Upload folder using huggingface_hub
Browse files- .ipynb_checkpoints/README-checkpoint.md +217 -0
- README.md +119 -12
- config.json +2 -2
- generation_config.json +1 -1
- openvino_decoder_model.bin +2 -2
- openvino_decoder_model.xml +0 -0
- openvino_detokenizer.bin +2 -2
- openvino_detokenizer.xml +102 -66
- openvino_encoder_model.bin +1 -1
- openvino_encoder_model.xml +0 -0
- openvino_tokenizer.bin +2 -2
- openvino_tokenizer.xml +211 -225
- preprocessor_config.json +1 -0
- tokenizer_config.json +1 -0
.ipynb_checkpoints/README-checkpoint.md
ADDED
|
@@ -0,0 +1,217 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
language:
|
| 3 |
+
- en
|
| 4 |
+
- zh
|
| 5 |
+
- de
|
| 6 |
+
- es
|
| 7 |
+
- ru
|
| 8 |
+
- ko
|
| 9 |
+
- fr
|
| 10 |
+
- ja
|
| 11 |
+
- pt
|
| 12 |
+
- tr
|
| 13 |
+
- pl
|
| 14 |
+
- ca
|
| 15 |
+
- nl
|
| 16 |
+
- ar
|
| 17 |
+
- sv
|
| 18 |
+
- it
|
| 19 |
+
- id
|
| 20 |
+
- hi
|
| 21 |
+
- fi
|
| 22 |
+
- vi
|
| 23 |
+
- he
|
| 24 |
+
- uk
|
| 25 |
+
- el
|
| 26 |
+
- ms
|
| 27 |
+
- cs
|
| 28 |
+
- ro
|
| 29 |
+
- da
|
| 30 |
+
- hu
|
| 31 |
+
- ta
|
| 32 |
+
- no
|
| 33 |
+
- th
|
| 34 |
+
- ur
|
| 35 |
+
- hr
|
| 36 |
+
- bg
|
| 37 |
+
- lt
|
| 38 |
+
- la
|
| 39 |
+
- mi
|
| 40 |
+
- ml
|
| 41 |
+
- cy
|
| 42 |
+
- sk
|
| 43 |
+
- te
|
| 44 |
+
- fa
|
| 45 |
+
- lv
|
| 46 |
+
- bn
|
| 47 |
+
- sr
|
| 48 |
+
- az
|
| 49 |
+
- sl
|
| 50 |
+
- kn
|
| 51 |
+
- et
|
| 52 |
+
- mk
|
| 53 |
+
- br
|
| 54 |
+
- eu
|
| 55 |
+
- is
|
| 56 |
+
- hy
|
| 57 |
+
- ne
|
| 58 |
+
- mn
|
| 59 |
+
- bs
|
| 60 |
+
- kk
|
| 61 |
+
- sq
|
| 62 |
+
- sw
|
| 63 |
+
- gl
|
| 64 |
+
- mr
|
| 65 |
+
- pa
|
| 66 |
+
- si
|
| 67 |
+
- km
|
| 68 |
+
- sn
|
| 69 |
+
- yo
|
| 70 |
+
- so
|
| 71 |
+
- af
|
| 72 |
+
- oc
|
| 73 |
+
- ka
|
| 74 |
+
- be
|
| 75 |
+
- tg
|
| 76 |
+
- sd
|
| 77 |
+
- gu
|
| 78 |
+
- am
|
| 79 |
+
- yi
|
| 80 |
+
- lo
|
| 81 |
+
- uz
|
| 82 |
+
- fo
|
| 83 |
+
- ht
|
| 84 |
+
- ps
|
| 85 |
+
- tk
|
| 86 |
+
- nn
|
| 87 |
+
- mt
|
| 88 |
+
- sa
|
| 89 |
+
- lb
|
| 90 |
+
- my
|
| 91 |
+
- bo
|
| 92 |
+
- tl
|
| 93 |
+
- mg
|
| 94 |
+
- as
|
| 95 |
+
- tt
|
| 96 |
+
- haw
|
| 97 |
+
- ln
|
| 98 |
+
- ha
|
| 99 |
+
- ba
|
| 100 |
+
- jw
|
| 101 |
+
- su
|
| 102 |
+
tags:
|
| 103 |
+
- audio
|
| 104 |
+
- automatic-speech-recognition
|
| 105 |
+
- hf-asr-leaderboard
|
| 106 |
+
pipeline_tag: automatic-speech-recognition
|
| 107 |
+
license: apache-2.0
|
| 108 |
+
license_link: https://choosealicense.com/licenses/apache-2.0/
|
| 109 |
+
---
|
| 110 |
+
|
| 111 |
+
# whisper-base-int4-ov
|
| 112 |
+
* Model creator: [OpenAI](https://huggingface.co/openai)
|
| 113 |
+
* Original model: [whisper-base](https://huggingface.co/openai/whisper-base)
|
| 114 |
+
|
| 115 |
+
## Description
|
| 116 |
+
This is [whisper-base](https://huggingface.co/openai/whisper-base) model converted to the [OpenVINO™ IR](https://docs.openvino.ai/2025/documentation/openvino-ir-format.html) (Intermediate Representation) format with weights compressed to INT4 by [NNCF](https://github.com/openvinotoolkit/nncf).
|
| 117 |
+
|
| 118 |
+
|
| 119 |
+
## Quantization Parameters
|
| 120 |
+
|
| 121 |
+
Weight compression was performed using `nncf.compress_weights` with the following parameters:
|
| 122 |
+
|
| 123 |
+
* mode: **INT4_ASYM**
|
| 124 |
+
* ratio: **1.0**
|
| 125 |
+
* group_size: **128**
|
| 126 |
+
|
| 127 |
+
For more information on quantization, check the [OpenVINO model optimization guide](https://docs.openvino.ai/2025/openvino-workflow/model-optimization-guide/weight-compression.html).
|
| 128 |
+
|
| 129 |
+
|
| 130 |
+
## Compatibility
|
| 131 |
+
|
| 132 |
+
The provided OpenVINO™ IR model is compatible with:
|
| 133 |
+
|
| 134 |
+
* OpenVINO version 2025.1.0 and higher
|
| 135 |
+
* Optimum Intel 1.23.0 and higher
|
| 136 |
+
|
| 137 |
+
|
| 138 |
+
## Running Model Inference with [Optimum Intel](https://huggingface.co/docs/optimum/intel/index)
|
| 139 |
+
|
| 140 |
+
1. Install packages required for using [Optimum Intel](https://huggingface.co/docs/optimum/intel/index) integration with the OpenVINO backend:
|
| 141 |
+
|
| 142 |
+
```
|
| 143 |
+
pip install optimum[openvino]
|
| 144 |
+
```
|
| 145 |
+
|
| 146 |
+
2. Run model inference:
|
| 147 |
+
|
| 148 |
+
```
|
| 149 |
+
from datasets import load_dataset
|
| 150 |
+
from transformers import AutoProcessor
|
| 151 |
+
from optimum.intel.openvino import OVModelForSpeechSeq2Seq
|
| 152 |
+
|
| 153 |
+
model_id = "OpenVINO/whisper-base-int4-ov"
|
| 154 |
+
tokenizer = AutoProcessor.from_pretrained(model_id)
|
| 155 |
+
model = OVModelForSpeechSeq2Seq.from_pretrained(model_id)
|
| 156 |
+
|
| 157 |
+
dataset = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation", trust_remote_code=True)
|
| 158 |
+
sample = dataset[0]
|
| 159 |
+
|
| 160 |
+
input_features = tokenizer(
|
| 161 |
+
sample["audio"]["array"],
|
| 162 |
+
sampling_rate=sample["audio"]["sampling_rate"],
|
| 163 |
+
return_tensors="pt",
|
| 164 |
+
).input_features
|
| 165 |
+
|
| 166 |
+
outputs = model.generate(input_features)
|
| 167 |
+
text = tokenizer.batch_decode(outputs)[0]
|
| 168 |
+
print(text)
|
| 169 |
+
```
|
| 170 |
+
|
| 171 |
+
## Running Model Inference with [OpenVINO GenAI](https://github.com/openvinotoolkit/openvino.genai)
|
| 172 |
+
|
| 173 |
+
1. Install packages required for using OpenVINO GenAI.
|
| 174 |
+
```
|
| 175 |
+
pip install huggingface_hub
|
| 176 |
+
pip install -U --pre --extra-index-url https://storage.openvinotoolkit.org/simple/wheels/nightly openvino openvino-tokenizers openvino-genai
|
| 177 |
+
```
|
| 178 |
+
|
| 179 |
+
2. Download model from HuggingFace Hub
|
| 180 |
+
|
| 181 |
+
```
|
| 182 |
+
import huggingface_hub as hf_hub
|
| 183 |
+
|
| 184 |
+
model_id = "OpenVINO/whisper-base-int4-ov"
|
| 185 |
+
model_path = "whisper-base-int4-ov"
|
| 186 |
+
|
| 187 |
+
hf_hub.snapshot_download(model_id, local_dir=model_path)
|
| 188 |
+
|
| 189 |
+
```
|
| 190 |
+
|
| 191 |
+
3. Run model inference:
|
| 192 |
+
|
| 193 |
+
```
|
| 194 |
+
import openvino_genai as ov_genai
|
| 195 |
+
import datasets
|
| 196 |
+
|
| 197 |
+
device = "CPU"
|
| 198 |
+
pipe = ov_genai.WhisperPipeline(model_path, device)
|
| 199 |
+
|
| 200 |
+
dataset = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation", trust_remote_code=True)
|
| 201 |
+
sample = dataset[0]["audio"]["array"]
|
| 202 |
+
print(pipe.generate(sample))
|
| 203 |
+
```
|
| 204 |
+
|
| 205 |
+
More GenAI usage examples can be found in OpenVINO GenAI library [docs](https://github.com/openvinotoolkit/openvino.genai/blob/master/src/README.md) and [samples](https://github.com/openvinotoolkit/openvino.genai?tab=readme-ov-file#openvino-genai-samples)
|
| 206 |
+
|
| 207 |
+
## Limitations
|
| 208 |
+
|
| 209 |
+
Check the original model card for [original model card](https://huggingface.co/openai/whisper-base) for limitations.
|
| 210 |
+
|
| 211 |
+
## Legal information
|
| 212 |
+
|
| 213 |
+
The original model is distributed under [apache-2.0](https://choosealicense.com/licenses/apache-2.0/) license. More details can be found in [original model card](https://huggingface.co/openai/whisper-base).
|
| 214 |
+
|
| 215 |
+
## Disclaimer
|
| 216 |
+
|
| 217 |
+
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.
