openvino-ci commited on
Commit
3b7fcd4
·
verified ·
1 Parent(s): 78f75d6

Upload folder using huggingface_hub

Browse files
README.md CHANGED
@@ -1,21 +1,32 @@
1
  ---
 
 
 
 
 
 
 
2
  license: mit
3
- license_link: https://choosealicense.com/licenses/mit/
4
- base_model:
5
- - distil-whisper/distil-large-v2
6
  ---
7
- # distil-large-v2-fp16-ov
8
- * Model creator: [Whisper Distillation](https://huggingface.co/distil-whisper)
 
9
  * Original model: [distil-large-v2](https://huggingface.co/distil-whisper/distil-large-v2)
10
 
 
11
  ## Description
 
 
12
 
13
  ## Compatibility
14
 
15
  The provided OpenVINO™ IR model is compatible with:
16
 
17
- * OpenVINO version 2024.5.0 and higher
18
- * Optimum Intel 1.21.0 and higher
 
19
 
20
  ## Running Model Inference with [Optimum Intel](https://huggingface.co/docs/optimum/intel/index)
21
 
@@ -28,24 +39,25 @@ pip install optimum[openvino]
28
  2. Run model inference:
29
 
30
  ```
 
31
  from transformers import AutoProcessor
32
  from optimum.intel.openvino import OVModelForSpeechSeq2Seq
33
 
34
- model_id = "OpenVINO/distil-large-v2-fp16-ov"
35
  tokenizer = AutoProcessor.from_pretrained(model_id)
36
  model = OVModelForSpeechSeq2Seq.from_pretrained(model_id)
37
 
38
  dataset = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation", trust_remote_code=True)
39
  sample = dataset[0]
40
 
41
- input_features = processor(
42
  sample["audio"]["array"],
43
  sampling_rate=sample["audio"]["sampling_rate"],
44
  return_tensors="pt",
45
  ).input_features
46
 
47
  outputs = model.generate(input_features)
48
- text = processor.batch_decode(outputs)[0]
49
  print(text)
50
  ```
51
 
@@ -62,8 +74,8 @@ pip install -U --pre --extra-index-url https://storage.openvinotoolkit.org/simpl
62
  ```
63
  import huggingface_hub as hf_hub
64
 
65
- model_id = "OpenVINO/distil-large-v2-fp16-ov"
66
- model_path = "distil-large-v2-fp16-ov"
67
 
68
  hf_hub.snapshot_download(model_id, local_dir=model_path)
69
 
@@ -79,7 +91,7 @@ device = "CPU"
79
  pipe = ov_genai.WhisperPipeline(model_path, device)
80
 
81
  dataset = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation", trust_remote_code=True)
82
- sample = dataset[0]["audio]["array"]
83
  print(pipe.generate(sample))
84
  ```
85
 
@@ -91,7 +103,7 @@ Check the original model card for [original model card](https://huggingface.co/d
91
 
92
  ## Legal information
93
 
94
- The original model is distributed under [mit](https://choosealicense.com/licenses/mit/) license. More details can be found in [original model card](https://huggingface.co/distil-whisper/distil-large-v2).
95
 
96
  ## Disclaimer
97
 
 
1
  ---
2
+ language:
3
+ - en
4
+ tags:
5
+ - audio
6
+ - automatic-speech-recognition
7
+ - transformers.js
8
+ pipeline_tag: automatic-speech-recognition
9
  license: mit
10
+ license_link: https://github.com/huggingface/distil-whisper/blob/main/LICENSE
11
+ library_name: transformers
 
12
  ---
13
+
14
+ # distil-whisper-large-v2-fp16-ov
15
+ * Model creator: [Distil-whisper](https://huggingface.co/distil-whisper)
16
  * Original model: [distil-large-v2](https://huggingface.co/distil-whisper/distil-large-v2)
17
 
18
+
19
  ## Description
20
+ This is [distil-large-v2](https://huggingface.co/distil-whisper/distil-large-v2) model converted to the [OpenVINO™ IR](https://docs.openvino.ai/2025/documentation/openvino-ir-format.html) (Intermediate Representation) format with weights compressed to FP16.
21
+
22
 
23
  ## Compatibility
24
 
25
  The provided OpenVINO™ IR model is compatible with:
26
 
27
+ * OpenVINO version 2025.1.0 and higher
28
+ * Optimum Intel 1.23.0 and higher
29
+
30
 
31
  ## Running Model Inference with [Optimum Intel](https://huggingface.co/docs/optimum/intel/index)
32
 
 
39
  2. Run model inference:
40
 
41
  ```
42
+ from datasets import load_dataset
43
  from transformers import AutoProcessor
44
  from optimum.intel.openvino import OVModelForSpeechSeq2Seq
45
 
46
+ model_id = "OpenVINO/distil-whisper-large-v2-fp16-ov"
47
  tokenizer = AutoProcessor.from_pretrained(model_id)
48
  model = OVModelForSpeechSeq2Seq.from_pretrained(model_id)
49
 
50
  dataset = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation", trust_remote_code=True)
51
  sample = dataset[0]
52
 
53
+ input_features = tokenizer(
54
  sample["audio"]["array"],
55
  sampling_rate=sample["audio"]["sampling_rate"],
56
  return_tensors="pt",
57
  ).input_features
58
 
59
  outputs = model.generate(input_features)
60
+ text = tokenizer.batch_decode(outputs)[0]
61
  print(text)
62
  ```
63
 
 
74
  ```
75
  import huggingface_hub as hf_hub
76
 
77
+ model_id = "OpenVINO/distil-whisper-large-v2-fp16-ov"
78
+ model_path = "distil-whisper-large-v2-fp16-ov"
79
 
80
  hf_hub.snapshot_download(model_id, local_dir=model_path)
81
 
 
91
  pipe = ov_genai.WhisperPipeline(model_path, device)
92
 
93
  dataset = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation", trust_remote_code=True)
94
+ sample = dataset[0]["audio"]["array"]
95
  print(pipe.generate(sample))
96
  ```
97
 
 
103
 
104
  ## Legal information
105
 
106
+ The original model is distributed under [mit](https://github.com/huggingface/distil-whisper/blob/main/LICENSE) license. More details can be found in [original model card](https://huggingface.co/distil-whisper/distil-large-v2).
107
 
108
  ## Disclaimer
109
 
config.json CHANGED
@@ -1,6 +1,5 @@
1
  {
2
  "_attn_implementation_autoset": true,
3
- "_name_or_path": "OpenVINO/distil-whisper-large-v2-fp16-ov",
4
  "activation_dropout": 0.0,
5
  "activation_function": "gelu",
6
  "apply_spec_augment": false,
@@ -54,7 +53,13 @@
54
  "num_mel_bins": 80,
55
  "pad_token_id": 50257,
56
  "scale_embedding": false,
57
- "transformers_version": "4.46.3",
 
 
 
 
 
 
58
  "use_cache": true,
59
  "use_weighted_layer_sum": false,
60
  "vocab_size": 51865
 
1
  {
2
  "_attn_implementation_autoset": true,
 
3
  "activation_dropout": 0.0,
4
  "activation_function": "gelu",
5
  "apply_spec_augment": false,
 
53
  "num_mel_bins": 80,
54
  "pad_token_id": 50257,
55
  "scale_embedding": false,
56
+ "torch_dtype": "float32",
57
+ "transformers.js_config": {
58
+ "use_external_data_format": {
59
+ "encoder_model.onnx": true
60
+ }
61
+ },
62
+ "transformers_version": "4.51.3",
63
  "use_cache": true,
64
  "use_weighted_layer_sum": false,
65
  "vocab_size": 51865
generation_config.json CHANGED
@@ -222,6 +222,6 @@
222
  "transcribe": 50359,
223
  "translate": 50358
224
  },
225
- "transformers_version": "4.46.3",
226
  "use_scan": false
227
  }
 
222
  "transcribe": 50359,
223
  "translate": 50358
224
  },
225
+ "transformers_version": "4.51.3",
226
  "use_scan": false
227
  }
openvino_decoder_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:00af96fab694bfd689077a57cbd86826df37d6324d7e2716cb47c0c7092a54d9
3
- size 238871234
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:17191f85e4c93c0055651761397bac32383bdccad513e8a04afa01640ea197b4
3
+ size 477742248
openvino_decoder_model.xml CHANGED
The diff for this file is too large to render. See raw diff
 
openvino_detokenizer.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:369dbe475202046c9ed61f55e4c32ce00c4f821af815f8a7f70523fba0155491
3
- size 528297
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2542d1fe6c4c5d838e7b7c61b24996b94ae0cb54e64ff6dbaab2b9dd5ddd7ca0
3
+ size 736181
openvino_detokenizer.xml CHANGED
@@ -1,16 +1,16 @@
1
  <?xml version="1.0"?>
2
  <net name="detokenizer" version="11">
3
  <layers>
4
- <layer id="0" name="Parameter_43401" type="Parameter" version="opset1">
5
  <data shape="?,?" element_type="i64" />
6
  <output>
7
- <port id="0" precision="I64" names="Parameter_43401">
8
  <dim>-1</dim>
9
  <dim>-1</dim>
10
  </port>
11
  </output>
12
  </layer>
13
- <layer id="1" name="Convert_43417" type="Convert" version="opset1">
14
  <data destination_type="i32" />
15
  <input>
16
  <port id="0" precision="I64">
@@ -25,54 +25,59 @@
25
  </port>
26
  </output>
27
  </layer>
28
- <layer id="2" name="Constant_43356" type="Const" version="opset1">
29
- <data element_type="u8" shape="528248" offset="0" size="528248" />
30
  <output>
31
- <port id="0" precision="U8">
32
- <dim>528248</dim>
33
  </port>
34
  </output>
35
  </layer>
36
- <layer id="3" name="StringTensorUnpack_43357" type="StringTensorUnpack" version="extension">
37
- <data mode="begins_ends" />
38
- <input>
39
- <port id="0" precision="U8">
40
- <dim>528248</dim>
41
- </port>
42
- </input>
43
  <output>
44
- <port id="1" precision="I32">
45
- <dim>-1</dim>
46
  </port>
47
- <port id="2" precision="I32">
48
- <dim>-1</dim>
 
 
 
 
 
49
  </port>
50
- <port id="3" precision="U8">
51
- <dim>-1</dim>
 
 
 
 
 
