././@PaxHeader
0000000
0000000
0000000
00000000034
00000000000
011452
0000000
0000000
28 mtime=1727720481.7657332
openai-whisper-20240930/
0000755
0001751
0000177
00000000000
00000000000
015517
00runner
docker
0000000
0000000
././@PaxHeader
0000000
0000000
0000000
00000000026
00000000000
011453
0000000
0000000
22 mtime=1727720469.0
openai-whisper-20240930/LICENSE
0000644
0001751
0000177
00000002047
00000000000
016527
00runner
docker
0000000
0000000
MIT License
Copyright (c) 2022 OpenAI
Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.
././@PaxHeader
0000000
0000000
0000000
00000000026
00000000000
011453
0000000
0000000
22 mtime=1727720469.0
openai-whisper-20240930/MANIFEST.in
0000644
0001751
0000177
00000000175
00000000000
017260
00runner
docker
0000000
0000000
include requirements.txt
include README.md
include LICENSE
include whisper/assets/*
include whisper/normalizers/english.json
././@PaxHeader
0000000
0000000
0000000
00000000034
00000000000
011452
0000000
0000000
28 mtime=1727720481.7657332
openai-whisper-20240930/PKG-INFO
0000644
0001751
0000177
00000022163
00000000000
016620
00runner
docker
0000000
0000000
Metadata-Version: 2.1
Name: openai-whisper
Version: 20240930
Summary: Robust Speech Recognition via Large-Scale Weak Supervision
Home-page: https://github.com/openai/whisper
Author: OpenAI
License: MIT
Description: # Whisper
[[Blog]](https://openai.com/blog/whisper)
[[Paper]](https://arxiv.org/abs/2212.04356)
[[Model card]](https://github.com/openai/whisper/blob/main/model-card.md)
[[Colab example]](https://colab.research.google.com/github/openai/whisper/blob/master/notebooks/LibriSpeech.ipynb)
Whisper is a general-purpose speech recognition model. It is trained on a large dataset of diverse audio and is also a multitasking model that can perform multilingual speech recognition, speech translation, and language identification.
## Approach
A Transformer sequence-to-sequence model is trained on various speech processing tasks, including multilingual speech recognition, speech translation, spoken language identification, and voice activity detection. These tasks are jointly represented as a sequence of tokens to be predicted by the decoder, allowing a single model to replace many stages of a traditional speech-processing pipeline. The multitask training format uses a set of special tokens that serve as task specifiers or classification targets.
## Setup
We used Python 3.9.9 and [PyTorch](https://pytorch.org/) 1.10.1 to train and test our models, but the codebase is expected to be compatible with Python 3.8-3.11 and recent PyTorch versions. The codebase also depends on a few Python packages, most notably [OpenAI's tiktoken](https://github.com/openai/tiktoken) for their fast tokenizer implementation. You can download and install (or update to) the latest release of Whisper with the following command:
pip install -U openai-whisper
Alternatively, the following command will pull and install the latest commit from this repository, along with its Python dependencies:
pip install git+https://github.com/openai/whisper.git
To update the package to the latest version of this repository, please run:
pip install --upgrade --no-deps --force-reinstall git+https://github.com/openai/whisper.git
It also requires the command-line tool [`ffmpeg`](https://ffmpeg.org/) to be installed on your system, which is available from most package managers:
```bash
# on Ubuntu or Debian
sudo apt update && sudo apt install ffmpeg
# on Arch Linux
sudo pacman -S ffmpeg
# on MacOS using Homebrew (https://brew.sh/)
brew install ffmpeg
# on Windows using Chocolatey (https://chocolatey.org/)
choco install ffmpeg
# on Windows using Scoop (https://scoop.sh/)
scoop install ffmpeg
```
You may need [`rust`](http://rust-lang.org) installed as well, in case [tiktoken](https://github.com/openai/tiktoken) does not provide a pre-built wheel for your platform. If you see installation errors during the `pip install` command above, please follow the [Getting started page](https://www.rust-lang.org/learn/get-started) to install Rust development environment. Additionally, you may need to configure the `PATH` environment variable, e.g. `export PATH="$HOME/.cargo/bin:$PATH"`. If the installation fails with `No module named 'setuptools_rust'`, you need to install `setuptools_rust`, e.g. by running:
```bash
pip install setuptools-rust
```
## Available models and languages
There are six model sizes, four with English-only versions, offering speed and accuracy tradeoffs.