|
README.md
CHANGED
|
@@ -1,33 +1,139 @@
|
|
| 1 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2 |
license: apache-2.0
|
| 3 |
license_link: https://choosealicense.com/licenses/apache-2.0/
|
| 4 |
-
base_model: openai/whisper-base
|
| 5 |
-
base_model_relation: quantized
|
| 6 |
---
|
|
|
|
| 7 |
# whisper-base-int4-ov
|
| 8 |
-
* Model creator: [
|
| 9 |
* Original model: [whisper-base](https://huggingface.co/openai/whisper-base)
|
| 10 |
|
| 11 |
## Description
|
| 12 |
-
This is [whisper-base](https://huggingface.co/openai/whisper-base) model converted to the [OpenVINO™ IR](https://docs.openvino.ai/
|
|
|
|
| 13 |
|
| 14 |
## Quantization Parameters
|
| 15 |
|
| 16 |
Weight compression was performed using `nncf.compress_weights` with the following parameters:
|
| 17 |
|
| 18 |
-
* mode: **
|
| 19 |
-
* ratio: **1**
|
| 20 |
* group_size: **128**
|
| 21 |
|
| 22 |
-
For more information on quantization, check the [OpenVINO model optimization guide](https://docs.openvino.ai/
|
| 23 |
|
| 24 |
|
| 25 |
## Compatibility
|
| 26 |
|
| 27 |
The provided OpenVINO™ IR model is compatible with:
|
| 28 |
|
| 29 |
-
* OpenVINO version
|
| 30 |
-
* Optimum Intel 1.
|
|
|
|
| 31 |
|
| 32 |
## Running Model Inference with [Optimum Intel](https://huggingface.co/docs/optimum/intel/index)
|
| 33 |
|
|
@@ -40,6 +146,7 @@ pip install optimum[openvino]
|
|
| 40 |
2. Run model inference:
|
| 41 |
|
| 42 |
```
|
|
|
|
| 43 |
from transformers import AutoProcessor
|
| 44 |
from optimum.intel.openvino import OVModelForSpeechSeq2Seq
|
| 45 |
|
|
@@ -50,14 +157,14 @@ model = OVModelForSpeechSeq2Seq.from_pretrained(model_id)
|
|
| 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 =
|
| 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 =
|
| 61 |
print(text)
|
| 62 |
```
|
| 63 |
|
|
@@ -91,7 +198,7 @@ device = "CPU"
|
|
| 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 |
|
|
|
|
| 1 |
---
|
| 2 |
+
language:
|
| 3 |
+
- en
|
| 4 |
+
- zh
|
| 5 |
+
- de
|
| 6 |
+
- es
|
| 7 |
+
- ru
|
| 8 |
+
- ko
|
| 9 |
+
- fr
|
| 10 |
+
- ja
|
| 11 |
+
- pt
|
| 12 |
+
- tr
|
| 13 |
+
- pl
|
| 14 |
+
- ca
|
| 15 |
+
- nl
|
| 16 |
+
- ar
|
| 17 |
+
- sv
|
| 18 |
+
- it
|
| 19 |
+
- id
|
| 20 |
+
- hi
|
| 21 |
+
- fi
|
| 22 |
+
- vi
|
| 23 |
+
- he
|
| 24 |
+
- uk
|
| 25 |
+
- el
|
| 26 |
+
- ms
|
| 27 |
+
- cs
|
| 28 |
+
- ro
|
| 29 |
+
- da
|
| 30 |
+
- hu
|
| 31 |
+
- ta
|
| 32 |
+
- no
|
| 33 |
+
- th
|
| 34 |
+
- ur
|
| 35 |
+
- hr
|
| 36 |
+
- bg
|
| 37 |
+
- lt
|
| 38 |
+
- la
|
| 39 |
+
- mi
|
| 40 |
+
- ml
|
| 41 |
+
- cy
|
| 42 |
+
- sk
|
| 43 |
+
- te
|
| 44 |
+
- fa
|
| 45 |
+
- lv
|
| 46 |
+
- bn
|
| 47 |
+
- sr
|
| 48 |
+
- az
|
| 49 |
+
- sl
|
| 50 |
+
- kn
|
| 51 |
+
- et
|
| 52 |
+
- mk
|
| 53 |
+
- br
|
| 54 |
+
- eu
|
| 55 |
+
- is
|
| 56 |
+
- hy
|
| 57 |
+
- ne
|
| 58 |
+
- mn
|
| 59 |
+
- bs
|
| 60 |
+
- kk
|
| 61 |
+
- sq
|
| 62 |
+
- sw
|
| 63 |
+
- gl
|
| 64 |
+
- mr
|
| 65 |
+
- pa
|
| 66 |
+
- si
|
| 67 |
+
- km
|
| 68 |
+
- sn
|
| 69 |
+
- yo
|
| 70 |
+
- so
|
| 71 |
+
- af
|
| 72 |
+
- oc
|
| 73 |
+
- ka
|
| 74 |
+
- be
|
| 75 |
+
- tg
|
| 76 |
+
- sd
|
| 77 |
+
- gu
|
| 78 |
+
- am
|
| 79 |
+
- yi
|
| 80 |
+
- lo
|
| 81 |
+
- uz
|
| 82 |
+
- fo
|
| 83 |
+
- ht
|
| 84 |
+
- ps
|
| 85 |
+
- tk
|
| 86 |
+
- nn
|
| 87 |
+
- mt
|
| 88 |
+
- sa
|
| 89 |
+
- lb
|
| 90 |
+
- my
|
| 91 |
+
- bo
|
| 92 |
+
- tl
|
| 93 |
+
- mg
|
| 94 |
+
- as
|
| 95 |
+
- tt
|
| 96 |
+
- haw
|
| 97 |
+
- ln
|
| 98 |
+
- ha
|
| 99 |
+
- ba
|
| 100 |
+
- jw
|
| 101 |
+
- su
|
| 102 |
+
tags:
|
| 103 |
+
- audio
|
| 104 |
+
- automatic-speech-recognition
|
| 105 |
+
- hf-asr-leaderboard
|
| 106 |
+
pipeline_tag: automatic-speech-recognition
|
| 107 |
license: apache-2.0
|
| 108 |
license_link: https://choosealicense.com/licenses/apache-2.0/
|
|
|
|
|
|
|
| 109 |
---
|
| 110 |
+
|
| 111 |
# whisper-base-int4-ov
|
| 112 |
+
* Model creator: [OpenAI](https://huggingface.co/openai)
|
| 113 |
* Original model: [whisper-base](https://huggingface.co/openai/whisper-base)
|
| 114 |
|
| 115 |
## Description
|
| 116 |
+
This is [whisper-base](https://huggingface.co/openai/whisper-base) model converted to the [OpenVINO™ IR](https://docs.openvino.ai/2025/documentation/openvino-ir-format.html) (Intermediate Representation) format with weights compressed to INT4 by [NNCF](https://github.com/openvinotoolkit/nncf).
|
| 117 |
+
|
| 118 |
|
| 119 |
## Quantization Parameters
|
| 120 |
|
| 121 |
Weight compression was performed using `nncf.compress_weights` with the following parameters:
|
| 122 |
|
| 123 |
+
* mode: **INT4_ASYM**
|
| 124 |
+
* ratio: **1.0**
|
| 125 |
* group_size: **128**
|
| 126 |
|
| 127 |
+
For more information on quantization, check the [OpenVINO model optimization guide](https://docs.openvino.ai/2025/openvino-workflow/model-optimization-guide/weight-compression.html).
|
| 128 |
|
| 129 |
|
| 130 |
## Compatibility
|
| 131 |
|
| 132 |
The provided OpenVINO™ IR model is compatible with:
|
| 133 |
|
| 134 |
+
* OpenVINO version 2025.1.0 and higher
|
| 135 |
+
* Optimum Intel 1.23.0 and higher
|
| 136 |
+
|
| 137 |
|
| 138 |
## Running Model Inference with [Optimum Intel](https://huggingface.co/docs/optimum/intel/index)
|
| 139 |
|
|
|
|
| 146 |
2. Run model inference:
|
| 147 |
|
| 148 |
```
|
| 149 |
+
from datasets import load_dataset
|
| 150 |
from transformers import AutoProcessor
|
| 151 |
from optimum.intel.openvino import OVModelForSpeechSeq2Seq
|
| 152 |
|
|
|
|
| 157 |
dataset = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation", trust_remote_code=True)
|
| 158 |
sample = dataset[0]
|
| 159 |
|
| 160 |
+
input_features = tokenizer(
|
| 161 |
sample["audio"]["array"],
|
| 162 |
sampling_rate=sample["audio"]["sampling_rate"],
|
| 163 |
return_tensors="pt",
|
| 164 |
).input_features
|
| 165 |
|
| 166 |
outputs = model.generate(input_features)
|
| 167 |
+
text = tokenizer.batch_decode(outputs)[0]
|
| 168 |
print(text)
|
| 169 |
```
|
| 170 |
|
|
|
|
| 198 |
pipe = ov_genai.WhisperPipeline(model_path, device)
|
| 199 |
|
| 200 |
dataset = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation", trust_remote_code=True)
|
| 201 |
+
sample = dataset[0]["audio"]["array"]
|
| 202 |
print(pipe.generate(sample))
|
| 203 |
```
|
| 204 |
|
config.json
CHANGED
|
@@ -1,6 +1,5 @@
|
|
| 1 |
{
|
| 2 |
"_attn_implementation_autoset": true,
|
| 3 |
-
"_name_or_path": "OpenVINO/whisper-base-int4-ov",
|
| 4 |
"activation_dropout": 0.0,
|
| 5 |
"activation_function": "gelu",
|
| 6 |
"apply_spec_augment": false,
|
|
@@ -54,7 +53,8 @@
|
|
| 54 |
"num_mel_bins": 80,
|
| 55 |
"pad_token_id": 50257,
|
| 56 |
"scale_embedding": false,
|
| 57 |
-
"
|
|
|
|
| 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_version": "4.51.3",
|
| 58 |
"use_cache": true,
|
| 59 |
"use_weighted_layer_sum": false,
|
| 60 |
"vocab_size": 51865
|
generation_config.json
CHANGED
|
@@ -252,5 +252,5 @@
|
|
| 252 |
"transcribe": 50359,
|
| 253 |
"translate": 50358
|
| 254 |
},
|
| 255 |
-
"transformers_version": "4.