52
  </port>
53
  </output>
54
  </layer>
55
- <layer id="4" name="VocabDecoder_43402" type="VocabDecoder" version="extension">
56
- <data skip_tokens="50257, 50258, 50259, 50260, 50261, 50262, 50263, 50264, 50265, 50266, 50267, 50268, 50269, 50270, 50271, 50272, 50273, 50274, 50275, 50276, 50277, 50278, 50279, 50280, 50281, 50282, 50283, 50284, 50285, 50286, 50287, 50288, 50289, 50290, 50291, 50292, 50293, 50294, 50295, 50296, 50297, 50298, 50299, 50300, 50301, 50302, 50303, 50304, 50305, 50306, 50307, 50308, 50309, 50310, 50311, 50312, 50313, 50314, 50315, 50316, 50317, 50318, 50319, 50320, 50321, 50322, 50323, 50324, 50325, 50326, 50327, 50328, 50329, 50330, 50331, 50332, 50333, 50334, 50335, 50336, 50337, 50338, 50339, 50340, 50341, 50342, 50343, 50344, 50345, 50346, 50347, 50348, 50349, 50350, 50351, 50352, 50353, 50354, 50355, 50356, 50357, 50358, 50359, 50360, 50361, 50362, 50363" />
57
  <input>
58
  <port id="0" precision="I32">
59
  <dim>-1</dim>
60
  <dim>-1</dim>
61
  </port>
62
  <port id="1" precision="I32">
63
- <dim>-1</dim>
64
  </port>
65
  <port id="2" precision="I32">
66
- <dim>-1</dim>
67
  </port>
68
  <port id="3" precision="U8">
69
- <dim>-1</dim>
70
  </port>
71
- </input>
72
- <output>
73
  <port id="4" precision="I32">
74
- <dim>-1</dim>
75
  </port>
 
 
76
  <port id="5" precision="I32">
77
  <dim>-1</dim>
78
  </port>
@@ -82,12 +87,15 @@
82
  <port id="7" precision="I32">
83
  <dim>-1</dim>
84
  </port>
85
- <port id="8" precision="U8">
 
 
 
86
  <dim>-1</dim>
87
  </port>
88
  </output>
89
  </layer>
90
- <layer id="5" name="FuzeRagged_43403" type="FuzeRagged" version="extension">
91
  <input>
92
  <port id="0" precision="I32">
93
  <dim>-1</dim>
@@ -111,23 +119,48 @@
111
  </port>
112
  </output>
113
  </layer>
114
- <layer id="6" name="Constant_43405" type="Const" version="opset1">
115
- <data element_type="u8" shape="47" offset="528248" size="47" />
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
116
  <output>
117
  <port id="0" precision="U8">
118
- <dim>47</dim>
119
  </port>
120
  </output>
121
  </layer>
122
- <layer id="7" name="Constant_43407" type="Const" version="opset1">
123
- <data element_type="u8" shape="2" offset="528295" size="2" />
124
  <output>
125
  <port id="0" precision="U8">
126
  <dim>2</dim>
127
  </port>
128
  </output>
129
  </layer>
130
- <layer id="8" name="RegexNormalization_43408" type="RegexNormalization" version="extension">
131
  <data global_replace="true" />
132
  <input>
133
  <port id="0" precision="I32">
@@ -140,7 +173,7 @@
140
  <dim>-1</dim>
141
  </port>
142
  <port id="3" precision="U8">
143
- <dim>47</dim>
144
  </port>
145
  <port id="4" precision="U8">
146
  <dim>2</dim>
@@ -158,8 +191,7 @@
158
  </port>
159
  </output>
160
  </layer>
161
- <layer id="9" name="StringTensorPack_43409" type="StringTensorPack" version="extension">
162
- <data mode="begins_ends" />
163
  <input>
164
  <port id="0" precision="I32">
165
  <dim>-1</dim>
@@ -172,12 +204,12 @@
172
  </port>
173
  </input>
174
  <output>
175
- <port id="3" precision="STRING" names="string_output">
176
  <dim>-1</dim>
177
  </port>
178
  </output>
179
  </layer>
180
- <layer id="10" name="Result_43410" type="Result" version="opset1">
181
  <input>
182
  <port id="0" precision="STRING">
183
  <dim>-1</dim>
@@ -187,24 +219,27 @@
187
  </layers>
188
  <edges>
189
  <edge from-layer="0" from-port="0" to-layer="1" to-port="0" />
190
- <edge from-layer="1" from-port="1" to-layer="4" to-port="0" />
191
- <edge from-layer="2" from-port="0" to-layer="3" to-port="0" />
192
- <edge from-layer="3" from-port="1" to-layer="4" to-port="1" />
193
- <edge from-layer="3" from-port="2" to-layer="4" to-port="2" />
194
- <edge from-layer="3" from-port="3" to-layer="4" to-port="3" />
195
- <edge from-layer="4" from-port="7" to-layer="5" to-port="3" />
196
- <edge from-layer="4" from-port="8" to-layer="8" to-port="2" />
197
- <edge from-layer="4" from-port="6" to-layer="5" to-port="2" />
198
- <edge from-layer="4" from-port="5" to-layer="5" to-port="1" />
199
- <edge from-layer="4" from-port="4" to-layer="5" to-port="0" />
200
- <edge from-layer="5" from-port="4" to-layer="8" to-port="0" />
201
- <edge from-layer="5" from-port="5" to-layer="8" to-port="1" />
202
- <edge from-layer="6" from-port="0" to-layer="8" to-port="3" />
203
- <edge from-layer="7" from-port="0" to-layer="8" to-port="4" />
204
- <edge from-layer="8" from-port="5" to-layer="9" to-port="0" />
205
- <edge from-layer="8" from-port="6" to-layer="9" to-port="1" />
206
- <edge from-layer="8" from-port="7" to-layer="9" to-port="2" />
207
- <edge from-layer="9" from-port="3" to-layer="10" to-port="0" />
 
 
 
208
  </edges>
209
  <rt_info>
210
  <add_attention_mask value="True" />
@@ -215,21 +250,22 @@
215
  <detokenizer_input_type value="i64" />
216
  <eos_token_id value="50257" />
217
  <handle_special_tokens_with_re />
 
218
  <number_of_inputs value="1" />
219
- <openvino_tokenizers_version value="2024.5.0.0" />
220
- <openvino_version value="2024.5.0" />
 
221
  <original_tokenizer_class value="&lt;class 'transformers.models.whisper.tokenization_whisper_fast.WhisperTokenizerFast'>" />
222
  <pad_token_id value="50257" />
223
- <sentencepiece_version value="0.2.0" />
224
  <skip_special_tokens value="True" />
225
  <streaming_detokenizer value="False" />
226
- <tiktoken_version value="0.8.0" />
227
  <tokenizer_output_type value="i64" />
228
- <tokenizers_version value="0.20.3" />
229
- <transformers_version value="4.46.3" />
230
  <use_max_padding value="False" />
231
  <use_sentencepiece_backend value="False" />
232
- <utf8_replace_mode />
233
  <with_detokenizer value="True" />
234
  </rt_info>
235
  </net>
 
1
  <?xml version="1.0"?>
2
  <net name="detokenizer" version="11">
3
  <layers>
4
+ <layer id="0" name="Parameter_36326" type="Parameter" version="opset1">
5
  <data shape="?,?" element_type="i64" />
6
  <output>
7
+ <port id="0" precision="I64" names="Parameter_36326">
8
  <dim>-1</dim>
9
  <dim>-1</dim>
10
  </port>
11
  </output>
12
  </layer>
13
+ <layer id="1" name="Convert_36521" type="Convert" version="opset1">
14
  <data destination_type="i32" />
15
  <input>
16
  <port id="0" precision="I64">
 
25
  </port>
26
  </output>
27
  </layer>
28
+ <layer id="2" name="Constant_36328" type="Const" version="opset1">
29
+ <data element_type="i32" shape="51865" offset="0" size="207460" />
30
  <output>
31
+ <port id="0" precision="I32">
32
+ <dim>51865</dim>
33
  </port>
34
  </output>
35
  </layer>
36
+ <layer id="3" name="Constant_36330" type="Const" version="opset1">
37
+ <data element_type="i32" shape="51865" offset="207460" size="207460" />
 
 
 
 
 
38
  <output>
39
+ <port id="0" precision="I32">
40
+ <dim>51865</dim>
41
  </port>
42
+ </output>
43
+ </layer>
44
+ <layer id="4" name="Constant_36332" type="Const" version="opset1">
45
+ <data element_type="u8" shape="320780" offset="414920" size="320780" />
46
+ <output>
47
+ <port id="0" precision="U8">
48
+ <dim>320780</dim>
49
  </port>
50
+ </output>
51
+ </layer>
52
+ <layer id="5" name="Slice_36337" type="Const" version="opset1">
53
+ <data element_type="i32" shape="107" offset="735700" size="428" />
54
+ <output>
55
+ <port id="0" precision="I32">
56
+ <dim>107</dim>
57
  </port>
58
  </output>
59
  </layer>
60
+ <layer id="6" name="VocabDecoder_36339" 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>51865</dim>
69
  </port>
70
  <port id="2" precision="I32">
71
+ <dim>51865</dim>
72
  </port>
73
  <port id="3" precision="U8">
74
+ <dim>320780</dim>
75
  </port>
 
 
76
  <port id="4" precision="I32">
77
+ <dim>107</dim>
78
  </port>
79
+ </input>
80
+ <output>
81
  <port id="5" precision="I32">
82
  <dim>-1</dim>
83
  </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_36340" type="FuzeRagged" version="extension">
99
  <input>
100
  <port id="0" precision="I32">
101
  <dim>-1</dim>
 
119
  </port>
120
  </output>
121
  </layer>
122
+ <layer id="8" name="UTF8Validate_36341" 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="Constant_36343" type="Const" version="opset1">
148
+ <data element_type="u8" shape="51" offset="736128" size="51" />
149
  <output>
150
  <port id="0" precision="U8">
151
+ <dim>51</dim>
152
  </port>
153
  </output>
154
  </layer>
155
+ <layer id="10" name="Constant_36345" type="Const" version="opset1">
156
+ <data element_type="u8" shape="2" offset="736179" size="2" />
157
  <output>
158
  <port id="0" precision="U8">
159
  <dim>2</dim>
160
  </port>
161
  </output>
162
  </layer>
163
+ <layer id="11" name="RegexNormalization_36346" type="RegexNormalization" version="extension">
164
  <data global_replace="true" />
165
  <input>
166
  <port id="0" precision="I32">
 
173
  <dim>-1</dim>
174
  </port>
175
  <port id="3" precision="U8">
176
+ <dim>51</dim>
177
  </port>
178
  <port id="4" precision="U8">
179
  <dim>2</dim>
 
191
  </port>
192
  </output>
193
  </layer>
194
+ <layer id="12" name="StringTensorPack_36347" type="StringTensorPack" version="opset15">
 
195
  <input>
196
  <port id="0" precision="I32">
197
  <dim>-1</dim>
 
204
  </port>
205
  </input>
206
  <output>
207
+ <port id="3" precision="STRING" names="Result_36348,string_output">
208
  <dim>-1</dim>
209
  </port>
210
  </output>
211
  </layer>
212
+ <layer id="13" name="Result_36348" type="Result" version="opset1" output_names="Result_36348,string_output">
213
  <input>
214
  <port id="0" precision="STRING">
215
  <dim>-1</dim>
 
