ </output>
577
+ </layer>
578
+ <layer id="43" name="Convert_71965.0" type="Convert" version="opset1">
579
+ <data destination_type="i64" />
580
+ <input>
581
+ <port id="0" precision="I32">
582
+ <dim>-1</dim>
583
+ <dim>-1</dim>
584
+ </port>
585
+ </input>
586
+ <output>
587
+ <port id="1" precision="I64" names="attention_mask">
588
+ <dim>-1</dim>
589
+ <dim>-1</dim>
590
+ </port>
591
+ </output>
592
+ </layer>
593
+ <layer id="45" name="RaggedToDense_71964.0" type="Convert" version="opset1">
594
+ <data destination_type="i64" />
595
+ <input>
596
+ <port id="0" precision="I32">
597
+ <dim>-1</dim>
598
+ <dim>-1</dim>
599
+ </port>
600
+ </input>
601
+ <output>
602
+ <port id="1" precision="I64" names="input_ids">
603
+ <dim>-1</dim>
604
+ <dim>-1</dim>
605
+ </port>
606
+ </output>
607
+ </layer>
608
+ <layer id="46" name="Result_71966" type="Result" version="opset1" output_names="input_ids">
609
+ <input>
610
+ <port id="0" precision="I64">
611
+ <dim>-1</dim>
612
+ <dim>-1</dim>
613
+ </port>
614
+ </input>
615
+ </layer>
616
+ <layer id="44" name="Result_71967" type="Result" version="opset1" output_names="attention_mask">
617
+ <input>
618
+ <port id="0" precision="I64">
619
+ <dim>-1</dim>
620
+ <dim>-1</dim>
621
+ </port>
622
+ </input>
623
+ </layer>
624
+ </layers>
625
+ <edges>
626
+ <edge from-layer="0" from-port="0" to-layer="2" to-port="0" />
627
+ <edge from-layer="1" from-port="0" to-layer="8" to-port="0" />
628
+ <edge from-layer="2" from-port="1" to-layer="3" to-port="0" />
629
+ <edge from-layer="2" from-port="3" to-layer="15" to-port="4" />
630
+ <edge from-layer="2" from-port="2" to-layer="15" to-port="3" />
631
+ <edge from-layer="2" from-port="1" to-layer="15" to-port="2" />
632
+ <edge from-layer="3" from-port="1" to-layer="6" to-port="0" />
633
+ <edge from-layer="4" from-port="0" to-layer="6" to-port="1" />
634
+ <edge from-layer="5" from-port="0" to-layer="6" to-port="2" />
635
+ <edge from-layer="6" from-port="3" to-layer="11" to-port="0" />
636
+ <edge from-layer="6" from-port="3" to-layer="8" to-port="1" />
637
+ <edge from-layer="7" from-port="0" to-layer="8" to-port="2" />
638
+ <edge from-layer="8" from-port="3" to-layer="15" to-port="0" />
639
+ <edge from-layer="9" from-port="0" to-layer="13" to-port="0" />
640
+ <edge from-layer="10" from-port="0" to-layer="11" to-port="1" />
641
+ <edge from-layer="11" from-port="2" to-layer="13" to-port="1" />
642
+ <edge from-layer="12" from-port="0" to-layer="13" to-port="2" />
643
+ <edge from-layer="13" from-port="3" to-layer="15" to-port="1" />
644
+ <edge from-layer="14" from-port="0" to-layer="15" to-port="5" />
645
+ <edge from-layer="15" from-port="8" to-layer="16" to-port="0" />
646
+ <edge from-layer="15" from-port="7" to-layer="18" to-port="1" />
647
+ <edge from-layer="15" from-port="6" to-layer="18" to-port="0" />
648
+ <edge from-layer="15" from-port="11" to-layer="16" to-port="3" />
649
+ <edge from-layer="15" from-port="10" to-layer="16" to-port="2" />
650
+ <edge from-layer="15" from-port="9" to-layer="16" to-port="1" />
651
+ <edge from-layer="16" from-port="4" to-layer="18" to-port="2" />
652
+ <edge from-layer="16" from-port="5" to-layer="18" to-port="3" />
653
+ <edge from-layer="16" from-port="6" to-layer="18" to-port="4" />
654
+ <edge from-layer="16" from-port="7" to-layer="18" to-port="5" />
655
+ <edge from-layer="17" from-port="0" to-layer="18" to-port="6" />
656
+ <edge from-layer="18" from-port="11" to-layer="32" to-port="4" />