|
| 256 |
}
|
|
|
|
| 252 |
"transcribe": 50359,
|
| 253 |
"translate": 50358
|
| 254 |
},
|
| 255 |
+
"transformers_version": "4.51.3"
|
| 256 |
}
|
openvino_decoder_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:83acf4fbb10cdfdedb25ccc4a2df74010500f69a3332f0530fbba5aee659c7d8
|
| 3 |
+
size 40228835
|
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:
|
| 3 |
-
size
|
|
|
|
| 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="
|
| 5 |
<data shape="?,?" element_type="i64" />
|
| 6 |
<output>
|
| 7 |
-
<port id="0" precision="I64" names="
|
| 8 |
<dim>-1</dim>
|
| 9 |
<dim>-1</dim>
|
| 10 |
</port>
|
| 11 |
</output>
|
| 12 |
</layer>
|
| 13 |
-
<layer id="1" name="
|
| 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="
|
| 29 |
-
<data element_type="
|
| 30 |
<output>
|
| 31 |
-
<port id="0" precision="
|
| 32 |
-
<dim>
|
| 33 |
</port>
|
| 34 |
</output>
|
| 35 |
</layer>
|
| 36 |
-
<layer id="3" name="
|
| 37 |
-
<data
|
| 38 |
-
<input>
|
| 39 |
-
<port id="0" precision="U8">
|
| 40 |
-
<dim>528248</dim>
|
| 41 |
-
</port>
|
| 42 |
-
</input>
|
| 43 |
<output>
|
| 44 |
-
<port id="
|
| 45 |
-
<dim
|
| 46 |
</port>
|
| 47 |
-
|
| 48 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 49 |
</port>
|
| 50 |
-
|
| 51 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 52 |
</port>
|
| 53 |
</output>
|
| 54 |
</layer>
|
| 55 |
-
<layer id="
|
| 56 |
-
<data skip_tokens="
|
| 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
|
| 64 |
</port>
|
| 65 |
<port id="2" precision="I32">
|
| 66 |
-
<dim
|
| 67 |
</port>
|
| 68 |
<port id="3" precision="U8">
|
| 69 |
-
<dim
|
| 70 |
</port>
|
| 71 |
-
</input>
|
| 72 |
-
<output>
|
| 73 |
<port id="4" precision="I32">
|
| 74 |
-
<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="
|
|
|
|
|
|
|
|
|
|
| 86 |
<dim>-1</dim>
|
| 87 |
</port>
|
| 88 |
</output>
|
| 89 |
</layer>
|
| 90 |
-
<layer id="
|
| 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="
|
| 115 |
-
<data
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 116 |
<output>
|
| 117 |
<port id="0" precision="U8">
|
| 118 |
-
<dim>
|
| 119 |
</port>
|
| 120 |
</output>
|
| 121 |
</layer>
|
| 122 |
-
<layer id="
|
| 123 |
-
<data element_type="u8" shape="2" offset="
|
| 124 |
<output>
|
| 125 |
<port id="0" precision="U8">
|
| 126 |
<dim>2</dim>
|
| 127 |
</port>
|
| 128 |
</output>
|
| 129 |
</layer>
|
| 130 |
-
<layer id="
|
| 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>
|
| 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="
|
| 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="
|
| 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="
|
| 191 |
-
<edge from-layer="2" from-port="0" to-layer="
|
| 192 |
-
<edge from-layer="3" from-port="
|
| 193 |
-
<edge from-layer="
|
| 194 |
-
<edge from-layer="
|
| 195 |
-
<edge from-layer="
|
| 196 |
-
<edge from-layer="
|
| 197 |
-
<edge from-layer="
|
| 198 |
-
<edge from-layer="
|
| 199 |
-
<edge from-layer="
|
| 200 |
-
<edge from-layer="
|
| 201 |
-
<edge from-layer="
|
| 202 |
-
<edge from-layer="
|
| 203 |
-
<edge from-layer="
|
| 204 |
-
<edge from-layer="8" from-port="5" to-layer="
|
| 205 |
-
<edge from-layer="
|
| 206 |
-
<edge from-layer="
|
| 207 |
-
<edge from-layer="
|
|
|
|
|
|
|
|
|
|
| 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="
|
| 220 |
-
<openvino_version value="
|
|
|
|
| 221 |
<original_tokenizer_class value="<class 'transformers.models.whisper.tokenization_whisper_fast.WhisperTokenizerFast'>" />
|
| 222 |
<pad_token_id value="50257" />
|
| 223 |
-
<
|
| 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.
|
| 229 |
-
<transformers_version value="4.
|
| 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_24340" type="Parameter" version="opset1">
|
| 5 |
<data shape="?,?" element_type="i64" />
|
| 6 |
<output>
|
| 7 |
+
<port id="0" precision="I64" names="Parameter_24340">
|
| 8 |
<dim>-1</dim>
|
| 9 |
<dim>-1</dim>
|
| 10 |
</port>
|
| 11 |
</output>
|
| 12 |
</layer>
|
| 13 |
+
<layer id="1" name="Convert_24535" 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_24342" 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_24344" 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_24346" 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_24351" 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_24353" 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_24354" 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_24355" 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_24357" 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_24359" 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_24360" 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_24361" 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_24362,string_output">
|
| 208 |
<dim>-1</dim>
|
| 209 |
</port>
|
| 210 |
</output>
|
| 211 |
</layer>
|
| 212 |
+
<layer id="13" name="Result_24362" type="Result" version="opset1" output_names="Result_24362,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="{"type": "TemplateProcessing", "single": [{"SpecialToken": {"id": "<|startoftranscript|>", "type_id": 0}}, {"SpecialToken": {"id": "<|notimestamps|>", "type_id": 0}}, {"Sequence": {"id": "A", "type_id": 0}}, {"SpecialToken": {"id": "<|endoftext|>", "type_id": 0}}], "pair": [{"SpecialToken": {"id": "<|startoftranscript|>", "type_id": 0}}, {"SpecialToken": {"id": "<|notimestamps|>", "type_id": 0}}, {"Sequence": {"id": "A", "type_id": 0}}, {"Sequence": {"id": "B", "type_id": 1}}, {"SpecialToken": {"id": "<|endoftext|>", "type_id": 1}}], "special_tokens": {"<|endoftext|>": {"id": "<|endoftext|>", "ids": [50257], "tokens": ["<|endoftext|>"]}, "<|notimestamps|>": {"id": "<|notimestamps|>", "ids": [50363], "tokens": ["<|notimestamps|>"]}, "<|startoftranscript|>": {"id": "<|startoftranscript|>", "ids": [50258], "tokens": ["<|startoftranscript|>"]}}}" />
|
| 258 |
<original_tokenizer_class value="<class 'transformers.models.whisper.tokenization_whisper_fast.WhisperTokenizerFast'>" />
|
| 259 |
<pad_token_id value="50257" />
|
| 260 |
+
<processed_post_processor_template value="{"single": {"ids": [50258, 50363, -1, 50257], "type_ids": [0, 0, 0, 0]}, "pair": {"ids": [50258, 50363, -1, -2, 50257], "type_ids": [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:
|
| 3 |
size 14451348
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a2bd6864a7ea31cf1347f4430d4b7613a61d7a748032a6548d31eab7404cf869
|
| 3 |
size 14451348
|
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:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:78b41c1d8b56734ca7a0942b5517e352074c375282398f5644efe1b02b439129
|
| 3 |
+
size 1902147
|
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="
|
| 5 |
<data shape="?" element_type="string" />
|
| 6 |
<output>
|
| 7 |
-
<port id="0" precision="STRING" names="
|
| 8 |
<dim>-1</dim>
|
| 9 |
</port>
|
| 10 |
</output>
|
| 11 |
</layer>
|
| 12 |
-
<layer id="1" name="
|
| 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="
|
| 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="
|
| 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="
|
| 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="
|
| 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="
|
| 45 |
<data element_type="i32" shape="1" offset="12" size="4" />
|
| 46 |
<output>
|
| 47 |
<port id="0" precision="I32">
|
|
@@ -49,14 +49,13 @@
|
|
| 49 |
</port>
|
| 50 |
</output>
|
| 51 |
</layer>
|
| 52 |
-
<layer id="7" name="
|
| 53 |
<data element_type="i64" shape="" offset="16" size="8" />
|
| 54 |
<output>
|
| 55 |
<port id="0" precision="I64" />
|
| 56 |
</output>
|
| 57 |
</layer>
|
| 58 |
-
<layer id="8" name="
|
| 59 |
-
<data mode="begins_ends" />
|
| 60 |
<input>
|
| 61 |
<port id="0" precision="STRING">
|
| 62 |
<dim>-1</dim>
|
|
@@ -74,7 +73,7 @@
|
|
| 74 |
</port>
|
| 75 |
</output>
|
| 76 |
</layer>
|
| 77 |
-
<layer id="9" name="
|
| 78 |
<data output_type="i64" />
|
| 79 |
<input>
|
| 80 |
<port id="0" precision="I32">
|
|
@@ -87,19 +86,19 @@
|
|
| 87 |
</port>
|
| 88 |
</output>
|
| 89 |
</layer>
|
| 90 |
-
<layer id="10" name="
|
| 91 |
<data element_type="i64" shape="" offset="16" size="8" />
|
| 92 |