219
  </layers>
220
  <edges>
221
  <edge from-layer="0" from-port="0" to-layer="1" to-port="0" />
222
+ <edge from-layer="1" from-port="1" to-layer="6" to-port="0" />
223
+ <edge from-layer="2" from-port="0" to-layer="6" to-port="1" />
224
+ <edge from-layer="3" from-port="0" to-layer="6" to-port="2" />
225
+ <edge from-layer="4" from-port="0" to-layer="6" to-port="3" />
226
+ <edge from-layer="5" from-port="0" to-layer="6" to-port="4" />
227
+ <edge from-layer="6" from-port="5" to-layer="7" to-port="0" />
228
+ <edge from-layer="6" from-port="6" to-layer="7" to-port="1" />
229
+ <edge from-layer="6" from-port="7" to-layer="7" to-port="2" />
230
+ <edge from-layer="6" from-port="8" to-layer="7" to-port="3" />
231
+ <edge from-layer="6" from-port="9" to-layer="8" to-port="2" />
232
+ <edge from-layer="7" from-port="5" to-layer="8" to-port="1" />
233
+ <edge from-layer="7" from-port="4" to-layer="8" to-port="0" />
234
+ <edge from-layer="8" from-port="3" to-layer="11" to-port="0" />
235
+ <edge from-layer="8" from-port="4" to-layer="11" to-port="1" />
236
+ <edge from-layer="8" from-port="5" to-layer="11" to-port="2" />
237
+ <edge from-layer="9" from-port="0" to-layer="11" to-port="3" />
238
+ <edge from-layer="10" from-port="0" to-layer="11" to-port="4" />
239
+ <edge from-layer="11" from-port="5" to-layer="12" to-port="0" />
240
+ <edge from-layer="11" from-port="6" to-layer="12" to-port="1" />
241
+ <edge from-layer="11" from-port="7" to-layer="12" to-port="2" />
242
+ <edge from-layer="12" from-port="3" to-layer="13" to-port="0" />
243
  </edges>
244
  <rt_info>
245
  <add_attention_mask value="True" />
 
250
  <detokenizer_input_type value="i64" />
251
  <eos_token_id value="50257" />
252
  <handle_special_tokens_with_re />
253
+ <max_length />
254
  <number_of_inputs value="1" />
255
+ <openvino_tokenizers_version value="2025.2.0.0-565-130827ab189" />
256
+ <openvino_version value="2025.2.0-19120-87425bc78ca-releases/2025/2" />
257
+ <original_post_processor_template value="{&quot;type&quot;: &quot;TemplateProcessing&quot;, &quot;single&quot;: [{&quot;SpecialToken&quot;: {&quot;id&quot;: &quot;&lt;|startoftranscript|>&quot;, &quot;type_id&quot;: 0}}, {&quot;SpecialToken&quot;: {&quot;id&quot;: &quot;&lt;|en|>&quot;, &quot;type_id&quot;: 0}}, {&quot;SpecialToken&quot;: {&quot;id&quot;: &quot;&lt;|transcribe|>&quot;, &quot;type_id&quot;: 0}}, {&quot;SpecialToken&quot;: {&quot;id&quot;: &quot;&lt;|notimestamps|>&quot;, &quot;type_id&quot;: 0}}, {&quot;Sequence&quot;: {&quot;id&quot;: &quot;A&quot;, &quot;type_id&quot;: 0}}, {&quot;SpecialToken&quot;: {&quot;id&quot;: &quot;&lt;|endoftext|>&quot;, &quot;type_id&quot;: 0}}], &quot;pair&quot;: [{&quot;SpecialToken&quot;: {&quot;id&quot;: &quot;&lt;|startoftranscript|>&quot;, &quot;type_id&quot;: 0}}, {&quot;SpecialToken&quot;: {&quot;id&quot;: &quot;&lt;|en|>&quot;, &quot;type_id&quot;: 0}}, {&quot;SpecialToken&quot;: {&quot;id&quot;: &quot;&lt;|transcribe|>&quot;, &quot;type_id&quot;: 0}}, {&quot;SpecialToken&quot;: {&quot;id&quot;: &quot;&lt;|notimestamps|>&quot;, &quot;type_id&quot;: 0}}, {&quot;Sequence&quot;: {&quot;id&quot;: &quot;A&quot;, &quot;type_id&quot;: 0}}, {&quot;Sequence&quot;: {&quot;id&quot;: &quot;B&quot;, &quot;type_id&quot;: 1}}, {&quot;SpecialToken&quot;: {&quot;id&quot;: &quot;&lt;|endoftext|>&quot;, &quot;type_id&quot;: 1}}], &quot;special_tokens&quot;: {&quot;&lt;|endoftext|>&quot;: {&quot;id&quot;: &quot;&lt;|endoftext|>&quot;, &quot;ids&quot;: [50257], &quot;tokens&quot;: [&quot;&lt;|endoftext|>&quot;]}, &quot;&lt;|en|>&quot;: {&quot;id&quot;: &quot;&lt;|en|>&quot;, &quot;ids&quot;: [50259], &quot;tokens&quot;: [&quot;&lt;|en|>&quot;]}, &quot;&lt;|notimestamps|>&quot;: {&quot;id&quot;: &quot;&lt;|notimestamps|>&quot;, &quot;ids&quot;: [50363], &quot;tokens&quot;: [&quot;&lt;|notimestamps|>&quot;]}, &quot;&lt;|startoftranscript|>&quot;: {&quot;id&quot;: &quot;&lt;|startoftranscript|>&quot;, &quot;ids&quot;: [50258], &quot;tokens&quot;: [&quot;&lt;|startoftranscript|>&quot;]}, &quot;&lt;|transcribe|>&quot;: {&quot;id&quot;: &quot;&lt;|transcribe|>&quot;, &quot;ids&quot;: [50359], &quot;tokens&quot;: [&quot;&lt;|transcribe|>&quot;]}}}" />
258
  <original_tokenizer_class value="&lt;class 'transformers.models.whisper.tokenization_whisper_fast.WhisperTokenizerFast'>" />
259
  <pad_token_id value="50257" />
260
+ <processed_post_processor_template value="{&quot;single&quot;: {&quot;ids&quot;: [50258, 50259, 50359, 50363, -1, 50257], &quot;type_ids&quot;: [0, 0, 0, 0, 0, 0]}, &quot;pair&quot;: {&quot;ids&quot;: [50258, 50259, 50359, 50363, -1, -2, 50257], &quot;type_ids&quot;: [0, 0, 0, 0, 0, 1, 1]}}" />
261
  <skip_special_tokens value="True" />
262
  <streaming_detokenizer value="False" />
 
263
  <tokenizer_output_type value="i64" />
264
+ <tokenizers_version value="0.21.1" />
265
+ <transformers_version value="4.51.3" />
266
  <use_max_padding value="False" />
267
  <use_sentencepiece_backend value="False" />
268
+ <utf8_replace_mode value="replace" />
269
  <with_detokenizer value="True" />
270
  </rt_info>
271
  </net>
openvino_encoder_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:97144d935de784916946a39840bc2222450abda523dca83a8c057d6e42f9c254
3
- size 1273569380
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c152639430fc0eea3cd0340a11c9a271f6ab83bc5d758946866fc949447386c1
3
+ size 2547138660
openvino_encoder_model.xml CHANGED
The diff for this file is too large to render. See raw diff
 