657
+ <edge from-layer="18" from-port="10" to-layer="32" to-port="3" />
658
+ <edge from-layer="18" from-port="8" to-layer="32" to-port="1" />
659
+ <edge from-layer="18" from-port="9" to-layer="32" to-port="2" />
660
+ <edge from-layer="18" from-port="7" to-layer="32" to-port="0" />
661
+ <edge from-layer="19" from-port="0" to-layer="32" to-port="5" />
662
+ <edge from-layer="20" from-port="0" to-layer="32" to-port="6" />
663
+ <edge from-layer="21" from-port="0" to-layer="32" to-port="7" />
664
+ <edge from-layer="22" from-port="0" to-layer="32" to-port="8" />
665
+ <edge from-layer="23" from-port="0" to-layer="32" to-port="9" />
666
+ <edge from-layer="24" from-port="0" to-layer="32" to-port="10" />
667
+ <edge from-layer="25" from-port="0" to-layer="32" to-port="11" />
668
+ <edge from-layer="26" from-port="0" to-layer="32" to-port="12" />
669
+ <edge from-layer="27" from-port="0" to-layer="32" to-port="13" />
670
+ <edge from-layer="28" from-port="0" to-layer="32" to-port="14" />
671
+ <edge from-layer="29" from-port="0" to-layer="32" to-port="15" />
672
+ <edge from-layer="30" from-port="0" to-layer="32" to-port="16" />
673
+ <edge from-layer="31" from-port="0" to-layer="32" to-port="17" />
674
+ <edge from-layer="32" from-port="20" to-layer="36" to-port="2" />
675
+ <edge from-layer="32" from-port="18" to-layer="36" to-port="0" />
676
+ <edge from-layer="32" from-port="19" to-layer="36" to-port="1" />
677
+ <edge from-layer="33" from-port="0" to-layer="36" to-port="3" />
678
+ <edge from-layer="34" from-port="0" to-layer="36" to-port="4" />
679
+ <edge from-layer="35" from-port="0" to-layer="36" to-port="5" />
680
+ <edge from-layer="36" from-port="6" to-layer="41" to-port="0" />
681
+ <edge from-layer="36" from-port="8" to-layer="41" to-port="2" />
682
+ <edge from-layer="36" from-port="7" to-layer="41" to-port="1" />
683
+ <edge from-layer="36" from-port="6" to-layer="37" to-port="1" />
684
+ <edge from-layer="36" from-port="7" to-layer="37" to-port="0" />
685
+ <edge from-layer="37" from-port="2" to-layer="39" to-port="0" />
686
+ <edge from-layer="38" from-port="0" to-layer="39" to-port="1" />
687
+ <edge from-layer="39" from-port="2" to-layer="41" to-port="3" />
688
+ <edge from-layer="40" from-port="0" to-layer="41" to-port="4" />
689
+ <edge from-layer="41" from-port="6" to-layer="42" to-port="0" />
690
+ <edge from-layer="41" from-port="5" to-layer="45" to-port="0" />
691
+ <edge from-layer="42" from-port="1" to-layer="43" to-port="0" />
692
+ <edge from-layer="43" from-port="1" to-layer="44" to-port="0" />
693
+ <edge from-layer="45" from-port="1" to-layer="46" to-port="0" />
694
+ </edges>
695
+ <rt_info>
696
+ <add_attention_mask value="True" />
697
+ <add_prefix_space />
698
+ <add_special_tokens value="True" />
699
+ <clean_up_tokenization_spaces />
700
+ <detokenizer_input_type value="i64" />
701
+ <eos_token_id value="151645" />
702
+ <handle_special_tokens_with_re />
703
+ <max_length />
704
+ <number_of_inputs value="1" />
705
+ <openvino_tokenizers_version value="2025.4.0.0-617-b2d52c80e23" />
706
+ <openvino_version value="2025.4.0-20128-67fddb40e31" />
707
+ <original_post_processor_template value="{&quot;type&quot;: &quot;ByteLevel&quot;, &quot;add_prefix_space&quot;: false, &quot;trim_offsets&quot;: false, &quot;use_regex&quot;: false}" />
708
+ <original_tokenizer_class value="&lt;class 'transformers.models.qwen2.tokenization_qwen2_fast.Qwen2TokenizerFast'>" />
709
+ <pad_token_id value="151643" />