<output>
|
| 93 |
<port id="0" precision="I64" />
|
| 94 |
</output>
|
| 95 |
</layer>
|
| 96 |
-
<layer id="11" name="
|
| 97 |
<data element_type="i64" shape="" offset="16" size="8" />
|
| 98 |
<output>
|
| 99 |
<port id="0" precision="I64" />
|
| 100 |
</output>
|
| 101 |
</layer>
|
| 102 |
-
<layer id="12" name="
|
| 103 |
<data batch_dims="0" />
|
| 104 |
<input>
|
| 105 |
<port id="0" precision="I64">
|
|
@@ -112,13 +111,13 @@
|
|
| 112 |
<port id="3" precision="I64" />
|
| 113 |
</output>
|
| 114 |
</layer>
|
| 115 |
-
<layer id="13" name="
|
| 116 |
<data element_type="i64" shape="" offset="24" size="8" />
|
| 117 |
<output>
|
| 118 |
<port id="0" precision="I64" />
|
| 119 |
</output>
|
| 120 |
</layer>
|
| 121 |
-
<layer id="14" name="
|
| 122 |
<data output_type="i32" />
|
| 123 |
<input>
|
| 124 |
<port id="0" precision="I64" />
|
|
@@ -131,19 +130,19 @@
|
|
| 131 |
</port>
|
| 132 |
</output>
|
| 133 |
</layer>
|
| 134 |
-
<layer id="15" name="
|
| 135 |
<data element_type="i64" shape="" offset="24" size="8" />
|
| 136 |
<output>
|
| 137 |
<port id="0" precision="I64" />
|
| 138 |
</output>
|
| 139 |
</layer>
|
| 140 |
-
<layer id="16" name="
|
| 141 |
<data element_type="i64" shape="" offset="24" size="8" />
|
| 142 |
<output>
|
| 143 |
<port id="0" precision="I64" />
|
| 144 |
</output>
|
| 145 |
</layer>
|
| 146 |
-
<layer id="17" name="
|
| 147 |
<data auto_broadcast="numpy" />
|
| 148 |
<input>
|
| 149 |
<port id="0" precision="I64" />
|
|
@@ -153,13 +152,13 @@
|
|
| 153 |
<port id="2" precision="I64" />
|
| 154 |
</output>
|
| 155 |
</layer>
|
| 156 |
-
<layer id="18" name="
|
| 157 |
<data element_type="i64" shape="" offset="24" size="8" />
|
| 158 |
<output>
|
| 159 |
<port id="0" precision="I64" />
|
| 160 |
</output>
|
| 161 |
</layer>
|
| 162 |
-
<layer id="19" name="
|
| 163 |
<data output_type="i32" />
|
| 164 |
<input>
|
| 165 |
<port id="0" precision="I64" />
|
|
@@ -172,7 +171,7 @@
|
|
| 172 |
</port>
|
| 173 |
</output>
|
| 174 |
</layer>
|
| 175 |
-
<layer id="20" name="
|
| 176 |
<data element_type="u8" shape="26479" offset="32" size="26479" />
|
| 177 |
<output>
|
| 178 |
<port id="0" precision="U8">
|
|
@@ -180,7 +179,7 @@
|
|
| 180 |
</port>
|
| 181 |
</output>
|
| 182 |
</layer>
|
| 183 |
-
<layer id="21" name="
|
| 184 |
<input>
|
| 185 |
<port id="0" precision="I32">
|
| 186 |
<dim>-1</dim>
|
|
@@ -222,15 +221,15 @@
|
|
| 222 |
</port>
|
| 223 |
</output>
|
| 224 |
</layer>
|
| 225 |
-
<layer id="22" name="
|
| 226 |
-
<data element_type="u8" shape="
|
| 227 |
<output>
|
| 228 |
<port id="0" precision="U8">
|
| 229 |
-
<dim>
|
| 230 |
</port>
|
| 231 |
</output>
|
| 232 |
</layer>
|
| 233 |
-
<layer id="23" name="
|
| 234 |
<data behaviour="isolate" invert="false" max_splits="-1" />
|
| 235 |
<input>
|
| 236 |
<port id="0" precision="I32">
|
|
@@ -252,7 +251,7 @@
|
|
| 252 |
<dim>-1</dim>
|
| 253 |
</port>
|
| 254 |
<port id="6" precision="U8">
|
| 255 |
-
<dim>
|
| 256 |
</port>
|
| 257 |
</input>
|
| 258 |
<output>
|
|
@@ -276,123 +275,111 @@
|
|
| 276 |
</port>
|
| 277 |
</output>
|
| 278 |
</layer>
|
| 279 |
-
<layer id="24" name="
|
| 280 |
-
<data element_type="
|
| 281 |
<output>
|
| 282 |
-
<port id="0" precision="
|
| 283 |
-
<dim>
|
| 284 |
</port>
|
| 285 |
</output>
|
| 286 |
</layer>
|
| 287 |
-
<layer id="25" name="
|
| 288 |
-
<data
|
| 289 |
-
<input>
|
| 290 |
-
<port id="0" precision="U8">
|
| 291 |
-
<dim>528248</dim>
|
| 292 |
-
</port>
|
| 293 |
-
</input>
|
| 294 |
<output>
|
| 295 |
-
<port id="
|
| 296 |
-
<dim
|
| 297 |
-
</port>
|
| 298 |
-
<port id="2" precision="I32">
|
| 299 |
-
<dim>-1</dim>
|
| 300 |
-
</port>
|
| 301 |
-
<port id="3" precision="U8">
|
| 302 |
-
<dim>-1</dim>
|
| 303 |
</port>
|
| 304 |
</output>
|
| 305 |
</layer>
|
| 306 |
-
<layer id="26" name="
|
| 307 |
-
<data element_type="u8" shape="
|
| 308 |
<output>
|
| 309 |
<port id="0" precision="U8">
|
| 310 |
-
<dim>
|
| 311 |
</port>
|
| 312 |
</output>
|
| 313 |
</layer>
|
| 314 |
-
<layer id="27" name="
|
| 315 |
-
<data
|
| 316 |
-
<input>
|
| 317 |
-
<port id="0" precision="U8">
|
| 318 |
-
<dim>369958</dim>
|
| 319 |
-
</port>
|
| 320 |
-
</input>
|
| 321 |
<output>
|
| 322 |
-
<port id="
|
| 323 |
-
<dim
|
| 324 |
-
</port>
|
| 325 |
-
<port id="2" precision="I32">
|
| 326 |
-
<dim>-1</dim>
|
| 327 |
-
</port>
|
| 328 |
-
<port id="3" precision="U8">
|
| 329 |
-
<dim>-1</dim>
|
| 330 |
</port>
|
| 331 |
</output>
|
| 332 |
</layer>
|
| 333 |
-
<layer id="28" name="
|
| 334 |
-
<data element_type="
|
| 335 |
<output>
|
| 336 |
-
<port id="0" precision="
|
| 337 |
-
<dim>
|
| 338 |
</port>
|
| 339 |
</output>
|
| 340 |
</layer>
|
| 341 |
-
<layer id="29" name="
|
| 342 |
-
<data
|
| 343 |
-
<
|
| 344 |
<port id="0" precision="U8">
|
| 345 |
-
<dim>
|
| 346 |
</port>
|
| 347 |
-
</
|
|
|
|
|
|
|
|
|
|
| 348 |
<output>
|
| 349 |
-
<port id="
|
| 350 |
-
<dim
|
| 351 |
-
</port>
|
| 352 |
-
<port id="2" precision="I32">
|
| 353 |
-
<dim>-1</dim>
|
| 354 |
-
</port>
|
| 355 |
-
<port id="3" precision="U8">
|
| 356 |
-
<dim>-1</dim>
|
| 357 |
</port>
|
| 358 |
</output>
|
| 359 |
</layer>
|
| 360 |
-
<layer id="
|
| 361 |
-
<data element_type="
|
| 362 |
<output>
|
| 363 |
-
<port id="0" precision="
|
| 364 |
-
<dim>
|
| 365 |
</port>
|
| 366 |
</output>
|
| 367 |
</layer>
|
| 368 |
-
<layer id="
|
| 369 |
-
<data
|
| 370 |
-
<
|
| 371 |
<port id="0" precision="U8">
|
| 372 |
-
<dim>
|
| 373 |
</port>
|
| 374 |
-
</
|
|
|
|
|
|
|
|
|
|
| 375 |
<output>
|
| 376 |
-
<port id="
|
| 377 |
-
<dim
|
| 378 |
</port>
|
| 379 |
-
|
| 380 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 381 |
</port>
|
| 382 |
-
|
| 383 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 384 |
</port>
|
| 385 |
</output>
|
| 386 |
</layer>
|
| 387 |
-
<layer id="
|
| 388 |
-
<data element_type="i32" shape="1608" offset="
|
| 389 |
<output>
|
| 390 |
<port id="0" precision="I32">
|
| 391 |
<dim>1608</dim>
|
| 392 |
</port>
|
| 393 |
</output>
|
| 394 |
</layer>
|
| 395 |
-
<layer id="
|
| 396 |
<data unk_token="" fuse_unk="false" suffix_indicator="" end_suffix="" byte_fallback="false" cache_capacity="20000" />
|
| 397 |
<input>
|
| 398 |
<port id="0" precision="I32">
|
|
@@ -411,40 +398,40 @@
|
|
| 411 |
<dim>-1</dim>
|
| 412 |
</port>
|
| 413 |
<port id="5" precision="I32">
|
| 414 |
-
<dim
|
| 415 |
</port>
|
| 416 |
<port id="6" precision="I32">
|
| 417 |
-
<dim
|
| 418 |
</port>
|
| 419 |
<port id="7" precision="U8">
|
| 420 |
-
<dim
|
| 421 |
</port>
|
| 422 |
<port id="8" precision="I32">
|
| 423 |
-
<dim
|
| 424 |
</port>
|
| 425 |
<port id="9" precision="I32">
|
| 426 |
-
<dim
|
| 427 |
</port>
|
| 428 |
<port id="10" precision="U8">
|
| 429 |
-
<dim
|
| 430 |
</port>
|
| 431 |
<port id="11" precision="I32">
|
| 432 |
-
<dim
|
| 433 |
</port>
|
| 434 |
<port id="12" precision="I32">
|
| 435 |
-
<dim
|
| 436 |
</port>
|
| 437 |
<port id="13" precision="U8">
|
| 438 |
-
<dim
|
| 439 |
</port>
|
| 440 |
<port id="14" precision="I32">
|
| 441 |
-
<dim
|
| 442 |
</port>
|
| 443 |
<port id="15" precision="I32">
|
| 444 |
-
<dim
|
| 445 |
</port>
|
| 446 |
<port id="16" precision="U8">
|
| 447 |
-
<dim
|
| 448 |
</port>
|
| 449 |
<port id="17" precision="I32">
|
| 450 |
<dim>1608</dim>
|
|
@@ -462,44 +449,30 @@
|
|
| 462 |
</port>
|
| 463 |
</output>
|
| 464 |
</layer>
|
| 465 |
-
<layer id="
|
| 466 |
-
<data
|
| 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 |
-
</input>
|
| 475 |
<output>
|
| 476 |
-
<port id="
|
| 477 |
-
<dim>-1</dim>
|
| 478 |
-
</port>
|
| 479 |
</output>
|
| 480 |
</layer>
|
| 481 |
-
<layer id="
|
| 482 |
-
<data element_type="
|
| 483 |
<output>
|
| 484 |
-
<port id="0" precision="
|
|
|
|
|
|
|
| 485 |
</output>
|
| 486 |
</layer>
|
| 487 |
-
<layer id="
|
| 488 |
-
<data
|
| 489 |
-
<input>
|
| 490 |
-
<port id="0" precision="I32">
|
| 491 |
-
<dim>-1</dim>
|
| 492 |
-
</port>
|
| 493 |
-
<port id="1" precision="I32" />
|
| 494 |
-
</input>
|
| 495 |
<output>
|
| 496 |
-
<port id="
|
| 497 |
-
<dim
|
| 498 |
</port>
|
| 499 |
</output>
|
| 500 |
</layer>
|
| 501 |
-
<layer id="
|
| 502 |
-
<data
|
| 503 |
<input>
|
| 504 |
<port id="0" precision="I32">
|
| 505 |