openvino_tokenizer.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:f98cfe9f9db7c4318bbf2d4f051c116f4a08df8e16cd0f2be19c36e261f1ec87
3
- size 1288265
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a42843d4d6b3060006af58ea6226cc97b0755bbccf9d9922d52b00ec55e88b09
3
+ size 1902145
openvino_tokenizer.xml CHANGED
@@ -1,27 +1,27 @@
1
  <?xml version="1.0"?>
2
  <net name="tokenizer" version="11">
3
  <layers>
4
- <layer id="0" name="Parameter_43275" type="Parameter" version="opset1">
5
  <data shape="?" element_type="string" />
6
  <output>
7
- <port id="0" precision="STRING" names="Parameter_43275">
8
  <dim>-1</dim>
9
  </port>
10
  </output>
11
  </layer>
12
- <layer id="1" name="Constant_43373" type="Const" version="opset1">
13
  <data element_type="i32" shape="" offset="0" size="4" />
14
  <output>
15
  <port id="0" precision="I32" />
16
  </output>
17
  </layer>
18
- <layer id="2" name="Constant_43374" type="Const" version="opset1">
19
  <data element_type="i32" shape="" offset="4" size="4" />
20
  <output>
21
  <port id="0" precision="I32" />
22
  </output>
23
  </layer>
24
- <layer id="3" name="Constant_43375" type="Const" version="opset1">
25
  <data element_type="i32" shape="1" offset="8" size="4" />
26
  <output>
27
  <port id="0" precision="I32">
@@ -29,19 +29,19 @@
29
  </port>
30
  </output>
31
  </layer>
32
- <layer id="4" name="Constant_43376" type="Const" version="opset1">
33
  <data element_type="i32" shape="" offset="0" size="4" />
34
  <output>
35
  <port id="0" precision="I32" />
36
  </output>
37
  </layer>
38
- <layer id="5" name="Constant_43377" type="Const" version="opset1">
39
  <data element_type="i32" shape="" offset="4" size="4" />
40
  <output>
41
  <port id="0" precision="I32" />
42
  </output>
43
  </layer>
44
- <layer id="6" name="Constant_43378" type="Const" version="opset1">
45
  <data element_type="i32" shape="1" offset="12" size="4" />
46
  <output>
47
  <port id="0" precision="I32">
@@ -49,19 +49,19 @@
49
  </port>
50
  </output>
51
  </layer>
52
- <layer id="7" name="Constant_43379" type="Const" version="opset1">
53
  <data element_type="i32" shape="" offset="0" size="4" />
54
  <output>
55
  <port id="0" precision="I32" />
56
  </output>
57
  </layer>
58
- <layer id="8" name="Constant_43380" type="Const" version="opset1">
59
  <data element_type="i32" shape="" offset="4" size="4" />
60
  <output>
61
  <port id="0" precision="I32" />
62
  </output>
63
  </layer>
64
- <layer id="9" name="Constant_43381" type="Const" version="opset1">
65
  <data element_type="i32" shape="1" offset="16" size="4" />
66
  <output>
67
  <port id="0" precision="I32">
@@ -69,19 +69,19 @@
69
  </port>
70
  </output>
71
  </layer>
72
- <layer id="10" name="Constant_43382" type="Const" version="opset1">
73
  <data element_type="i32" shape="" offset="0" size="4" />
74
  <output>
75
  <port id="0" precision="I32" />
76
  </output>
77
  </layer>
78
- <layer id="11" name="Constant_43383" type="Const" version="opset1">
79
  <data element_type="i32" shape="" offset="4" size="4" />
80
  <output>
81
  <port id="0" precision="I32" />
82
  </output>
83
  </layer>
84
- <layer id="12" name="Constant_43384" type="Const" version="opset1">
85
  <data element_type="i32" shape="1" offset="20" size="4" />
86
  <output>
87
  <port id="0" precision="I32">
@@ -89,14 +89,13 @@
89
  </port>
90
  </output>
91
  </layer>
92
- <layer id="13" name="Constant_43281" type="Const" version="opset1">
93
  <data element_type="i64" shape="" offset="24" size="8" />
94
  <output>
95
  <port id="0" precision="I64" />
96
  </output>
97
  </layer>
98
- <layer id="14" name="StringTensorUnpack_43276" type="StringTensorUnpack" version="extension">
99
- <data mode="begins_ends" />
100
  <input>
101
  <port id="0" precision="STRING">
102
  <dim>-1</dim>
@@ -114,7 +113,7 @@
114
  </port>
115
  </output>
116
  </layer>
117
- <layer id="15" name="ShapeOf_43277" type="ShapeOf" version="opset3">
118
  <data output_type="i64" />
119
  <input>
120
  <port id="0" precision="I32">
@@ -127,19 +126,19 @@
127
  </port>
128
  </output>
129
  </layer>
130
- <layer id="16" name="Constant_43278" type="Const" version="opset1">
131
  <data element_type="i64" shape="" offset="24" size="8" />
132
  <output>
133
  <port id="0" precision="I64" />
134
  </output>
135
  </layer>
136
- <layer id="17" name="Constant_43279" type="Const" version="opset1">
137
  <data element_type="i64" shape="" offset="24" size="8" />
138
  <output>
139
  <port id="0" precision="I64" />
140
  </output>
141
  </layer>
142
- <layer id="18" name="Gather_43280" type="Gather" version="opset8">
143
  <data batch_dims="0" />
144
  <input>
145
  <port id="0" precision="I64">
@@ -152,13 +151,13 @@
152
  <port id="3" precision="I64" />
153
  </output>
154
  </layer>
155
- <layer id="19" name="Constant_43282" type="Const" version="opset1">
156
  <data element_type="i64" shape="" offset="32" size="8" />
157
  <output>
158
  <port id="0" precision="I64" />
159
  </output>
160
  </layer>
161
- <layer id="20" name="Range_43283" type="Range" version="opset4">
162
  <data output_type="i32" />
163
  <input>
164
  <port id="0" precision="I64" />
@@ -171,19 +170,19 @@
171
  </port>
172
  </output>
173
  </layer>
174
- <layer id="21" name="Constant_43284" type="Const" version="opset1">
175
  <data element_type="i64" shape="" offset="32" size="8" />
176
  <output>
177
  <port id="0" precision="I64" />
178
  </output>
179
  </layer>
180
- <layer id="22" name="Constant_43285" type="Const" version="opset1">
181
  <data element_type="i64" shape="" offset="32" size="8" />
182
  <output>
183
  <port id="0" precision="I64" />
184
  </output>
185
  </layer>
186
- <layer id="23" name="Add_43286" type="Add" version="opset1">
187
  <data auto_broadcast="numpy" />
188
  <input>
189
  <port id="0" precision="I64" />
@@ -193,13 +192,13 @@
193
  <port id="2" precision="I64" />
194
  </output>
195
  </layer>
196
- <layer id="24" name="Constant_43287" type="Const" version="opset1">
197
  <data element_type="i64" shape="" offset="32" size="8" />
198
  <output>
199
  <port id="0" precision="I64" />
200
  </output>
201
  </layer>
202
- <layer id="25" name="Range_43288" type="Range" version="opset4">
203
  <data output_type="i32" />
204
  <input>
205
  <port id="0" precision="I64" />
@@ -212,7 +211,7 @@
212
  </port>
213
  </output>
214
  </layer>
215
- <layer id="26" name="Constant_43350" type="Const" version="opset1">
216
  <data element_type="u8" shape="26479" offset="40" size="26479" />
217
  <output>
218
  <port id="0" precision="U8">
@@ -220,7 +219,7 @@
220
  </port>
221
  </output>
222
  </layer>
223
- <layer id="27" name="SpecialTokensSplit_43351" type="SpecialTokensSplit" version="extension">
224
  <input>
225
  <port id="0" precision="I32">
226
  <dim>-1</dim>
@@ -262,15 +261,15 @@
262
  </port>
263
  </output>
264
  </layer>
265
- <layer id="28" name="Constant_43353" type="Const" version="opset1">
266
- <data element_type="u8" shape="64" offset="26519" size="64" />
267
  <output>
268
  <port id="0" precision="U8">
269
- <dim>64</dim>
270
  </port>
271
  </output>
272
  </layer>
273
- <layer id="29" name="RegexSplit_43354" type="RegexSplit" version="extension">
274
  <data behaviour="isolate" invert="false" max_splits="-1" />
275
  <input>
276
  <port id="0" precision="I32">
@@ -292,7 +291,7 @@
292
  <dim>-1</dim>
293
  </port>
294
  <port id="6" precision="U8">
295
- <dim>64</dim>
296
  </port>
297
  </input>
298
  <output>
@@ -316,123 +315,111 @@
316
  </port>
317
  </output>
318
  </layer>
319
- <layer id="30" name="Constant_43356" type="Const" version="opset1">
320
- <data element_type="u8" shape="528248" offset="26583" size="528248" />
321
  <output>
322
- <port id="0" precision="U8">
323
- <dim>528248</dim>
324
  </port>
325
  </output>
326
  </layer>
327
- <layer id="31" name="StringTensorUnpack_43357" type="StringTensorUnpack" version="extension">
328
- <data mode="begins_ends" />
329
- <input>
330
- <port id="0" precision="U8">
331
- <dim>528248</dim>
332
- </port>
333
- </input>
334
  <output>
335
- <port id="1" precision="I32">
336
- <dim>-1</dim>
337
- </port>
338
- <port id="2" precision="I32">
339
- <dim>-1</dim>
340
- </port>
341
- <port id="3" precision="U8">
342
- <dim>-1</dim>
343
  </port>
344
  </output>
345
  </layer>
346
- <layer id="32" name="Constant_43362" type="Const" version="opset1">
347
- <data element_type="u8" shape="369953" offset="554831" size="369953" />
348
  <output>
349
  <port id="0" precision="U8">
350
- <dim>369953</dim>
351
  </port>
352
  </output>
353
  </layer>
354
- <layer id="33" name="StringTensorUnpack_43363" type="StringTensorUnpack" version="extension">
355
- <data mode="begins_ends" />
356
- <input>
357
- <port id="0" precision="U8">
358
- <dim>369953</dim>
359
- </port>
360
- </input>
361
  <output>
362
- <port id="1" precision="I32">
363
- <dim>-1</dim>
364
- </port>
365
- <port id="2" precision="I32">
366
- <dim>-1</dim>
367
- </port>
368
- <port id="3" precision="U8">
369
- <dim>-1</dim>
370
  </port>
371
  </output>
372
  </layer>
373
- <layer id="34" name="Constant_43365" type="Const" version="opset1">
374
- <data element_type="u8" shape="336854" offset="924784" size="336854" />
375
  <output>
376
- <port id="0" precision="U8">
377
- <dim>336854</dim>
378
  </port>
379
  </output>
380
  </layer>
381
- <layer id="35" name="StringTensorUnpack_43366" type="StringTensorUnpack" version="extension">
382
- <data mode="begins_ends" />
383
- <input>
384
  <port id="0" precision="U8">
385
- <dim>336854</dim>
386
  </port>
387
- </input>
 
 
 
388
  <output>
389
- <port id="1" precision="I32">
390
- <dim>-1</dim>
391
- </port>
392
- <port id="2" precision="I32">
393
- <dim>-1</dim>
394
- </port>
395
- <port id="3" precision="U8">
396
- <dim>-1</dim>
397
  </port>
398
  </output>
399
  </layer>
400
- <layer id="36" name="Constant_43359" type="Const" version="opset1">
401
- <data element_type="u8" shape="20163" offset="1261638" size="20163" />
402
  <output>
403
- <port id="0" precision="U8">
404
- <dim>20163</dim>
405
  </port>
406
  </output>
407
  </layer>
408
- <layer id="37" name="StringTensorUnpack_43360" type="StringTensorUnpack" version="extension">
409
- <data mode="begins_ends" />
410
- <input>
411
  <port id="0" precision="U8">
412
- <dim>20163</dim>
413
  </port>
414
- </input>
 
 
 
415
  <output>
416
- <port id="1" precision="I32">
417
- <dim>-1</dim>
418
  </port>
419
- <port id="2" precision="I32">
420
- <dim>-1</dim>
 
 
 
 
 
421
  </port>
422
- <port id="3" precision="U8">
423
- <dim>-1</dim>
 
 
 
 
 
424
  </port>
425
  </output>
426
  </layer>
427
- <layer id="38" name="Constant_43367" type="Const" version="opset1">
428
- <data element_type="i32" shape="1608" offset="1281801" size="6432" />
429
  <output>
430
  <port id="0" precision="I32">
431
  <dim>1608</dim>
432
  </port>
433
  </output>
434
  </layer>
435
- <layer id="39" name="BPETokenizer_43368" type="BPETokenizer" version="extension">
436
  <data unk_token="" fuse_unk="false" suffix_indicator="" end_suffix="" byte_fallback="false" cache_capacity="20000" />
437
  <input>
438
  <port id="0" precision="I32">
@@ -451,40 +438,40 @@
451
  <dim>-1</dim>
452
  </port>
453
  <port id="5" precision="I32">
454
- <dim>-1</dim>
455
  </port>
456
  <port id="6" precision="I32">
457
- <dim>-1</dim>
458
  </port>
459
  <port id="7" precision="U8">
460
- <dim>-1</dim>
461
  </port>
462
  <port id="8" precision="I32">
463
- <dim>-1</dim>
464
  </port>
465
  <port id="9" precision="I32">
466
- <dim>-1</dim>
467
  </port>
468
  <port id="10" precision="U8">
469
- <dim>-1</dim>
470
  </port>
471
  <port id="11" precision="I32">
472
- <dim>-1</dim>
473
  </port>
474
  <port id="12" precision="I32">
475
- <dim>-1</dim>
476
  </port>
477
  <port id="13" precision="U8">
478
- <dim>-1</dim>
479
  </port>
480
  <port id="14" precision="I32">
481
- <dim>-1</dim>
482
  </port>
483
  <port id="15" precision="I32">
484
- <dim>-1</dim>
485
  </port>
486
  <port id="16" precision="U8">
487
- <dim>-1</dim>
488
  </port>
489
  <port id="17" precision="I32">
490
  <dim>1608</dim>
@@ -502,44 +489,30 @@
502
  </port>
503
  </output>
504
  </layer>
505
- <layer id="40" name="Subtract_43369" type="Subtract" version="opset1">
506
- <data auto_broadcast="numpy" />
507
- <input>
508
- <port id="0" precision="I32">
509
- <dim>-1</dim>
510
- </port>
511
- <port id="1" precision="I32">
512
- <dim>-1</dim>
513
- </port>
514
- </input>
515
  <output>
516
- <port id="2" precision="I32">
517
- <dim>-1</dim>
518
- </port>
519
  </output>
520
  </layer>
521
- <layer id="41" name="Constant_43370" type="Const" version="opset1">
522
- <data element_type="i32" shape="" offset="1288233" size="4" />
523
  <output>
524
- <port id="0" precision="I32" />
 