710
+ <sentencepiece_version value="0.2.1" />
711
+ <skip_special_tokens value="True" />
712
+ <streaming_detokenizer value="False" />
713
+ <tiktoken_version value="0.12.0" />
714
+ <tokenizer_output_type value="i64" />
715
+ <tokenizers_version value="0.21.4" />
716
+ <transformers_version value="4.55.4" />
717
+ <use_max_padding value="False" />
718
+ <use_sentencepiece_backend value="False" />
719
+ <utf8_replace_mode value="replace" />
720
+ <with_detokenizer value="True" />
721
+ </rt_info>
722
+ </net>
special_tokens_map.json ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<|im_start|>",
4
+ "<|im_end|>",
5
+ "<|object_ref_start|>",
6
+ "<|object_ref_end|>",
7
+ "<|box_start|>",
8
+ "<|box_end|>",
9
+ "<|quad_start|>",
10
+ "<|quad_end|>",
11
+ "<|vision_start|>",
12
+ "<|vision_end|>",
13
+ "<|vision_pad|>",
14
+ "<|image_pad|>",
15
+ "<|video_pad|>"
16
+ ],
17
+ "eos_token": {
18
+ "content": "<|im_end|>",
19
+ "lstrip": false,
20
+ "normalized": false,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ },
24
+ "pad_token": {
25
+ "content": "<|endoftext|>",
26
+ "lstrip": false,
27
+ "normalized": false,
28
+ "rstrip": false,
29
+ "single_word": false
30
+ }
31
+ }
tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:aeb13307a71acd8fe81861d94ad54ab689df773318809eed3cbe794b4492dae4
3
+ size 11422654
tokenizer_config.json ADDED
@@ -0,0 +1,239 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": false,
3
+ "add_prefix_space": false,
4
+ "added_tokens_decoder": {
5
+ "151643": {
6
+ "content": "<|endoftext|>",
7
+ "lstrip": false,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false,
11
+ "special": true
12
+ },
13
+ "151644": {
14
+ "content": "<|im_start|>",
15
+ "lstrip": false,
16
+ "normalized": false,
17
+ "rstrip": false,
18
+ "single_word": false,
19
+ "special": true
20
+ },
21
+ "151645": {
22
+ "content": "<|im_end|>",
23
+ "lstrip": false,
24
+ "normalized": false,
25
+ "rstrip": false,
26
+ "single_word": false,
27
+ "special": true
28
+ },
29
+ "151646": {
30
+ "content": "<|object_ref_start|>",
31
+ "lstrip": false,
32
+ "normalized": false,
33
+ "rstrip": false,
34
+ "single_word": false,
35
+ "special": true
36
+ },
37
+ "151647": {
38
+ "content": "<|object_ref_end|>",
39
+ "lstrip": false,
40
+ "normalized": false,
41
+ "rstrip": false,
42
+ "single_word": false,
43
+ "special": true
44
+ },
45
+ "151648": {
46
+ "content": "<|box_start|>",
47
+ "lstrip": false,
48
+ "normalized": false,
49
+ "rstrip": false,
50
+ "single_word": false,
51
+ "special": true
52
+ },
53
+ "151649": {
54
+ "content": "<|box_end|>",
55
+ "lstrip": false,
56
+ "normalized": false,
57
+ "rstrip": false,
58
+ "single_word": false,
59
+ "special": true
60
+ },
61
+ "151650": {
62
+ "content": "<|quad_start|>",
63
+ "lstrip": false,
64
+ "normalized": false,
65
+ "rstrip": false,
66
+ "single_word": false,
67
+ "special": true
68
+ },
69
+ "151651": {
70
+ "content": "<|quad_end|>",
71
+ "lstrip": false,
72
+ "normalized": false,
73
+ "rstrip": false,
74
+ "single_word": false,
75
+ "special": true
76
+ },
77
+ "151652": {
78
+ "content": "<|vision_start|>",
79
+ "lstrip": false,
80
+ "normalized": false,
81
+ "rstrip": false,
82
+ "single_word": false,
83
+ "special": true
84
+ },
85
+ "151653": {
86
+ "content": "<|vision_end|>",
87
+ "lstrip": false,
88
+ "normalized": false,
89
+ "rstrip": false,
90
+ "single_word": false,
91
+ "special": true
92
+ },
93
+ "151654": {
94
+ "content": "<|vision_pad|>",