<dim>-1</dim>
|
|
@@ -507,42 +480,58 @@
|
|
| 507 |
<port id="1" precision="I32">
|
| 508 |
<dim>-1</dim>
|
| 509 |
</port>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 510 |
</input>
|
| 511 |
<output>
|
| 512 |
-
<port id="
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 513 |
<dim>-1</dim>
|
| 514 |
</port>
|
| 515 |
</output>
|
| 516 |
</layer>
|
| 517 |
-
<layer id="
|
| 518 |
<data element_type="i32" shape="" offset="0" size="4" />
|
| 519 |
<output>
|
| 520 |
<port id="0" precision="I32" />
|
| 521 |
</output>
|
| 522 |
</layer>
|
| 523 |
-
<layer id="
|
| 524 |
<data element_type="i32" shape="" offset="4" size="4" />
|
| 525 |
<output>
|
| 526 |
<port id="0" precision="I32" />
|
| 527 |
</output>
|
| 528 |
</layer>
|
| 529 |
-
<layer id="
|
| 530 |
-
<data element_type="i32" shape="1" offset="
|
| 531 |
<output>
|
| 532 |
<port id="0" precision="I32">
|
| 533 |
<dim>1</dim>
|
| 534 |
</port>
|
| 535 |
</output>
|
| 536 |
</layer>
|
| 537 |
-
<layer id="
|
| 538 |
-
<data element_type="i32" shape="4" offset="
|
| 539 |
<output>
|
| 540 |
<port id="0" precision="I32">
|
| 541 |
<dim>4</dim>
|
| 542 |
</port>
|
| 543 |
</output>
|
| 544 |
</layer>
|
| 545 |
-
<layer id="
|
| 546 |
<input>
|
| 547 |
<port id="0" precision="I32" />
|
| 548 |
<port id="1" precision="I32" />
|
|
@@ -593,7 +582,7 @@
|
|
| 593 |
</port>
|
| 594 |
</output>
|
| 595 |
</layer>
|
| 596 |
-
<layer id="
|
| 597 |
<data auto_broadcast="numpy" />
|
| 598 |
<input>
|
| 599 |
<port id="0" precision="I32">
|
|
@@ -609,13 +598,13 @@
|
|
| 609 |
</port>
|
| 610 |
</output>
|
| 611 |
</layer>
|
| 612 |
-
<layer id="
|
| 613 |
<data element_type="i32" shape="" offset="0" size="4" />
|
| 614 |
<output>
|
| 615 |
<port id="0" precision="I32" />
|
| 616 |
</output>
|
| 617 |
</layer>
|
| 618 |
-
<layer id="
|
| 619 |
<data keep_dims="false" />
|
| 620 |
<input>
|
| 621 |
<port id="0" precision="I32">
|
|
@@ -627,14 +616,14 @@
|
|
| 627 |
<port id="2" precision="I32" />
|
| 628 |
</output>
|
| 629 |
</layer>
|
| 630 |
-
<layer id="
|
| 631 |
-
<data element_type="i32" shape="" offset="
|
| 632 |
<output>
|
| 633 |
<port id="0" precision="I32" />
|
| 634 |
</output>
|
| 635 |
</layer>
|
| 636 |
-
<layer id="
|
| 637 |
-
<data pad_right="true" />
|
| 638 |
<input>
|
| 639 |
<port id="0" precision="I32">
|
| 640 |
<dim>-1</dim>
|
|
@@ -659,7 +648,7 @@
|
|
| 659 |
</port>
|
| 660 |
</output>
|
| 661 |
</layer>
|
| 662 |
-
<layer id="
|
| 663 |
<data destination_type="i32" />
|
| 664 |
<input>
|
| 665 |
<port id="0" precision="BOOL">
|
|
@@ -674,7 +663,7 @@
|
|
| 674 |
</port>
|
| 675 |
</output>
|
| 676 |
</layer>
|
| 677 |
-
<layer id="
|
| 678 |
<data destination_type="i64" />
|
| 679 |
<input>
|
| 680 |
<port id="0" precision="I32">
|
|
@@ -689,7 +678,7 @@
|
|
| 689 |
</port>
|
| 690 |
</output>
|
| 691 |
</layer>
|
| 692 |
-
<layer id="
|
| 693 |
<data destination_type="i64" />
|
| 694 |
<input>
|
| 695 |
<port id="0" precision="I32">
|
|
@@ -704,7 +693,7 @@
|
|
| 704 |
</port>
|
| 705 |
</output>
|
| 706 |
</layer>
|
| 707 |
-
<layer id="
|
| 708 |
<input>
|
| 709 |
<port id="0" precision="I64">
|
| 710 |
<dim>-1</dim>
|
|
@@ -712,7 +701,7 @@
|
|
| 712 |
</port>
|
| 713 |
</input>
|
| 714 |
</layer>
|
| 715 |
-
<layer id="
|
| 716 |
<input>
|
| 717 |
<port id="0" precision="I64">
|
| 718 |
<dim>-1</dim>
|
|
@@ -723,12 +712,12 @@
|
|
| 723 |
</layers>
|
| 724 |
<edges>
|
| 725 |
<edge from-layer="0" from-port="0" to-layer="8" to-port="0" />
|
| 726 |
-
<edge from-layer="1" from-port="0" to-layer="
|
| 727 |
-
<edge from-layer="2" from-port="0" to-layer="
|
| 728 |
-
<edge from-layer="3" from-port="0" to-layer="
|
| 729 |
-
<edge from-layer="4" from-port="0" to-layer="
|
| 730 |
-
<edge from-layer="5" from-port="0" to-layer="
|
| 731 |
-
<edge from-layer="6" from-port="0" to-layer="
|
| 732 |
<edge from-layer="7" from-port="0" to-layer="14" to-port="0" />
|
| 733 |
<edge from-layer="8" from-port="1" to-layer="9" to-port="0" />
|
| 734 |
<edge from-layer="8" from-port="3" to-layer="21" to-port="4" />
|
|
@@ -737,8 +726,8 @@
|
|
| 737 |
<edge from-layer="9" from-port="1" to-layer="12" to-port="0" />
|
| 738 |
<edge from-layer="10" from-port="0" to-layer="12" to-port="1" />
|
| 739 |
<edge from-layer="11" from-port="0" to-layer="12" to-port="2" />
|
| 740 |
-
<edge from-layer="12" from-port="3" to-layer="14" to-port="1" />
|
| 741 |
<edge from-layer="12" from-port="3" to-layer="17" to-port="0" />
|
|
|
|
| 742 |
<edge from-layer="13" from-port="0" to-layer="14" to-port="2" />
|
| 743 |
<edge from-layer="14" from-port="3" to-layer="21" to-port="0" />
|
| 744 |
<edge from-layer="15" from-port="0" to-layer="19" to-port="0" />
|
|
@@ -747,62 +736,58 @@
|
|
| 747 |
<edge from-layer="18" from-port="0" to-layer="19" to-port="2" />
|
| 748 |
<edge from-layer="19" from-port="3" to-layer="21" to-port="1" />
|
| 749 |
<edge from-layer="20" from-port="0" to-layer="21" to-port="5" />
|
|
|
|
| 750 |
<edge from-layer="21" from-port="7" to-layer="23" to-port="1" />
|
| 751 |
<edge from-layer="21" from-port="8" to-layer="23" to-port="2" />
|
| 752 |
<edge from-layer="21" from-port="9" to-layer="23" to-port="3" />
|
| 753 |
<edge from-layer="21" from-port="10" to-layer="23" to-port="4" />
|
| 754 |
<edge from-layer="21" from-port="11" to-layer="23" to-port="5" />
|
| 755 |
-
<edge from-layer="21" from-port="6" to-layer="23" to-port="0" />
|
| 756 |
<edge from-layer="22" from-port="0" to-layer="23" to-port="6" />
|
| 757 |
-
<edge from-layer="23" from-port="
|
| 758 |
-
<edge from-layer="23" from-port="11" to-layer="
|
| 759 |
-
<edge from-layer="23" from-port="10" to-layer="
|
| 760 |
-
<edge from-layer="23" from-port="
|
| 761 |
-
<edge from-layer="23" from-port="
|
| 762 |
-
<edge from-layer="24" from-port="0" to-layer="
|
| 763 |
-
<edge from-layer="25" from-port="
|
| 764 |
-
<edge from-layer="
|
| 765 |
-
<edge from-layer="
|
| 766 |
-
<edge from-layer="
|
| 767 |
-
<edge from-layer="
|
| 768 |
-
<edge from-layer="
|
| 769 |
-
<edge from-layer="
|
| 770 |
-
<edge from-layer="
|
| 771 |
-
<edge from-layer="
|
| 772 |
-
<edge from-layer="
|
| 773 |
-
<edge from-layer="
|
| 774 |
-
<edge from-layer="
|
| 775 |
-
<edge from-layer="
|
| 776 |
-
<edge from-layer="
|
| 777 |
-
<edge from-layer="
|
| 778 |
-
<edge from-layer="
|
| 779 |
-
<edge from-layer="
|
| 780 |
-
<edge from-layer="
|
| 781 |
-
<edge from-layer="
|
| 782 |
-
<edge from-layer="
|
| 783 |
-
<edge from-layer="
|
| 784 |
-
<edge from-layer="
|
| 785 |
-
<edge from-layer="
|
| 786 |
-
<edge from-layer="
|
| 787 |
-
<edge from-layer="
|
| 788 |
-
<edge from-layer="
|
| 789 |
-
<edge from-layer="
|
| 790 |
-
<edge from-layer="
|
| 791 |
-
<edge from-layer="
|
| 792 |
-
<edge from-layer="
|
| 793 |
-
<edge from-layer="
|
| 794 |
-
<edge from-layer="
|
| 795 |
-
<edge from-layer="
|
| 796 |
-
<edge from-layer="
|
| 797 |
-
<edge from-layer="
|
| 798 |
-
<edge from-layer="
|
| 799 |
-
<edge from-layer="
|
| 800 |
-
<edge from-layer="
|
| 801 |
-
<edge from-layer="
|
| 802 |
-
<edge from-layer="47" from-port="5" to-layer="51" to-port="0" />
|
| 803 |
-
<edge from-layer="48" from-port="1" to-layer="49" to-port="0" />
|
| 804 |
-
<edge from-layer="49" from-port="1" to-layer="50" to-port="0" />
|
| 805 |
-
<edge from-layer="51" from-port="1" to-layer="52" to-port="0" />
|
| 806 |
</edges>
|
| 807 |
<rt_info>
|
| 808 |
<add_attention_mask value="True" />
|
|
@@ -813,21 +798,22 @@
|
|
| 813 |
<detokenizer_input_type value="i64" />
|
| 814 |
<eos_token_id value="50257" />
|
| 815 |
<handle_special_tokens_with_re />
|
|
|
|
| 816 |
<number_of_inputs value="1" />
|
| 817 |
-
<openvino_tokenizers_version value="
|
| 818 |
-
<openvino_version value="
|
|
|
|
| 819 |
<original_tokenizer_class value="<class 'transformers.models.whisper.tokenization_whisper_fast.WhisperTokenizerFast'>" />
|
| 820 |
<pad_token_id value="50257" />
|
| 821 |
-
<
|
| 822 |
<skip_special_tokens value="True" />
|
| 823 |
<streaming_detokenizer value="False" />
|
| 824 |
-
<tiktoken_version value="0.8.0" />
|
| 825 |
<tokenizer_output_type value="i64" />
|
| 826 |
-
<tokenizers_version value="0.
|
| 827 |
-
<transformers_version value="4.