 
525
  </output>
526
  </layer>
527
- <layer id="42" name="Minimum_43371" type="Minimum" version="opset1">
528
- <data auto_broadcast="numpy" />
529
- <input>
530
- <port id="0" precision="I32">
531
- <dim>-1</dim>
532
- </port>
533
- <port id="1" precision="I32" />
534
- </input>
535
  <output>
536
- <port id="2" precision="I32">
537
- <dim>-1</dim>
538
  </port>
539
  </output>
540
  </layer>
541
- <layer id="43" name="Add_43372" type="Add" version="opset1">
542
- <data auto_broadcast="numpy" />
543
  <input>
544
  <port id="0" precision="I32">
545
  <dim>-1</dim>
@@ -547,42 +520,58 @@
547
  <port id="1" precision="I32">
548
  <dim>-1</dim>
549
  </port>
 
 
 
 
 
 
 
 
 
 
550
  </input>
551
  <output>
552
- <port id="2" precision="I32">
 
 
 
 
 
 
553
  <dim>-1</dim>
554
  </port>
555
  </output>
556
  </layer>
557
- <layer id="44" name="Constant_43385" type="Const" version="opset1">
558
  <data element_type="i32" shape="" offset="0" size="4" />
559
  <output>
560
  <port id="0" precision="I32" />
561
  </output>
562
  </layer>
563
- <layer id="45" name="Constant_43386" type="Const" version="opset1">
564
  <data element_type="i32" shape="" offset="4" size="4" />
565
  <output>
566
  <port id="0" precision="I32" />
567
  </output>
568
  </layer>
569
- <layer id="46" name="Constant_43387" type="Const" version="opset1">
570
- <data element_type="i32" shape="1" offset="1288237" size="4" />
571
  <output>
572
  <port id="0" precision="I32">
573
  <dim>1</dim>
574
  </port>
575
  </output>
576
  </layer>
577
- <layer id="47" name="Constant_43388" type="Const" version="opset1">
578
- <data element_type="i32" shape="6" offset="1288241" size="24" />
579
  <output>
580
  <port id="0" precision="I32">
581
  <dim>6</dim>
582
  </port>
583
  </output>
584
  </layer>
585
- <layer id="48" name="CombineSegments_43389" type="CombineSegments" version="extension">
586
  <input>
587
  <port id="0" precision="I32" />
588
  <port id="1" precision="I32" />
@@ -643,7 +632,7 @@
643
  </port>
644
  </output>
645
  </layer>
646
- <layer id="49" name="Subtract_43390" type="Subtract" version="opset1">
647
  <data auto_broadcast="numpy" />
648
  <input>
649
  <port id="0" precision="I32">
@@ -659,13 +648,13 @@
659
  </port>
660
  </output>
661
  </layer>
662
- <layer id="50" name="Constant_43391" type="Const" version="opset1">
663
  <data element_type="i32" shape="" offset="0" size="4" />
664
  <output>
665
  <port id="0" precision="I32" />
666
  </output>
667
  </layer>
668
- <layer id="51" name="ReduceMax_43392" type="ReduceMax" version="opset1">
669
  <data keep_dims="false" />
670
  <input>
671
  <port id="0" precision="I32">
@@ -677,14 +666,14 @@
677
  <port id="2" precision="I32" />
678
  </output>
679
  </layer>
680
- <layer id="52" name="Constant_43393" type="Const" version="opset1">
681
- <data element_type="i32" shape="" offset="1288237" size="4" />
682
  <output>
683
  <port id="0" precision="I32" />
684
  </output>
685
  </layer>
686
- <layer id="53" name="RaggedToDense_43394" type="RaggedToDense" version="extension">
687
- <data pad_right="true" />
688
  <input>
689
  <port id="0" precision="I32">
690
  <dim>-1</dim>
@@ -709,7 +698,7 @@
709
  </port>
710
  </output>
711
  </layer>
712
- <layer id="54" name="Convert_43395" type="Convert" version="opset1">
713
  <data destination_type="i32" />
714
  <input>
715
  <port id="0" precision="BOOL">
@@ -724,7 +713,7 @@
724
  </port>
725
  </output>
726
  </layer>
727
- <layer id="55" name="Convert_43395" type="Convert" version="opset1">
728
  <data destination_type="i64" />
729
  <input>
730
  <port id="0" precision="I32">
@@ -739,7 +728,7 @@
739
  </port>
740
  </output>
741
  </layer>
742
- <layer id="57" name="RaggedToDense_43394.0" type="Convert" version="opset1">
743
  <data destination_type="i64" />
744
  <input>
745
  <port id="0" precision="I32">
@@ -754,7 +743,7 @@
754
  </port>
755
  </output>
756
  </layer>
757
- <layer id="58" name="Result_43398" type="Result" version="opset1">
758
  <input>
759
  <port id="0" precision="I64">
760
  <dim>-1</dim>
@@ -762,7 +751,7 @@
762
  </port>
763
  </input>
764
  </layer>
765
- <layer id="56" name="Result_43400" type="Result" version="opset1">
766
  <input>
767
  <port id="0" precision="I64">
768
  <dim>-1</dim>
@@ -773,18 +762,18 @@
773
  </layers>
774
  <edges>
775
  <edge from-layer="0" from-port="0" to-layer="14" to-port="0" />
776
- <edge from-layer="1" from-port="0" to-layer="48" to-port="0" />
777
- <edge from-layer="2" from-port="0" to-layer="48" to-port="1" />
778
- <edge from-layer="3" from-port="0" to-layer="48" to-port="2" />
779
- <edge from-layer="4" from-port="0" to-layer="48" to-port="3" />
780
- <edge from-layer="5" from-port="0" to-layer="48" to-port="4" />
781
- <edge from-layer="6" from-port="0" to-layer="48" to-port="5" />
782
- <edge from-layer="7" from-port="0" to-layer="48" to-port="6" />
783
- <edge from-layer="8" from-port="0" to-layer="48" to-port="7" />
784
- <edge from-layer="9" from-port="0" to-layer="48" to-port="8" />
785
- <edge from-layer="10" from-port="0" to-layer="48" to-port="9" />
786
- <edge from-layer="11" from-port="0" to-layer="48" to-port="10" />
787
- <edge from-layer="12" from-port="0" to-layer="48" to-port="11" />
788
  <edge from-layer="13" from-port="0" to-layer="20" to-port="0" />
789
  <edge from-layer="14" from-port="1" to-layer="15" to-port="0" />
790
  <edge from-layer="14" from-port="3" to-layer="27" to-port="4" />
@@ -803,62 +792,58 @@
803
  <edge from-layer="24" from-port="0" to-layer="25" to-port="2" />
804
  <edge from-layer="25" from-port="3" to-layer="27" to-port="1" />
805
  <edge from-layer="26" from-port="0" to-layer="27" to-port="5" />
806
- <edge from-layer="27" from-port="7" to-layer="29" to-port="1" />
807
- <edge from-layer="27" from-port="6" to-layer="29" to-port="0" />
808
  <edge from-layer="27" from-port="11" to-layer="29" to-port="5" />
809
- <edge from-layer="27" from-port="10" to-layer="29" to-port="4" />
810
- <edge from-layer="27" from-port="9" to-layer="29" to-port="3" />
811
  <edge from-layer="27" from-port="8" to-layer="29" to-port="2" />
 
 
812
  <edge from-layer="28" from-port="0" to-layer="29" to-port="6" />
813
- <edge from-layer="29" from-port="11" to-layer="39" to-port="4" />
814
- <edge from-layer="29" from-port="10" to-layer="39" to-port="3" />
815
- <edge from-layer="29" from-port="9" to-layer="39" to-port="2" />
816
- <edge from-layer="29" from-port="8" to-layer="39" to-port="1" />
817
- <edge from-layer="29" from-port="7" to-layer="39" to-port="0" />
818
- <edge from-layer="30" from-port="0" to-layer="31" to-port="0" />
819
- <edge from-layer="31" from-port="3" to-layer="39" to-port="7" />
820
- <edge from-layer="31" from-port="2" to-layer="39" to-port="6" />
821
- <edge from-layer="31" from-port="1" to-layer="39" to-port="5" />
822
- <edge from-layer="32" from-port="0" to-layer="33" to-port="0" />
823
- <edge from-layer="33" from-port="2" to-layer="39" to-port="9" />
824
- <edge from-layer="33" from-port="3" to-layer="39" to-port="10" />
825
- <edge from-layer="33" from-port="1" to-layer="39" to-port="8" />
826
- <edge from-layer="34" from-port="0" to-layer="35" to-port="0" />
827
- <edge from-layer="35" from-port="3" to-layer="39" to-port="13" />
828
- <edge from-layer="35" from-port="1" to-layer="39" to-port="11" />
829
- <edge from-layer="35" from-port="2" to-layer="39" to-port="12" />
830
- <edge from-layer="36" from-port="0" to-layer="37" to-port="0" />
831
- <edge from-layer="37" from-port="3" to-layer="39" to-port="16" />
832
- <edge from-layer="37" from-port="2" to-layer="39" to-port="15" />
833
- <edge from-layer="37" from-port="1" to-layer="39" to-port="14" />
834
- <edge from-layer="38" from-port="0" to-layer="39" to-port="17" />
835
- <edge from-layer="39" from-port="18" to-layer="43" to-port="0" />
836
- <edge from-layer="39" from-port="20" to-layer="48" to-port="14" />
837
- <edge from-layer="39" from-port="18" to-layer="48" to-port="12" />
838
- <edge from-layer="39" from-port="19" to-layer="40" to-port="0" />
839
- <edge from-layer="39" from-port="18" to-layer="40" to-port="1" />
840
- <edge from-layer="40" from-port="2" to-layer="42" to-port="0" />
841
- <edge from-layer="41" from-port="0" to-layer="42" to-port="1" />
842
- <edge from-layer="42" from-port="2" to-layer="43" to-port="1" />
843
- <edge from-layer="43" from-port="2" to-layer="48" to-port="13" />
844
- <edge from-layer="44" from-port="0" to-layer="48" to-port="15" />
845
- <edge from-layer="45" from-port="0" to-layer="48" to-port="16" />
846
- <edge from-layer="46" from-port="0" to-layer="48" to-port="17" />
847
- <edge from-layer="47" from-port="0" to-layer="48" to-port="18" />
848
- <edge from-layer="48" from-port="20" to-layer="49" to-port="0" />
849
- <edge from-layer="48" from-port="19" to-layer="49" to-port="1" />
850
- <edge from-layer="48" from-port="19" to-layer="53" to-port="0" />
851
- <edge from-layer="48" from-port="20" to-layer="53" to-port="1" />
852
- <edge from-layer="48" from-port="21" to-layer="53" to-port="2" />
853
- <edge from-layer="49" from-port="2" to-layer="51" to-port="0" />
854
- <edge from-layer="50" from-port="0" to-layer="51" to-port="1" />
855
- <edge from-layer="51" from-port="2" to-layer="53" to-port="3" />
856
- <edge from-layer="52" from-port="0" to-layer="53" to-port="4" />
857
- <edge from-layer="53" from-port="6" to-layer="54" to-port="0" />
858
- <edge from-layer="53" from-port="5" to-layer="57" to-port="0" />
859
- <edge from-layer="54" from-port="1" to-layer="55" to-port="0" />
860
- <edge from-layer="55" from-port="1" to-layer="56" to-port="0" />
861
- <edge from-layer="57" from-port="1" to-layer="58" to-port="0" />
862
  </edges>
863
  <rt_info>
864
  <add_attention_mask value="True" />
@@ -869,21 +854,22 @@
869
  <detokenizer_input_type value="i64" />
870
  <eos_token_id value="50257" />
871
  <handle_special_tokens_with_re />
 