95
+ "lstrip": false,
96
+ "normalized": false,
97
+ "rstrip": false,
98
+ "single_word": false,
99
+ "special": true
100
+ },
101
+ "151655": {
102
+ "content": "<|image_pad|>",
103
+ "lstrip": false,
104
+ "normalized": false,
105
+ "rstrip": false,
106
+ "single_word": false,
107
+ "special": true
108
+ },
109
+ "151656": {
110
+ "content": "<|video_pad|>",
111
+ "lstrip": false,
112
+ "normalized": false,
113
+ "rstrip": false,
114
+ "single_word": false,
115
+ "special": true
116
+ },
117
+ "151657": {
118
+ "content": "<tool_call>",
119
+ "lstrip": false,
120
+ "normalized": false,
121
+ "rstrip": false,
122
+ "single_word": false,
123
+ "special": false
124
+ },
125
+ "151658": {
126
+ "content": "</tool_call>",
127
+ "lstrip": false,
128
+ "normalized": false,
129
+ "rstrip": false,
130
+ "single_word": false,
131
+ "special": false
132
+ },
133
+ "151659": {
134
+ "content": "<|fim_prefix|>",
135
+ "lstrip": false,
136
+ "normalized": false,
137
+ "rstrip": false,
138
+ "single_word": false,
139
+ "special": false
140
+ },
141
+ "151660": {
142
+ "content": "<|fim_middle|>",
143
+ "lstrip": false,
144
+ "normalized": false,
145
+ "rstrip": false,
146
+ "single_word": false,
147
+ "special": false
148
+ },
149
+ "151661": {
150
+ "content": "<|fim_suffix|>",
151
+ "lstrip": false,
152
+ "normalized": false,
153
+ "rstrip": false,
154
+ "single_word": false,
155
+ "special": false
156
+ },
157
+ "151662": {
158
+ "content": "<|fim_pad|>",
159
+ "lstrip": false,
160
+ "normalized": false,
161
+ "rstrip": false,
162
+ "single_word": false,
163
+ "special": false
164
+ },
165
+ "151663": {
166
+ "content": "<|repo_name|>",
167
+ "lstrip": false,
168
+ "normalized": false,
169
+ "rstrip": false,
170
+ "single_word": false,
171
+ "special": false
172
+ },
173
+ "151664": {
174
+ "content": "<|file_sep|>",
175
+ "lstrip": false,
176
+ "normalized": false,
177
+ "rstrip": false,
178
+ "single_word": false,
179
+ "special": false
180
+ },
181
+ "151665": {
182
+ "content": "<tool_response>",
183
+ "lstrip": false,
184
+ "normalized": false,
185
+ "rstrip": false,
186
+ "single_word": false,
187
+ "special": false
188
+ },
189
+ "151666": {
190
+ "content": "</tool_response>",
191
+ "lstrip": false,
192
+ "normalized": false,
193
+ "rstrip": false,
194
+ "single_word": false,
195
+ "special": false
196
+ },
197
+ "151667": {
198
+ "content": "<think>",
199
+ "lstrip": false,
200
+ "normalized": false,
201
+ "rstrip": false,
202
+ "single_word": false,
203
+ "special": false
204
+ },
205
+ "151668": {
206
+ "content": "</think>",
207
+ "lstrip": false,
208
+ "normalized": false,
209
+ "rstrip": false,
210
+ "single_word": false,
211
+ "special": false
212
+ }
213
+ },
214
+ "additional_special_tokens": [
215
+ "<|im_start|>",
216
+ "<|im_end|>",
217
+ "<|object_ref_start|>",
218
+ "<|object_ref_end|>",
219
+ "<|box_start|>",
220
+ "<|box_end|>",
221
+ "<|quad_start|>",
222
+ "<|quad_end|>",
223
+ "<|vision_start|>",
224
+ "<|vision_end|>",
225
+ "<|vision_pad|>",
226
+ "<|image_pad|>",
227
+ "<|video_pad|>"
228
+ ],
229
+ "bos_token": null,
230
+ "clean_up_tokenization_spaces": false,
231
+ "eos_token": "<|im_end|>",
232
+ "errors": "replace",
233
+ "extra_special_tokens": {},
234
+ "model_max_length": 131072,
235
+ "pad_token": "<|endoftext|>",
236
+ "split_special_tokens": false,
237
+ "tokenizer_class": "Qwen2Tokenizer",
238
+ "unk_token": null
239
+ }
vocab.json ADDED
The diff for this file is too large to render. See raw diff