|
| 828 |
<use_max_padding value="False" />
|
| 829 |
<use_sentencepiece_backend value="False" />
|
| 830 |
-
<utf8_replace_mode />
|
| 831 |
<with_detokenizer value="True" />
|
| 832 |
</rt_info>
|
| 833 |
</net>
|
|
|
|
| 1 |
<?xml version="1.0"?>
|
| 2 |
<net name="tokenizer" version="11">
|
| 3 |
<layers>
|
| 4 |
+
<layer id="0" name="Parameter_24206" type="Parameter" version="opset1">
|
| 5 |
<data shape="?" element_type="string" />
|
| 6 |
<output>
|
| 7 |
+
<port id="0" precision="STRING" names="Parameter_24206">
|
| 8 |
<dim>-1</dim>
|
| 9 |
</port>
|
| 10 |
</output>
|
| 11 |
</layer>
|
| 12 |
+
<layer id="1" name="Constant_24318" 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_24319" 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_24320" 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_24321" 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_24322" 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_24323" 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_24212" type="Const" version="opset1">
|
| 53 |
<data element_type="i64" shape="" offset="16" size="8" />
|
| 54 |
<output>
|
| 55 |
<port id="0" precision="I64" />
|
| 56 |
</output>
|
| 57 |
</layer>
|
| 58 |
+
<layer id="8" name="StringTensorUnpack_24207" type="StringTensorUnpack" version="opset15">
|
|
|
|
| 59 |
<input>
|
| 60 |
<port id="0" precision="STRING">
|
| 61 |
<dim>-1</dim>
|
|
|
|
| 73 |
</port>
|
| 74 |
</output>
|
| 75 |
</layer>
|
| 76 |
+
<layer id="9" name="ShapeOf_24208" type="ShapeOf" version="opset3">
|
| 77 |
<data output_type="i64" />
|
| 78 |
<input>
|
| 79 |
<port id="0" precision="I32">
|
|
|
|
| 86 |
</port>
|
| 87 |
</output>
|
| 88 |
</layer>
|
| 89 |
+
<layer id="10" name="Constant_24209" type="Const" version="opset1">
|
| 90 |
<data element_type="i64" shape="" offset="16" size="8" />
|
| 91 |
<output>
|
| 92 |
<port id="0" precision="I64" />
|
| 93 |
</output>
|
| 94 |
</layer>
|
| 95 |
+
<layer id="11" name="Constant_24210" type="Const" version="opset1">
|
| 96 |
<data element_type="i64" shape="" offset="16" size="8" />
|
| 97 |
<output>
|
| 98 |
<port id="0" precision="I64" />
|
| 99 |
</output>
|
| 100 |
</layer>
|
| 101 |
+
<layer id="12" name="Gather_24211" type="Gather" version="opset8">
|
| 102 |
<data batch_dims="0" />
|
| 103 |
<input>
|
| 104 |
<port id="0" precision="I64">
|
|
|
|
| 111 |
<port id="3" precision="I64" />
|
| 112 |
</output>
|
| 113 |
</layer>
|
| 114 |
+
<layer id="13" name="Constant_24213" type="Const" version="opset1">
|
| 115 |
<data element_type="i64" shape="" offset="24" size="8" />
|
| 116 |
<output>
|
| 117 |
<port id="0" precision="I64" />
|
| 118 |
</output>
|
| 119 |
</layer>
|
| 120 |
+
<layer id="14" name="Range_24214" type="Range" version="opset4">
|
| 121 |
<data output_type="i32" />
|
| 122 |
<input>
|
| 123 |
<port id="0" precision="I64" />
|
|
|
|
| 130 |
</port>
|
| 131 |
</output>
|
| 132 |
</layer>
|
| 133 |
+
<layer id="15" name="Constant_24215" type="Const" version="opset1">
|
| 134 |
<data element_type="i64" shape="" offset="24" size="8" />
|
| 135 |
<output>
|
| 136 |
<port id="0" precision="I64" />
|
| 137 |
</output>
|
| 138 |
</layer>
|
| 139 |
+
<layer id="16" name="Constant_24216" type="Const" version="opset1">
|
| 140 |
<data element_type="i64" shape="" offset="24" size="8" />
|
| 141 |
<output>
|
| 142 |
<port id="0" precision="I64" />
|
| 143 |
</output>
|
| 144 |
</layer>
|
| 145 |
+
<layer id="17" name="Add_24217" type="Add" version="opset1">
|
| 146 |
<data auto_broadcast="numpy" />
|
| 147 |
<input>
|
| 148 |
<port id="0" precision="I64" />
|
|
|
|
| 152 |
<port id="2" precision="I64" />
|
| 153 |
</output>
|
| 154 |
</layer>
|
| 155 |
+
<layer id="18" name="Constant_24218" type="Const" version="opset1">
|
| 156 |
<data element_type="i64" shape="" offset="24" size="8" />
|
| 157 |
<output>
|
| 158 |
<port id="0" precision="I64" />
|
| 159 |
</output>
|
| 160 |
</layer>
|
| 161 |
+
<layer id="19" name="Range_24219" type="Range" version="opset4">
|
| 162 |
<data output_type="i32" />
|
| 163 |
<input>
|
| 164 |
<port id="0" precision="I64" />
|
|
|
|
| 171 |
</port>
|
| 172 |
</output>
|
| 173 |
</layer>
|
| 174 |
+
<layer id="20" name="Constant_24281" type="Const" version="opset1">
|
| 175 |
<data element_type="u8" shape="26479" offset="32" size="26479" />
|
| 176 |
<output>
|
| 177 |
<port id="0" precision="U8">
|
|
|
|
| 179 |
</port>
|
| 180 |
</output>
|
| 181 |
</layer>
|
| 182 |
+
<layer id="21" name="SpecialTokensSplit_24282" type="SpecialTokensSplit" version="extension">
|
| 183 |
<input>
|
| 184 |
<port id="0" precision="I32">
|
| 185 |
<dim>-1</dim>
|
|
|
|
| 221 |
</port>
|
| 222 |
</output>
|
| 223 |
</layer>
|
| 224 |
+
<layer id="22" name="Constant_24284" type="Const" version="opset1">
|
| 225 |
+
<data element_type="u8" shape="74" offset="26511" size="74" />
|
| 226 |
<output>
|
| 227 |
<port id="0" precision="U8">
|
| 228 |
+
<dim>74</dim>
|
| 229 |
</port>
|
| 230 |
</output>
|
| 231 |
</layer>
|
| 232 |
+
<layer id="23" name="RegexSplit_24285" type="RegexSplit" version="extension">
|
| 233 |
<data behaviour="isolate" invert="false" max_splits="-1" />
|
| 234 |
<input>
|
| 235 |
<port id="0" precision="I32">
|
|
|
|
| 251 |
<dim>-1</dim>
|
| 252 |
</port>
|
| 253 |
<port id="6" precision="U8">
|
| 254 |
+
<dim>74</dim>
|
| 255 |
</port>
|
| 256 |
</input>
|
| 257 |
<output>
|
|
|
|
| 275 |
</port>
|
| 276 |
</output>
|
| 277 |
</layer>
|
| 278 |
+
<layer id="24" name="Constant_24287" type="Const" version="opset1">
|
| 279 |
+
<data element_type="i32" shape="51865" offset="26585" size="207460" />
|
| 280 |
<output>
|
| 281 |
+
<port id="0" precision="I32">
|
| 282 |
+
<dim>51865</dim>
|
| 283 |
</port>
|
| 284 |
</output>
|
| 285 |
</layer>
|
| 286 |
+
<layer id="25" name="Constant_24289" type="Const" version="opset1">
|
| 287 |
+
<data element_type="i32" shape="51865" offset="234045" size="207460" />
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 288 |
<output>
|
| 289 |
+
<port id="0" precision="I32">
|
| 290 |
+
<dim>51865</dim>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 291 |
</port>
|
| 292 |
</output>
|
| 293 |
</layer>
|
| 294 |
+
<layer id="26" name="Constant_24291" type="Const" version="opset1">
|
| 295 |
+
<data element_type="u8" shape="320780" offset="441505" size="320780" />
|
| 296 |
<output>
|
| 297 |
<port id="0" precision="U8">
|
| 298 |
+
<dim>320780</dim>
|
| 299 |
</port>
|
| 300 |
</output>
|
| 301 |
</layer>
|
| 302 |
+
<layer id="27" name="Constant_24299" type="Const" version="opset1">
|
| 303 |
+
<data element_type="i32" shape="50000" offset="762285" size="200000" />
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 304 |
<output>
|
| 305 |
+
<port id="0" precision="I32">
|
| 306 |
+
<dim>50000</dim>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 307 |
</port>
|
| 308 |
</output>
|
| 309 |
</layer>
|
| 310 |
+
<layer id="28" name="Constant_24301" type="Const" version="opset1">
|
| 311 |
+
<data element_type="i32" shape="50000" offset="962285" size="200000" />
|
| 312 |
<output>
|
| 313 |
+
<port id="0" precision="I32">
|
| 314 |
+
<dim>50000</dim>
|
| 315 |
</port>
|
| 316 |
</output>
|
| 317 |
</layer>
|
| 318 |
+
<layer id="29" name="Constant_24303" type="Const" version="opset1">
|
| 319 |
+
<data element_type="u8" shape="169950" offset="1162285" size="169950" />
|
| 320 |
+
<output>
|
| 321 |
<port id="0" precision="U8">
|
| 322 |
+
<dim>169950</dim>
|
| 323 |
</port>
|
| 324 |
+
</output>
|
| 325 |
+
</layer>
|
| 326 |
+
<layer id="30" name="Constant_24305" type="Const" version="opset1">
|
| 327 |
+
<data element_type="i32" shape="50000" offset="1332235" size="200000" />
|
| 328 |
<output>
|
| 329 |
+
<port id="0" precision="I32">
|
| 330 |
+
<dim>50000</dim>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 331 |
</port>
|
| 332 |
</output>
|
| 333 |
</layer>
|
| 334 |
+
<layer id="31" name="Constant_24307" type="Const" version="opset1">
|
| 335 |
+
<data element_type="i32" shape="50000" offset="1532235" size="200000" />
|
| 336 |
<output>
|
| 337 |
+
<port id="0" precision="I32">
|
| 338 |
+
<dim>50000</dim>
|
| 339 |
</port>
|
| 340 |
</output>
|
| 341 |
</layer>
|
| 342 |
+
<layer id="32" name="Constant_24309" type="Const" version="opset1">
|
| 343 |
+
<data element_type="u8" shape="136851" offset="1732235" size="136851" />