872
  <number_of_inputs value="1" />
873
- <openvino_tokenizers_version value="2024.5.0.0" />
874
- <openvino_version value="2024.5.0" />
 
875
  <original_tokenizer_class value="&lt;class 'transformers.models.whisper.tokenization_whisper_fast.WhisperTokenizerFast'>" />
876
  <pad_token_id value="50257" />
877
- <sentencepiece_version value="0.2.0" />
878
  <skip_special_tokens value="True" />
879
  <streaming_detokenizer value="False" />
880
- <tiktoken_version value="0.8.0" />
881
  <tokenizer_output_type value="i64" />
882
- <tokenizers_version value="0.20.3" />
883
- <transformers_version value="4.46.3" />
884
  <use_max_padding value="False" />
885
  <use_sentencepiece_backend value="False" />
886
- <utf8_replace_mode />
887
  <with_detokenizer value="True" />
888
  </rt_info>
889
  </net>
 
1
  <?xml version="1.0"?>
2
  <net name="tokenizer" version="11">
3
  <layers>
4
+ <layer id="0" name="Parameter_36186" type="Parameter" version="opset1">
5
  <data shape="?" element_type="string" />
6
  <output>
7
+ <port id="0" precision="STRING" names="Parameter_36186">
8
  <dim>-1</dim>
9
  </port>
10
  </output>
11
  </layer>
12
+ <layer id="1" name="Constant_36298" type="Const" version="opset1">
13
  <data element_type="i32" shape="" offset="0" size="4" />
14
  <output>
15
  <port id="0" precision="I32" />
16
  </output>
17
  </layer>
18
+ <layer id="2" name="Constant_36299" type="Const" version="opset1">
19
  <data element_type="i32" shape="" offset="4" size="4" />
20
  <output>
21
  <port id="0" precision="I32" />
22
  </output>
23
  </layer>
24
+ <layer id="3" name="Constant_36300" type="Const" version="opset1">
25
  <data element_type="i32" shape="1" offset="8" size="4" />
26
  <output>
27
  <port id="0" precision="I32">
 
29
  </port>
30
  </output>
31
  </layer>
32
+ <layer id="4" name="Constant_36301" type="Const" version="opset1">
33
  <data element_type="i32" shape="" offset="0" size="4" />
34
  <output>
35
  <port id="0" precision="I32" />
36
  </output>
37
  </layer>
38
+ <layer id="5" name="Constant_36302" type="Const" version="opset1">
39
  <data element_type="i32" shape="" offset="4" size="4" />
40
  <output>
41
  <port id="0" precision="I32" />
42
  </output>
43
  </layer>
44
+ <layer id="6" name="Constant_36303" type="Const" version="opset1">
45
  <data element_type="i32" shape="1" offset="12" size="4" />
46
  <output>
47
  <port id="0" precision="I32">
 
49
  </port>
50
  </output>
51
  </layer>
52
+ <layer id="7" name="Constant_36304" type="Const" version="opset1">
53
  <data element_type="i32" shape="" offset="0" size="4" />
54
  <output>
55
  <port id="0" precision="I32" />
56
  </output>
57
  </layer>
58
+ <layer id="8" name="Constant_36305" type="Const" version="opset1">
59
  <data element_type="i32" shape="" offset="4" size="4" />
60
  <output>
61
  <port id="0" precision="I32" />
62
  </output>
63
  </layer>
64
+ <layer id="9" name="Constant_36306" type="Const" version="opset1">
65
  <data element_type="i32" shape="1" offset="16" size="4" />
66
  <output>
67
  <port id="0" precision="I32">
 
69
  </port>
70
  </output>
71
  </layer>
72
+ <layer id="10" name="Constant_36307" type="Const" version="opset1">
73
  <data element_type="i32" shape="" offset="0" size="4" />
74
  <output>
75
  <port id="0" precision="I32" />
76
  </output>
77
  </layer>
78
+ <layer id="11" name="Constant_36308" type="Const" version="opset1">
79
  <data element_type="i32" shape="" offset="4" size="4" />
80
  <output>
81
  <port id="0" precision="I32" />
82
  </output>
83
  </layer>
84
+ <layer id="12" name="Constant_36309" type="Const" version="opset1">
85
  <data element_type="i32" shape="1" offset="20" size="4" />
86
  <output>
87
  <port id="0" precision="I32">
 
89
  </port>
90
  </output>
91
  </layer>
92
+ <layer id="13" name="Constant_36192" type="Const" version="opset1">
93
  <data element_type="i64" shape="" offset="24" size="8" />
94
  <output>
95
  <port id="0" precision="I64" />
96
  </output>
97
  </layer>
98
+ <layer id="14" name="StringTensorUnpack_36187" type="StringTensorUnpack" version="opset15">
 
99
  <input>
100
  <port id="0" precision="STRING">
101
  <dim>-1</dim>
 
113
  </port>
114
  </output>
115
  </layer>
116
+ <layer id="15" name="ShapeOf_36188" type="ShapeOf" version="opset3">
117
  <data output_type="i64" />
118
  <input>
119
  <port id="0" precision="I32">
 
126
  </port>
127
  </output>
128
  </layer>
129
+ <layer id="16" name="Constant_36189" type="Const" version="opset1">
130
  <data element_type="i64" shape="" offset="24" size="8" />
131
  <output>
132
  <port id="0" precision="I64" />
133
  </output>
134
  </layer>
135
+ <layer id="17" name="Constant_36190" type="Const" version="opset1">
136
  <data element_type="i64" shape="" offset="24" size="8" />
137
  <output>
138
  <port id="0" precision="I64" />
139
  </output>
140
  </layer>
141
+ <layer id="18" name="Gather_36191" type="Gather" version="opset8">
142
  <data batch_dims="0" />
143
  <input>
144
  <port id="0" precision="I64">
 
151
  <port id="3" precision="I64" />
152
  </output>
153
  </layer>
154
+ <layer id="19" name="Constant_36193" type="Const" version="opset1">
155
  <data element_type="i64" shape="" offset="32" size="8" />
156
  <output>
157
  <port id="0" precision="I64" />
158
  </output>
159
  </layer>
160
+ <layer id="20" name="Range_36194" type="Range" version="opset4">
161
  <data output_type="i32" />
162
  <input>
163
  <port id="0" precision="I64" />
 
170
  </port>
171
  </output>
172
  </layer>
173
+ <layer id="21" name="Constant_36195" type="Const" version="opset1">
174
  <data element_type="i64" shape="" offset="32" size="8" />
175
  <output>
176
  <port id="0" precision="I64" />
177
  </output>
178
  </layer>
179
+ <layer id="22" name="Constant_36196" type="Const" version="opset1">
180
  <data element_type="i64" shape="" offset="32" size="8" />
181
  <output>
182
  <port id="0" precision="I64" />
183
  </output>
184
  </layer>
185
+ <layer id="23" name="Add_36197" type="Add" version="opset1">
186
  <data auto_broadcast="numpy" />
187
  <input>
188
  <port id="0" precision="I64" />
 
192
  <port id="2" precision="I64" />
193
  </output>
194
  </layer>
195
+ <layer id="24" name="Constant_36198" type="Const" version="opset1">
196
  <data element_type="i64" shape="" offset="32" size="8" />
197
  <output>
198
  <port id="0" precision="I64" />
199
  </output>
200
  </layer>
201
+ <layer id="25" name="Range_36199" type="Range" version="opset4">
202
  <data output_type="i32" />
203
  <input>
204
  <port id="0" precision="I64" />
 
211
  </port>
212
  </output>
213
  </layer>
214
+ <layer id="26" name="Constant_36261" type="Const" version="opset1">
215
  <data element_type="u8" shape="26479" offset="40" size="26479" />
216
  <output>
217
  <port id="0" precision="U8">
 
219
  </port>
220
  </output>
221
  </layer>
222
+ <layer id="27" name="SpecialTokensSplit_36262" type="SpecialTokensSplit" version="extension">
223
  <input>
224
  <port id="0" precision="I32">
225
  <dim>-1</dim>
 
261
  </port>
262
  </output>
263
  </layer>
264
+ <layer id="28" name="Constant_36264" type="Const" version="opset1">
265
+ <data element_type="u8" shape="74" offset="26519" size="74" />
266
  <output>
267
  <port id="0" precision="U8">
268
+ <dim>74</dim>
269
  </port>
270
  </output>
271
  </layer>
272
+ <layer id="29" name="RegexSplit_36265" type="RegexSplit" version="extension">
273
  <data behaviour="isolate" invert="false" max_splits="-1" />
274
  <input>
275
  <port id="0" precision="I32">
 
291
  <dim>-1</dim>
292
  </port>
293
  <port id="6" precision="U8">
294
+ <dim>74</dim>
295
  </port>
296
  </input>
297
  <output>
 
315
  </port>
316
  </output>
317
  </layer>
318
+ <layer id="30" name="Constant_36267" type="Const" version="opset1">
319
+ <data element_type="i32" shape="51865" offset="26593" size="207460" />
320
  <output>
321
+ <port id="0" precision="I32">
322
+ <dim>51865</dim>
323
  </port>
324
  </output>
325
  </layer>
326
+ <layer id="31" name="Constant_36269" type="Const" version="opset1">
327
+ <data element_type="i32" shape="51865" offset="234053" size="207460" />
 
 
 
 
 
328
  <output>
329
+ <port id="0" precision="I32">
330
+ <dim>51865</dim>
 
 
 
 
 
 
331
  </port>
332
  </output>
333
  </layer>
334
+ <layer id="32" name="Constant_36271" type="Const" version="opset1">
335
+ <data element_type="u8" shape="320780" offset="441513" size="320780" />
336
  <output>
337
  <port id="0" precision="U8">
338
+ <dim>320780</dim>
339
  </port>
340
  </output>
341
  </layer>
342
+ <layer id="33" name="Constant_36279" type="Const" version="opset1">
343
+ <data element_type="i32" shape="49999" offset="762293" size="199996" />
 
 
 
 
 
344
  <output>
345
+ <port id="0" precision="I32">
346
+ <dim>49999</dim>
 
 
 
 
 
 
347
  </port>
348
  </output>
349
  </layer>
350
+ <layer id="34" name="Constant_36281" type="Const" version="opset1">
351
+ <data element_type="i32" shape="49999" offset="962289" size="199996" />
352
  <output>
353
+ <port id="0" precision="I32">
354
+ <dim>49999</dim>
355
  </port>
356
  </output>
357
  </layer>
358
+ <layer id="35" name="Constant_36283" type="Const" version="opset1">
359
+ <data element_type="u8" shape="169949" offset="1162285" size="169949" />
360
+ <output>
361
  <port id="0" precision="U8">
362
+ <dim>169949</dim>
363
  </port>
364
+ </output>
365
+ </layer>
366
+ <layer id="36" name="Constant_36285" type="Const" version="opset1">
367
+ <data element_type="i32" shape="49999" offset="1332234" size="199996" />
368
  <output>
369
+ <port id="0" precision="I32">
370
+ <dim>49999</dim>
 