Below are the names of the available models and their approximate memory requirements and inference speed relative to the large model.
The relative speeds below are measured by transcribing English speech on a A100, and the real-world speed may vary significantly depending on many factors including the language, the speaking speed, and the available hardware.
| Size | Parameters | English-only model | Multilingual model | Required VRAM | Relative speed |
|:------:|:----------:|:------------------:|:------------------:|:-------------:|:--------------:|
| tiny | 39 M | `tiny.en` | `tiny` | ~1 GB | ~10x |
| base | 74 M | `base.en` | `base` | ~1 GB | ~7x |
| small | 244 M | `small.en` | `small` | ~2 GB | ~4x |
| medium | 769 M | `medium.en` | `medium` | ~5 GB | ~2x |
| large | 1550 M | N/A | `large` | ~10 GB | 1x |
| turbo | 809 M | N/A | `turbo` | ~6 GB | ~8x |
The `.en` models for English-only applications tend to perform better, especially for the `tiny.en` and `base.en` models. We observed that the difference becomes less significant for the `small.en` and `medium.en` models.
Additionally, the `turbo` model is an optimized version of `large-v3` that offers faster transcription speed with a minimal degradation in accuracy.
Whisper's performance varies widely depending on the language. The figure below shows a performance breakdown of `large-v3` and `large-v2` models by language, using WERs (word error rates) or CER (character error rates, shown in *Italic*) evaluated on the Common Voice 15 and Fleurs datasets. Additional WER/CER metrics corresponding to the other models and datasets can be found in Appendix D.1, D.2, and D.4 of [the paper](https://arxiv.org/abs/2212.04356), as well as the BLEU (Bilingual Evaluation Understudy) scores for translation in Appendix D.3.
## Command-line usage
The following command will transcribe speech in audio files, using the `turbo` model:
whisper audio.flac audio.mp3 audio.wav --model turbo
The default setting (which selects the `small` model) works well for transcribing English. To transcribe an audio file containing non-English speech, you can specify the language using the `--language` option:
whisper japanese.wav --language Japanese
Adding `--task translate` will translate the speech into English:
whisper japanese.wav --language Japanese --task translate
Run the following to view all available options:
whisper --help
See [tokenizer.py](https://github.com/openai/whisper/blob/main/whisper/tokenizer.py) for the list of all available languages.
## Python usage
Transcription can also be performed within Python:
```python
import whisper
model = whisper.load_model("turbo")
result = model.transcribe("audio.mp3")
print(result["text"])
```
Internally, the `transcribe()` method reads the entire file and processes the audio with a sliding 30-second window, performing autoregressive sequence-to-sequence predictions on each window.
Below is an example usage of `whisper.detect_language()` and `whisper.decode()` which provide lower-level access to the model.
```python
import whisper
model = whisper.load_model("turbo")
# load audio and pad/trim it to fit 30 seconds
audio = whisper.load_audio("audio.mp3")
audio = whisper.pad_or_trim(audio)
# make log-Mel spectrogram and move to the same device as the model
mel = whisper.log_mel_spectrogram(audio).to(model.device)
# detect the spoken language
_, probs = model.detect_language(mel)
print(f"Detected language: {max(probs, key=probs.get)}")
# decode the audio
options = whisper.DecodingOptions()
result = whisper.decode(model, mel, options)
# print the recognized text
print(result.text)
```
## More examples
Please use the [
Show and tell](https://github.com/openai/whisper/discussions/categories/show-and-tell) category in Discussions for sharing more example usages of Whisper and third-party extensions such as web demos, integrations with other tools, ports for different platforms, etc.
## License
Whisper's code and model weights are released under the MIT License. See [LICENSE](https://github.com/openai/whisper/blob/main/LICENSE) for further details.