|
| 344 |
+
<output>
|
| 345 |
<port id="0" precision="U8">
|
| 346 |
+
<dim>136851</dim>
|
| 347 |
</port>
|
| 348 |
+
</output>
|
| 349 |
+
</layer>
|
| 350 |
+
<layer id="33" name="Constant_24293" type="Const" version="opset1">
|
| 351 |
+
<data element_type="i32" shape="1608" offset="1869086" size="6432" />
|
| 352 |
<output>
|
| 353 |
+
<port id="0" precision="I32">
|
| 354 |
+
<dim>1608</dim>
|
| 355 |
</port>
|
| 356 |
+
</output>
|
| 357 |
+
</layer>
|
| 358 |
+
<layer id="34" name="Constant_24295" type="Const" version="opset1">
|
| 359 |
+
<data element_type="i32" shape="1608" offset="1875518" size="6432" />
|
| 360 |
+
<output>
|
| 361 |
+
<port id="0" precision="I32">
|
| 362 |
+
<dim>1608</dim>
|
| 363 |
</port>
|
| 364 |
+
</output>
|
| 365 |
+
</layer>
|
| 366 |
+
<layer id="35" name="Constant_24297" type="Const" version="opset1">
|
| 367 |
+
<data element_type="u8" shape="13723" offset="1881950" size="13723" />
|
| 368 |
+
<output>
|
| 369 |
+
<port id="0" precision="U8">
|
| 370 |
+
<dim>13723</dim>
|
| 371 |
</port>
|
| 372 |
</output>
|
| 373 |
</layer>
|
| 374 |
+
<layer id="36" name="Constant_24310" type="Const" version="opset1">
|
| 375 |
+
<data element_type="i32" shape="1608" offset="1895673" size="6432" />
|
| 376 |
<output>
|
| 377 |
<port id="0" precision="I32">
|
| 378 |
<dim>1608</dim>
|
| 379 |
</port>
|
| 380 |
</output>
|
| 381 |
</layer>
|
| 382 |
+
<layer id="37" name="BPETokenizer_24311" type="BPETokenizer" version="extension">
|
| 383 |
<data unk_token="" fuse_unk="false" suffix_indicator="" end_suffix="" byte_fallback="false" cache_capacity="20000" />
|
| 384 |
<input>
|
| 385 |
<port id="0" precision="I32">
|
|
|
|
| 398 |
<dim>-1</dim>
|
| 399 |
</port>
|
| 400 |
<port id="5" precision="I32">
|
| 401 |
+
<dim>51865</dim>
|
| 402 |
</port>
|
| 403 |
<port id="6" precision="I32">
|
| 404 |
+
<dim>51865</dim>
|
| 405 |
</port>
|
| 406 |
<port id="7" precision="U8">
|
| 407 |
+
<dim>320780</dim>
|
| 408 |
</port>
|
| 409 |
<port id="8" precision="I32">
|
| 410 |
+
<dim>50000</dim>
|
| 411 |
</port>
|
| 412 |
<port id="9" precision="I32">
|
| 413 |
+
<dim>50000</dim>
|
| 414 |
</port>
|
| 415 |
<port id="10" precision="U8">
|
| 416 |
+
<dim>169950</dim>
|
| 417 |
</port>
|
| 418 |
<port id="11" precision="I32">
|
| 419 |
+
<dim>50000</dim>
|
| 420 |
</port>
|
| 421 |
<port id="12" precision="I32">
|
| 422 |
+
<dim>50000</dim>
|
| 423 |
</port>
|
| 424 |
<port id="13" precision="U8">
|
| 425 |
+
<dim>136851</dim>
|
| 426 |
</port>
|
| 427 |
<port id="14" precision="I32">
|
| 428 |
+
<dim>1608</dim>
|
| 429 |
</port>
|
| 430 |
<port id="15" precision="I32">
|
| 431 |
+
<dim>1608</dim>
|
| 432 |
</port>
|
| 433 |
<port id="16" precision="U8">
|
| 434 |
+
<dim>13723</dim>
|
| 435 |
</port>
|
| 436 |
<port id="17" precision="I32">
|
| 437 |
<dim>1608</dim>
|
|
|
|
| 449 |
</port>
|
| 450 |
</output>
|
| 451 |
</layer>
|
| 452 |
+
<layer id="38" name="Constant_24312" type="Const" version="opset1">
|
| 453 |
+
<data element_type="i32" shape="" offset="1902105" size="4" />
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 454 |
<output>
|
| 455 |
+
<port id="0" precision="I32" />
|
|
|
|
|
|
|
| 456 |
</output>
|
| 457 |
</layer>
|
| 458 |
+
<layer id="39" name="Constant_24314" type="Const" version="opset1">
|
| 459 |
+
<data element_type="u8" shape="5" offset="1902109" size="5" />
|
| 460 |
<output>
|
| 461 |
+
<port id="0" precision="U8">
|
| 462 |
+
<dim>5</dim>
|
| 463 |
+
</port>
|
| 464 |
</output>
|
| 465 |
</layer>
|
| 466 |
+
<layer id="40" name="Constant_24316" type="Const" version="opset1">
|
| 467 |
+
<data element_type="u8" shape="13" offset="1902114" size="13" />
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 468 |
<output>
|
| 469 |
+
<port id="0" precision="U8">
|
| 470 |
+
<dim>13</dim>
|
| 471 |
</port>
|
| 472 |
</output>
|
| 473 |
</layer>
|
| 474 |
+
<layer id="41" name="Truncate_24317" type="Truncate" version="extension">
|
| 475 |
+
<data m_num_inputs="1" />
|
| 476 |
<input>
|
| 477 |
<port id="0" precision="I32">
|
| 478 |
<dim>-1</dim>
|
|
|
|
| 480 |
<port id="1" precision="I32">
|
| 481 |
<dim>-1</dim>
|
| 482 |
</port>
|
| 483 |
+
<port id="2" precision="I32">
|
| 484 |
+
<dim>-1</dim>
|
| 485 |
+
</port>
|
| 486 |
+
<port id="3" precision="I32" />
|
| 487 |
+
<port id="4" precision="U8">
|
| 488 |
+
<dim>5</dim>
|
| 489 |
+
</port>
|
| 490 |
+
<port id="5" precision="U8">
|
| 491 |
+
<dim>13</dim>
|
| 492 |
+
</port>
|
| 493 |
</input>
|
| 494 |
<output>
|
| 495 |
+
<port id="6" precision="I32">
|
| 496 |
+
<dim>-1</dim>
|
| 497 |
+
</port>
|
| 498 |
+
<port id="7" precision="I32">
|
| 499 |
+
<dim>-1</dim>
|
| 500 |
+
</port>
|
| 501 |
+
<port id="8" precision="I32">
|
| 502 |
<dim>-1</dim>
|
| 503 |
</port>
|
| 504 |
</output>
|
| 505 |
</layer>
|
| 506 |
+
<layer id="42" name="Constant_24324" type="Const" version="opset1">
|
| 507 |
<data element_type="i32" shape="" offset="0" size="4" />
|
| 508 |
<output>
|
| 509 |
<port id="0" precision="I32" />
|
| 510 |
</output>
|
| 511 |
</layer>
|
| 512 |
+
<layer id="43" name="Constant_24325" type="Const" version="opset1">
|
| 513 |
<data element_type="i32" shape="" offset="4" size="4" />
|
| 514 |
<output>
|
| 515 |
<port id="0" precision="I32" />
|
| 516 |
</output>
|
| 517 |
</layer>
|
| 518 |
+
<layer id="44" name="Constant_24326" type="Const" version="opset1">
|
| 519 |
+
<data element_type="i32" shape="1" offset="1902127" size="4" />
|
| 520 |
<output>
|
| 521 |
<port id="0" precision="I32">
|
| 522 |
<dim>1</dim>
|
| 523 |
</port>
|
| 524 |
</output>
|
| 525 |
</layer>
|
| 526 |
+
<layer id="45" name="Constant_24327" type="Const" version="opset1">
|
| 527 |
+
<data element_type="i32" shape="4" offset="1902131" size="16" />
|
| 528 |
<output>
|
| 529 |
<port id="0" precision="I32">
|
| 530 |
<dim>4</dim>
|
| 531 |
</port>
|
| 532 |
</output>
|
| 533 |
</layer>
|
| 534 |
+
<layer id="46" name="CombineSegments_24328" type="CombineSegments" version="extension">
|
| 535 |
<input>
|
| 536 |
<port id="0" precision="I32" />
|
| 537 |
<port id="1" precision="I32" />
|
|
|
|
| 582 |
</port>
|
| 583 |
</output>
|
| 584 |
</layer>
|
| 585 |
+
<layer id="47" name="Subtract_24329" type="Subtract" version="opset1">
|
| 586 |
<data auto_broadcast="numpy" />
|
| 587 |
<input>
|
| 588 |
<port id="0" precision="I32">
|
|
|
|
| 598 |
</port>
|
| 599 |
</output>
|
| 600 |
</layer>
|
| 601 |
+
<layer id="48" name="Constant_24330" type="Const" version="opset1">
|
| 602 |
<data element_type="i32" shape="" offset="0" size="4" />
|
| 603 |
<output>
|
| 604 |
<port id="0" precision="I32" />
|
| 605 |
</output>
|
| 606 |
</layer>
|
| 607 |
+
<layer id="49" name="ReduceMax_24331" type="ReduceMax" version="opset1">
|
| 608 |
<data keep_dims="false" />
|
| 609 |
<input>
|
| 610 |
<port id="0" precision="I32">
|
|
|
|
| 616 |
<port id="2" precision="I32" />
|
| 617 |
</output>
|
| 618 |
</layer>
|
| 619 |
+
<layer id="50" name="Constant_24332" type="Const" version="opset1">
|
| 620 |
+
<data element_type="i32" shape="" offset="1902127" size="4" />
|
| 621 |
<output>
|
| 622 |
<port id="0" precision="I32" />
|
| 623 |
</output>
|
| 624 |
</layer>
|
| 625 |
+
<layer id="51" name="RaggedToDense_24333" type="RaggedToDense" version="extension">
|
| 626 |
+
<data pad_right="true" m_pad_max_length="false" />
|
| 627 |
<input>
|
| 628 |
<port id="0" precision="I32">
|
| 629 |
<dim>-1</dim>
|
|
|
|
| 648 |
</port>
|
| 649 |
</output>
|
| 650 |
</layer>
|
| 651 |
+
<layer id="52" name="Convert_24334" type="Convert" version="opset1">
|
| 652 |
<data destination_type="i32" />
|
| 653 |
<input>
|
| 654 |
<port id="0" precision="BOOL">
|
|
|
|
| 663 |
</port>
|
| 664 |
</output>
|
| 665 |
</layer>
|
| 666 |
+
<layer id="53" name="Convert_24334.0" type="Convert" version="opset1">
|
| 667 |
<data destination_type="i64" />
|
| 668 |
<input>
|
| 669 |
<port id="0" precision="I32">
|
|
|
|
| 678 |
</port>
|
| 679 |
</output>
|
| 680 |
</layer>
|
| 681 |
+
<layer id="55" name="RaggedToDense_24333.0" type="Convert" version="opset1">
|
| 682 |
<data destination_type="i64" />
|
| 683 |
<input>
|
| 684 |
<port id="0" precision="I32">
|
|
|
|
| 693 |
</port>
|
| 694 |
</output>
|
| 695 |
</layer>
|
| 696 |
+
<layer id="56" name="Result_24337" type="Result" version="opset1" output_names="input_ids">