 
 
 
 
 
371
  </port>
372
  </output>
373
  </layer>
374
+ <layer id="37" name="Constant_36287" type="Const" version="opset1">
375
+ <data element_type="i32" shape="49999" offset="1532230" size="199996" />
376
  <output>
377
+ <port id="0" precision="I32">
378
+ <dim>49999</dim>
379
  </port>
380
  </output>
381
  </layer>
382
+ <layer id="38" name="Constant_36289" type="Const" version="opset1">
383
+ <data element_type="u8" shape="136850" offset="1732226" size="136850" />
384
+ <output>
385
  <port id="0" precision="U8">
386
+ <dim>136850</dim>
387
  </port>
388
+ </output>
389
+ </layer>
390
+ <layer id="39" name="Constant_36273" type="Const" version="opset1">
391
+ <data element_type="i32" shape="1608" offset="1869076" size="6432" />
392
  <output>
393
+ <port id="0" precision="I32">
394
+ <dim>1608</dim>
395
  </port>
396
+ </output>
397
+ </layer>
398
+ <layer id="40" name="Constant_36275" type="Const" version="opset1">
399
+ <data element_type="i32" shape="1608" offset="1875508" size="6432" />
400
+ <output>
401
+ <port id="0" precision="I32">
402
+ <dim>1608</dim>
403
  </port>
404
+ </output>
405
+ </layer>
406
+ <layer id="41" name="Constant_36277" type="Const" version="opset1">
407
+ <data element_type="u8" shape="13723" offset="1881940" size="13723" />
408
+ <output>
409
+ <port id="0" precision="U8">
410
+ <dim>13723</dim>
411
  </port>
412
  </output>
413
  </layer>
414
+ <layer id="42" name="Constant_36290" type="Const" version="opset1">
415
+ <data element_type="i32" shape="1608" offset="1895663" size="6432" />
416
  <output>
417
  <port id="0" precision="I32">
418
  <dim>1608</dim>
419
  </port>
420
  </output>
421
  </layer>
422
+ <layer id="43" name="BPETokenizer_36291" type="BPETokenizer" version="extension">
423
  <data unk_token="" fuse_unk="false" suffix_indicator="" end_suffix="" byte_fallback="false" cache_capacity="20000" />
424
  <input>
425
  <port id="0" precision="I32">
 
438
  <dim>-1</dim>
439
  </port>
440
  <port id="5" precision="I32">
441
+ <dim>51865</dim>
442
  </port>
443
  <port id="6" precision="I32">
444
+ <dim>51865</dim>
445
  </port>
446
  <port id="7" precision="U8">
447
+ <dim>320780</dim>
448
  </port>
449
  <port id="8" precision="I32">
450
+ <dim>49999</dim>
451
  </port>
452
  <port id="9" precision="I32">
453
+ <dim>49999</dim>
454
  </port>
455
  <port id="10" precision="U8">
456
+ <dim>169949</dim>
457
  </port>
458
  <port id="11" precision="I32">
459
+ <dim>49999</dim>
460
  </port>
461
  <port id="12" precision="I32">
462
+ <dim>49999</dim>
463
  </port>
464
  <port id="13" precision="U8">
465
+ <dim>136850</dim>
466
  </port>
467
  <port id="14" precision="I32">
468
+ <dim>1608</dim>
469
  </port>
470
  <port id="15" precision="I32">
471
+ <dim>1608</dim>
472
  </port>
473
  <port id="16" precision="U8">
474
+ <dim>13723</dim>
475
  </port>
476
  <port id="17" precision="I32">
477
  <dim>1608</dim>
 
489
  </port>
490
  </output>
491
  </layer>
492
+ <layer id="44" name="Constant_36292" type="Const" version="opset1">
493
+ <data element_type="i32" shape="" offset="1902095" size="4" />
 
 
 
 
 
 
 
 
494
  <output>
495
+ <port id="0" precision="I32" />
 
 
496
  </output>
497
  </layer>
498
+ <layer id="45" name="Constant_36294" type="Const" version="opset1">
499
+ <data element_type="u8" shape="5" offset="1902099" size="5" />
500
  <output>
501
+ <port id="0" precision="U8">
502
+ <dim>5</dim>
503
+ </port>
504
  </output>
505
  </layer>
506
+ <layer id="46" name="Constant_36296" type="Const" version="opset1">
507
+ <data element_type="u8" shape="13" offset="1902104" size="13" />
 
 
 
 
 
 
508
  <output>
509
+ <port id="0" precision="U8">
510
+ <dim>13</dim>
511
  </port>
512
  </output>
513
  </layer>
514
+ <layer id="47" name="Truncate_36297" type="Truncate" version="extension">
515
+ <data m_num_inputs="1" />
516
  <input>
517
  <port id="0" precision="I32">
518
  <dim>-1</dim>
 
520
  <port id="1" precision="I32">
521
  <dim>-1</dim>
522
  </port>
523
+ <port id="2" precision="I32">
524
+ <dim>-1</dim>
525
+ </port>
526
+ <port id="3" precision="I32" />
527
+ <port id="4" precision="U8">
528
+ <dim>5</dim>
529
+ </port>
530
+ <port id="5" precision="U8">
531
+ <dim>13</dim>
532
+ </port>
533
  </input>
534
  <output>
535
+ <port id="6" precision="I32">
536
+ <dim>-1</dim>
537
+ </port>
538
+ <port id="7" precision="I32">
539
+ <dim>-1</dim>
540
+ </port>
541
+ <port id="8" precision="I32">
542
  <dim>-1</dim>
543
  </port>
544
  </output>
545
  </layer>
546
+ <layer id="48" name="Constant_36310" type="Const" version="opset1">
547
  <data element_type="i32" shape="" offset="0" size="4" />
548
  <output>
549
  <port id="0" precision="I32" />
550
  </output>
551
  </layer>
552
+ <layer id="49" name="Constant_36311" type="Const" version="opset1">
553
  <data element_type="i32" shape="" offset="4" size="4" />
554
  <output>
555
  <port id="0" precision="I32" />
556
  </output>
557
  </layer>
558
+ <layer id="50" name="Constant_36312" type="Const" version="opset1">
559
+ <data element_type="i32" shape="1" offset="1902117" size="4" />
560
  <output>
561
  <port id="0" precision="I32">
562
  <dim>1</dim>
563
  </port>
564
  </output>
565
  </layer>
566
+ <layer id="51" name="Constant_36313" type="Const" version="opset1">
567
+ <data element_type="i32" shape="6" offset="1902121" size="24" />
568
  <output>
569
  <port id="0" precision="I32">
570
  <dim>6</dim>
571
  </port>
572
  </output>
573
  </layer>
574
+ <layer id="52" name="CombineSegments_36314" type="CombineSegments" version="extension">
575
  <input>
576
  <port id="0" precision="I32" />
577
  <port id="1" precision="I32" />
 
632
  </port>
633
  </output>
634
  </layer>
635
+ <layer id="53" name="Subtract_36315" type="Subtract" version="opset1">
636
  <data auto_broadcast="numpy" />
637
  <input>
638
  <port id="0" precision="I32">
 
648
  </port>
649
  </output>
650
  </layer>
651
+ <layer id="54" name="Constant_36316" type="Const" version="opset1">
652
  <data element_type="i32" shape="" offset="0" size="4" />
653
  <output>
654
  <port id="0" precision="I32" />
655
  </output>
656
  </layer>
657
+ <layer id="55" name="ReduceMax_36317" type="ReduceMax" version="opset1">
658
  <data keep_dims="false" />
659
  <input>
660
  <port id="0" precision="I32">
 
666
  <port id="2" precision="I32" />
667
  </output>
668
  </layer>
669
+ <layer id="56" name="Constant_36318" type="Const" version="opset1">
670
+ <data element_type="i32" shape="" offset="1902117" size="4" />
671
  <output>
672
  <port id="0" precision="I32" />
673
  </output>
674
  </layer>
675
+ <layer id="57" name="RaggedToDense_36319" type="RaggedToDense" version="extension">
676
+ <data pad_right="true" m_pad_max_length="false" />
677
  <input>
678
  <port id="0" precision="I32">
679
  <dim>-1</dim>
 
698
  </port>
699
  </output>
700
  </layer>
701
+ <layer id="58" name="Convert_36320" type="Convert" version="opset1">
702
  <data destination_type="i32" />
703
  <input>
704
  <port id="0" precision="BOOL">
 
713
  </port>
714
  </output>
715
  </layer>
716
+ <layer id="59" name="Convert_36320.0" type="Convert" version="opset1">
717
  <data destination_type="i64" />
718
  <input>
719
  <port id="0" precision="I32">
 
728
  </port>
729
  </output>
730
  </layer>
731
+ <layer id="61" name="RaggedToDense_36319.0" type="Convert" version="opset1">
732
  <data destination_type="i64" />
733
  <input>
734
  <port id="0" precision="I32">
 
743
  </port>
744
  </output>
745
  </layer>
746
+ <layer id="62" name="Result_36323" type="Result" version="opset1" output_names="input_ids">
747
  <input>
748
  <port id="0" precision="I64">
749
  <dim>-1</dim>
 
751
  </port>
752
  </input>
753
  </layer>
754
+ <layer id="60" name="Result_36325" type="Result" version="opset1" output_names="attention_mask">
755
  <input>
756
  <port id="0" precision="I64">
757
  <dim>-1</dim>
 
762
  </layers>
763
  <edges>
764
  <edge from-layer="0" from-port="0" to-layer="14" to-port="0" />
765
+ <edge from-layer="1" from-port="0" to-layer="52" to-port="0" />
766
+ <edge from-layer="2" from-port="0" to-layer="52" to-port="1" />
767
+ <edge from-layer="3" from-port="0" to-layer="52" to-port="2" />
768
+ <edge from-layer="4" from-port="0" to-layer="52" to-port="3" />
769
+ <edge from-layer="5" from-port="0" to-layer="52" to-port="4" />
770
+ <edge from-layer="6" from-port="0" to-layer="52" to-port="5" />
771
+ <edge from-layer="7" from-port="0" to-layer="52" to-port="6" />
772
+ <edge from-layer="8" from-port="0" to-layer="52" to-port="7" />
773
+ <edge from-layer="9" from-port="0" to-layer="52" to-port="8" />
774
+ <edge from-layer="10" from-port="0" to-layer="52" to-port="9" />
775
+ <edge from-layer="11" from-port="0" to-layer="52" to-port="10" />
776
+ <edge from-layer="12" from-port="0" to-layer="52" to-port="11" />
777
  <edge from-layer="13" from-port="0" to-layer="20" to-port="0" />
778
  <edge from-layer="14" from-port="1" to-layer="15" to-port="0" />
779
  <edge from-layer="14" from-port="3" to-layer="27" to-port="4" />
 
792
  <edge from-layer="24" from-port="0" to-layer="25" to-port="2" />
793
  <edge from-layer="25" from-port="3" to-layer="27" to-port="1" />
794
  <edge from-layer="26" from-port="0" to-layer="27" to-port="5" />
 
 
795
  <edge from-layer="27" from-port="11" to-layer="29" to-port="5" />
796
+ <edge from-layer="27" from-port="6" to-layer="29" to-port="0" />
797
+ <edge from-layer="27" from-port="7" to-layer="29" to-port="1" />
798
  <edge from-layer="27" from-port="8" to-layer="29" to-port="2" />
799
+ <edge from-layer="27" from-port="9" to-layer="29" to-port="3" />
800
+ <edge from-layer="27" from-port="10" to-layer="29" to-port="4" />
801
  <edge from-layer="28" from-port="0" to-layer="29" to-port="6" />
802
+ <edge from-layer="29" from-port="11" to-layer="43" to-port="4" />
803
+ <edge from-layer="29" from-port="10" to-layer="43" to-port="3" />
804
+ <edge from-layer="29" from-port="9" to-layer="43" to-port="2" />
805
+ <edge from-layer="29" from-port="8" to-layer="43" to-port="1" />
806
+ <edge from-layer="29" from-port="7" to-layer="43" to-port="0" />
807
+ <edge from-layer="30" from-port="0" to-layer="43" to-port="5" />
808
+ <edge from-layer="31" from-port="0" to-layer="43" to-port="6" />
809
+ <edge from-layer="32" from-port="0" to-layer="43" to-port="7" />
810
+ <edge from-layer="33" from-port="0" to-layer="43" to-port="8" />
811
+ <edge from-layer="34" from-port="0" to-layer="43" to-port="9" />
812
+ <edge from-layer="35" from-port="0" to-layer="43" to-port="10" />
813
+ <edge from-layer="36" from-port="0" to-layer="43" to-port="11" />
814
+ <edge from-layer="37" from-port="0" to-layer="43" to-port="12" />
815
+ <edge from-layer="38" from-port="0" to-layer="43" to-port="13" />
816
+ <edge from-layer="39" from-port="0" to-layer="43" to-port="14" />
817
+ <edge from-layer="40" from-port="0" to-layer="43" to-port="15" />
818
+ <edge from-layer="41" from-port="0" to-layer="43" to-port="16" />
819
+ <edge from-layer="42" from-port="0" to-layer="43" to-port="17" />
820
+ <edge from-layer="43" from-port="20" to-layer="47" to-port="2" />
821
+ <edge from-layer="43" from-port="19" to-layer="47" to-port="1" />
822
+ <edge from-layer="43" from-port="18" to-layer="47" to-port="0" />
823
+ <edge from-layer="44" from-port="0" to-layer="47" to-port="3" />
824
+ <edge from-layer="45" from-port="0" to-layer="47" to-port="4" />
825
+ <edge from-layer="46" from-port="0" to-layer="47" to-port="5" />
826
+ <edge from-layer="47" from-port="6" to-layer="52" to-port="12" />
827
+ <edge from-layer="47" from-port="7" to-layer="52" to-port="13" />
828
+ <edge from-layer="47" from-port="8" to-layer="52" to-port="14" />
829
+ <edge from-layer="48" from-port="0" to-layer="52" to-port="15" />
830
+ <edge from-layer="49" from-port="0" to-layer="52" to-port="16" />
831
+ <edge from-layer="50" from-port="0" to-layer="52" to-port="17" />
832
+ <edge from-layer="51" from-port="0" to-layer="52" to-port="18" />
833
+ <edge from-layer="52" from-port="20" to-layer="53" to-port="0" />
834
+ <edge from-layer="52" from-port="19" to-layer="53" to-port="1" />
835
+ <edge from-layer="52" from-port="19" to-layer="57" to-port="0" />
836
+ <edge from-layer="52" from-port="20" to-layer="57" to-port="1" />
837
+ <edge from-layer="52" from-port="21" to-layer="57" to-port="2" />
838
+ <edge from-layer="53" from-port="2" to-layer="55" to-port="0" />
839
+ <edge from-layer="54" from-port="0" to-layer="55" to-port="1" />
840
+ <edge from-layer="55" from-port="2" to-layer="57" to-port="3" />
841
+ <edge from-layer="56" from-port="0" to-layer="57" to-port="4" />
842
+ <edge from-layer="57" from-port="6" to-layer="58" to-port="0" />
843
+ <edge from-layer="57" from-port="5" to-layer="61" to-port="0" />
844
+ <edge from-layer="58" from-port="1" to-layer="59" to-port="0" />
845
+ <edge from-layer="59" from-port="1" to-layer="60" to-port="0" />
846
+ <edge from-layer="61" from-port="1" to-layer="62" to-port="0" />
 
 
 
 
847
  </edges>
848
  <rt_info>
849
  <add_attention_mask value="True" />
 
854
  <detokenizer_input_type value="i64" />
855
  <eos_token_id value="50257" />
856
  <handle_special_tokens_with_re />
857
+ <max_length />
858
  <number_of_inputs value="1" />
859
+ <openvino_tokenizers_version value="2025.2.0.0-565-130827ab189" />
860
+ <openvino_version value="2025.2.0-19120-87425bc78ca-releases/2025/2" />
861
+ <original_post_processor_template value="{&quot;type&quot;: &quot;TemplateProcessing&quot;, &quot;single&quot;: [{&quot;SpecialToken&quot;: {&quot;id&quot;: &quot;&lt;|startoftranscript|>&quot;, &quot;type_id&quot;: 0}}, {&quot;SpecialToken&quot;: {&quot;id&quot;: &quot;&lt;|en|>&quot;, &quot;type_id&quot;: 0}}, {&quot;SpecialToken&quot;: {&quot;id&quot;: &quot;&lt;|transcribe|>&quot;, &quot;type_id&quot;: 0}}, {&quot;SpecialToken&quot;: {&quot;id&quot;: &quot;&lt;|notimestamps|>&quot;, &quot;type_id&quot;: 0}}, {&quot;Sequence&quot;: {&quot;id&quot;: &quot;A&quot;, &quot;type_id&quot;: 0}}, {&quot;SpecialToken&quot;: {&quot;id&quot;: &quot;&lt;|endoftext|>&quot;, &quot;type_id&quot;: 0}}], &quot;pair&quot;: [{&quot;SpecialToken&quot;: {&quot;id&quot;: &quot;&lt;|startoftranscript|>&quot;, &quot;type_id&quot;: 0}}, {&quot;SpecialToken&quot;: {&quot;id&quot;: &quot;&lt;|en|>&quot;, &quot;type_id&quot;: 0}}, {&quot;SpecialToken&quot;: {&quot;id&quot;: &quot;&lt;|transcribe|>&quot;, &quot;type_id&quot;: 0}}, {&quot;SpecialToken&quot;: {&quot;id&quot;: &quot;&lt;|notimestamps|>&quot;, &quot;type_id&quot;: 0}}, {&quot;Sequence&quot;: {&quot;id&quot;: &quot;A&quot;, &quot;type_id&quot;: 0}}, {&quot;Sequence&quot;: {&quot;id&quot;: &quot;B&quot;, &quot;type_id&quot;: 1}}, {&quot;SpecialToken&quot;: {&quot;id&quot;: &quot;&lt;|endoftext|>&quot;, &quot;type_id&quot;: 1}}], &quot;special_tokens&quot;: {&quot;&lt;|endoftext|>&quot;: {&quot;id&quot;: &quot;&lt;|endoftext|>&quot;, &quot;ids&quot;: [50257], &quot;tokens&quot;: [&quot;&lt;|endoftext|>&quot;]}, &quot;&lt;|en|>&quot;: {&quot;id&quot;: &quot;&lt;|en|>&quot;, &quot;ids&quot;: [50259], &quot;tokens&quot;: [&quot;&lt;|en|>&quot;]}, &quot;&lt;|notimestamps|>&quot;: {&quot;id&quot;: &quot;&lt;|notimestamps|>&quot;, &quot;ids&quot;: [50363], &quot;tokens&quot;: [&quot;&lt;|notimestamps|>&quot;]}, &quot;&lt;|startoftranscript|>&quot;: {&quot;id&quot;: &quot;&lt;|startoftranscript|>&quot;, &quot;ids&quot;: [50258], &quot;tokens&quot;: [&quot;&lt;|startoftranscript|>&quot;]}, &quot;&lt;|transcribe|>&quot;: {&quot;id&quot;: &quot;&lt;|transcribe|>&quot;, &quot;ids&quot;: [50359], &quot;tokens&quot;: [&quot;&lt;|transcribe|>&quot;]}}}" />
862
  <original_tokenizer_class value="&lt;class 'transformers.models.whisper.tokenization_whisper_fast.WhisperTokenizerFast'>" />
863
  <pad_token_id value="50257" />
864
+ <processed_post_processor_template value="{&quot;single&quot;: {&quot;ids&quot;: [50258, 50259, 50359, 50363, -1, 50257], &quot;type_ids&quot;: [0, 0, 0, 0, 0, 0]}, &quot;pair&quot;: {&quot;ids&quot;: [50258, 50259, 50359, 50363, -1, -2, 50257], &quot;type_ids&quot;: [0, 0, 0, 0, 0, 1, 1]}}" />
865
  <skip_special_tokens value="True" />
866
  <streaming_detokenizer value="False" />
 
867
  <tokenizer_output_type value="i64" />
868
+ <tokenizers_version value="0.21.1" />
869
+ <transformers_version value="4.51.3" />
870
  <use_max_padding value="False" />
871
  <use_sentencepiece_backend value="False" />
872
+ <utf8_replace_mode value="replace" />
873
  <with_detokenizer value="True" />
874
  </rt_info>
875
  </net>
preprocessor_config.json CHANGED
@@ -1,5 +1,6 @@
1
  {
2
  "chunk_length": 30,
 
3
  "feature_extractor_type": "WhisperFeatureExtractor",
4
  "feature_size": 80,
5
  "hop_length": 160,
 
1
  {
2
  "chunk_length": 30,
3
+ "dither": 0.0,
4
  "feature_extractor_type": "WhisperFeatureExtractor",
5
  "feature_size": 80,
6
  "hop_length": 160,
tokenizer_config.json CHANGED
@@ -12980,6 +12980,7 @@
12980
  "clean_up_tokenization_spaces": true,
12981
  "eos_token": "<|endoftext|>",
12982
  "errors": "replace",
 
12983
  "model_max_length": 1024,
12984
  "pad_token": "<|endoftext|>",
12985
  "processor_class": "WhisperProcessor",
 
12980
  "clean_up_tokenization_spaces": true,
12981
  "eos_token": "<|endoftext|>",
12982
  "errors": "replace",
12983
+ "extra_special_tokens": {},
12984
  "model_max_length": 1024,
12985
  "pad_token": "<|endoftext|>",
12986
  "processor_class": "WhisperProcessor",