|
| 697 |
<input>
|
| 698 |
<port id="0" precision="I64">
|
| 699 |
<dim>-1</dim>
|
|
|
|
| 701 |
</port>
|
| 702 |
</input>
|
| 703 |
</layer>
|
| 704 |
+
<layer id="54" name="Result_24339" type="Result" version="opset1" output_names="attention_mask">
|
| 705 |
<input>
|
| 706 |
<port id="0" precision="I64">
|
| 707 |
<dim>-1</dim>
|
|
|
|
| 712 |
</layers>
|
| 713 |
<edges>
|
| 714 |
<edge from-layer="0" from-port="0" to-layer="8" to-port="0" />
|
| 715 |
+
<edge from-layer="1" from-port="0" to-layer="46" to-port="0" />
|
| 716 |
+
<edge from-layer="2" from-port="0" to-layer="46" to-port="1" />
|
| 717 |
+
<edge from-layer="3" from-port="0" to-layer="46" to-port="2" />
|
| 718 |
+
<edge from-layer="4" from-port="0" to-layer="46" to-port="3" />
|
| 719 |
+
<edge from-layer="5" from-port="0" to-layer="46" to-port="4" />
|
| 720 |
+
<edge from-layer="6" from-port="0" to-layer="46" to-port="5" />
|
| 721 |
<edge from-layer="7" from-port="0" to-layer="14" to-port="0" />
|
| 722 |
<edge from-layer="8" from-port="1" to-layer="9" to-port="0" />
|
| 723 |
<edge from-layer="8" from-port="3" to-layer="21" to-port="4" />
|
|
|
|
| 726 |
<edge from-layer="9" from-port="1" to-layer="12" to-port="0" />
|
| 727 |
<edge from-layer="10" from-port="0" to-layer="12" to-port="1" />
|
| 728 |
<edge from-layer="11" from-port="0" to-layer="12" to-port="2" />
|
|
|
|
| 729 |
<edge from-layer="12" from-port="3" to-layer="17" to-port="0" />
|
| 730 |
+
<edge from-layer="12" from-port="3" to-layer="14" to-port="1" />
|
| 731 |
<edge from-layer="13" from-port="0" to-layer="14" to-port="2" />
|
| 732 |
<edge from-layer="14" from-port="3" to-layer="21" to-port="0" />
|
| 733 |
<edge from-layer="15" from-port="0" to-layer="19" to-port="0" />
|
|
|
|
| 736 |
<edge from-layer="18" from-port="0" to-layer="19" to-port="2" />
|
| 737 |
<edge from-layer="19" from-port="3" to-layer="21" to-port="1" />
|
| 738 |
<edge from-layer="20" from-port="0" to-layer="21" to-port="5" />
|
| 739 |
+
<edge from-layer="21" from-port="6" to-layer="23" to-port="0" />
|
| 740 |
<edge from-layer="21" from-port="7" to-layer="23" to-port="1" />
|
| 741 |
<edge from-layer="21" from-port="8" to-layer="23" to-port="2" />
|
| 742 |
<edge from-layer="21" from-port="9" to-layer="23" to-port="3" />
|
| 743 |
<edge from-layer="21" from-port="10" to-layer="23" to-port="4" />
|
| 744 |
<edge from-layer="21" from-port="11" to-layer="23" to-port="5" />
|
|
|
|
| 745 |
<edge from-layer="22" from-port="0" to-layer="23" to-port="6" />
|
| 746 |
+
<edge from-layer="23" from-port="9" to-layer="37" to-port="2" />
|
| 747 |
+
<edge from-layer="23" from-port="11" to-layer="37" to-port="4" />
|
| 748 |
+
<edge from-layer="23" from-port="10" to-layer="37" to-port="3" />
|
| 749 |
+
<edge from-layer="23" from-port="8" to-layer="37" to-port="1" />
|
| 750 |
+
<edge from-layer="23" from-port="7" to-layer="37" to-port="0" />
|
| 751 |
+
<edge from-layer="24" from-port="0" to-layer="37" to-port="5" />
|
| 752 |
+
<edge from-layer="25" from-port="0" to-layer="37" to-port="6" />
|
| 753 |
+
<edge from-layer="26" from-port="0" to-layer="37" to-port="7" />
|
| 754 |
+
<edge from-layer="27" from-port="0" to-layer="37" to-port="8" />
|
| 755 |
+
<edge from-layer="28" from-port="0" to-layer="37" to-port="9" />
|
| 756 |
+
<edge from-layer="29" from-port="0" to-layer="37" to-port="10" />
|
| 757 |
+
<edge from-layer="30" from-port="0" to-layer="37" to-port="11" />
|
| 758 |
+
<edge from-layer="31" from-port="0" to-layer="37" to-port="12" />
|
| 759 |
+
<edge from-layer="32" from-port="0" to-layer="37" to-port="13" />
|
| 760 |
+
<edge from-layer="33" from-port="0" to-layer="37" to-port="14" />
|
| 761 |
+
<edge from-layer="34" from-port="0" to-layer="37" to-port="15" />
|
| 762 |
+
<edge from-layer="35" from-port="0" to-layer="37" to-port="16" />
|
| 763 |
+
<edge from-layer="36" from-port="0" to-layer="37" to-port="17" />
|
| 764 |
+
<edge from-layer="37" from-port="19" to-layer="41" to-port="1" />
|
| 765 |
+
<edge from-layer="37" from-port="18" to-layer="41" to-port="0" />
|
| 766 |
+
<edge from-layer="37" from-port="20" to-layer="41" to-port="2" />
|
| 767 |
+
<edge from-layer="38" from-port="0" to-layer="41" to-port="3" />
|
| 768 |
+
<edge from-layer="39" from-port="0" to-layer="41" to-port="4" />
|
| 769 |
+
<edge from-layer="40" from-port="0" to-layer="41" to-port="5" />
|
| 770 |
+
<edge from-layer="41" from-port="6" to-layer="46" to-port="6" />
|
| 771 |
+
<edge from-layer="41" from-port="7" to-layer="46" to-port="7" />
|
| 772 |
+
<edge from-layer="41" from-port="8" to-layer="46" to-port="8" />
|
| 773 |
+
<edge from-layer="42" from-port="0" to-layer="46" to-port="9" />
|
| 774 |
+
<edge from-layer="43" from-port="0" to-layer="46" to-port="10" />
|
| 775 |
+
<edge from-layer="44" from-port="0" to-layer="46" to-port="11" />
|
| 776 |
+
<edge from-layer="45" from-port="0" to-layer="46" to-port="12" />
|
| 777 |
+
<edge from-layer="46" from-port="14" to-layer="47" to-port="0" />
|
| 778 |
+
<edge from-layer="46" from-port="13" to-layer="47" to-port="1" />
|
| 779 |
+
<edge from-layer="46" from-port="13" to-layer="51" to-port="0" />
|
| 780 |
+
<edge from-layer="46" from-port="14" to-layer="51" to-port="1" />
|
| 781 |
+
<edge from-layer="46" from-port="15" to-layer="51" to-port="2" />
|
| 782 |
+
<edge from-layer="47" from-port="2" to-layer="49" to-port="0" />
|
| 783 |
+
<edge from-layer="48" from-port="0" to-layer="49" to-port="1" />
|
| 784 |
+
<edge from-layer="49" from-port="2" to-layer="51" to-port="3" />
|
| 785 |
+
<edge from-layer="50" from-port="0" to-layer="51" to-port="4" />
|
| 786 |
+
<edge from-layer="51" from-port="6" to-layer="52" to-port="0" />
|
| 787 |
+
<edge from-layer="51" from-port="5" to-layer="55" to-port="0" />
|
| 788 |
+
<edge from-layer="52" from-port="1" to-layer="53" to-port="0" />
|
| 789 |
+
<edge from-layer="53" from-port="1" to-layer="54" to-port="0" />
|
| 790 |
+
<edge from-layer="55" from-port="1" to-layer="56" to-port="0" />
|
|
|
|
|
|
|
|
|
|
|
|
|
| 791 |
</edges>
|
| 792 |
<rt_info>
|
| 793 |
<add_attention_mask value="True" />
|
|
|
|
| 798 |
<detokenizer_input_type value="i64" />
|
| 799 |
<eos_token_id value="50257" />
|
| 800 |
<handle_special_tokens_with_re />
|
| 801 |
+
<max_length />
|
| 802 |
<number_of_inputs value="1" />
|
| 803 |
+
<openvino_tokenizers_version value="2025.2.0.0-565-130827ab189" />
|
| 804 |
+
<openvino_version value="2025.2.0-19120-87425bc78ca-releases/2025/2" />
|
| 805 |
+
<original_post_processor_template value="{"type": "TemplateProcessing", "single": [{"SpecialToken": {"id": "<|startoftranscript|>", "type_id": 0}}, {"SpecialToken": {"id": "<|notimestamps|>", "type_id": 0}}, {"Sequence": {"id": "A", "type_id": 0}}, {"SpecialToken": {"id": "<|endoftext|>", "type_id": 0}}], "pair": [{"SpecialToken": {"id": "<|startoftranscript|>", "type_id": 0}}, {"SpecialToken": {"id": "<|notimestamps|>", "type_id": 0}}, {"Sequence": {"id": "A", "type_id": 0}}, {"Sequence": {"id": "B", "type_id": 1}}, {"SpecialToken": {"id": "<|endoftext|>", "type_id": 1}}], "special_tokens": {"<|endoftext|>": {"id": "<|endoftext|>", "ids": [50257], "tokens": ["<|endoftext|>"]}, "<|notimestamps|>": {"id": "<|notimestamps|>", "ids": [50363], "tokens": ["<|notimestamps|>"]}, "<|startoftranscript|>": {"id": "<|startoftranscript|>", "ids": [50258], "tokens": ["<|startoftranscript|>"]}}}" />
|
| 806 |
<original_tokenizer_class value="<class 'transformers.models.whisper.tokenization_whisper_fast.WhisperTokenizerFast'>" />
|
| 807 |
<pad_token_id value="50257" />
|
| 808 |
+
<processed_post_processor_template value="{"single": {"ids": [50258, 50363, -1, 50257], "type_ids": [0, 0, 0, 0]}, "pair": {"ids": [50258, 50363, -1, -2, 50257], "type_ids": [0, 0, 0, 1, 1]}}" />
|
| 809 |
<skip_special_tokens value="True" />
|
| 810 |
<streaming_detokenizer value="False" />
|
|
|
|
| 811 |
<tokenizer_output_type value="i64" />
|
| 812 |
+
<tokenizers_version value="0.21.1" />
|
| 813 |
+
<transformers_version value="4.51.3" />
|
| 814 |
<use_max_padding value="False" />
|
| 815 |
<use_sentencepiece_backend value="False" />
|
| 816 |
+
<utf8_replace_mode value="replace" />
|
| 817 |
<with_detokenizer value="True" />
|
| 818 |
</rt_info>
|
| 819 |
</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",
|