Platform: UNKNOWN
Requires-Python: >=3.8
Description-Content-Type: text/markdown
Provides-Extra: dev
././@PaxHeader
0000000
0000000
0000000
00000000026
00000000000
011453
0000000
0000000
22 mtime=1727720469.0
openai-whisper-20240930/README.md
0000644
0001751
0000177
00000017206
00000000000
017004
00runner
docker
0000000
0000000
# Whisper
[[Blog]](https://openai.com/blog/whisper)
[[Paper]](https://arxiv.org/abs/2212.04356)
[[Model card]](https://github.com/openai/whisper/blob/main/model-card.md)
[[Colab example]](https://colab.research.google.com/github/openai/whisper/blob/master/notebooks/LibriSpeech.ipynb)
Whisper is a general-purpose speech recognition model. It is trained on a large dataset of diverse audio and is also a multitasking model that can perform multilingual speech recognition, speech translation, and language identification.
## Approach
A Transformer sequence-to-sequence model is trained on various speech processing tasks, including multilingual speech recognition, speech translation, spoken language identification, and voice activity detection. These tasks are jointly represented as a sequence of tokens to be predicted by the decoder, allowing a single model to replace many stages of a traditional speech-processing pipeline. The multitask training format uses a set of special tokens that serve as task specifiers or classification targets.
## Setup
We used Python 3.9.9 and [PyTorch](https://pytorch.org/) 1.10.1 to train and test our models, but the codebase is expected to be compatible with Python 3.8-3.11 and recent PyTorch versions. The codebase also depends on a few Python packages, most notably [OpenAI's tiktoken](https://github.com/openai/tiktoken) for their fast tokenizer implementation. You can download and install (or update to) the latest release of Whisper with the following command:
pip install -U openai-whisper
Alternatively, the following command will pull and install the latest commit from this repository, along with its Python dependencies:
pip install git+https://github.com/openai/whisper.git
To update the package to the latest version of this repository, please run:
pip install --upgrade --no-deps --force-reinstall git+https://github.com/openai/whisper.git
It also requires the command-line tool [`ffmpeg`](https://ffmpeg.org/) to be installed on your system, which is available from most package managers:
```bash
# on Ubuntu or Debian
sudo apt update && sudo apt install ffmpeg
# on Arch Linux
sudo pacman -S ffmpeg
# on MacOS using Homebrew (https://brew.sh/)
brew install ffmpeg
# on Windows using Chocolatey (https://chocolatey.org/)
choco install ffmpeg
# on Windows using Scoop (https://scoop.sh/)
scoop install ffmpeg
You may need [`rust`](http://rust-lang.org) installed as well, in case [tiktoken](https://github.com/openai/tiktoken) does not provide a pre-built wheel for your platform. If you see installation errors during the `pip install` command above, please follow the [Getting started page](https://www.rust-lang.org/learn/get-started) to install Rust development environment. Additionally, you may need to configure the `PATH` environment variable, e.g. `export PATH="$HOME/.cargo/bin:$PATH"`. If the installation fails with `No module named 'setuptools_rust'`, you need to install `setuptools_rust`, e.g. by running:
```bash
pip install setuptools-rust
## Available models and languages
There are six model sizes, four with English-only versions, offering speed and accuracy tradeoffs.
Below are the names of the available models and their approximate memory requirements and inference speed relative to the large model.
The relative speeds below are measured by transcribing English speech on a A100, and the real-world speed may vary significantly depending on many factors including the language, the speaking speed, and the available hardware.
| Size | Parameters | English-only model | Multilingual model | Required VRAM | Relative speed |
|:------:|:----------:|:------------------:|:------------------:|:-------------:|:--------------:|
| tiny | 39 M | `tiny.en` | `tiny` | ~1 GB | ~10x |
| base | 74 M | `base.en` | `base` | ~1 GB | ~7x |
| small | 244 M | `small.en` | `small` | ~2 GB | ~4x |
| medium | 769 M | `medium.en` | `medium` | ~5 GB | ~2x |
| large | 1550 M | N/A | `large` | ~10 GB | 1x |
| turbo | 809 M | N/A | `turbo` | ~6 GB | ~8x |
The `.en` models for English-only applications tend to perform better, especially for the `tiny.en` and `base.en` models. We observed that the difference becomes less significant for the `small.en` and `medium.en` models.
Additionally, the `turbo` model is an optimized version of `large-v3` that offers faster transcription speed with a minimal degradation in accuracy.
Whisper's performance varies widely depending on the language. The figure below shows a performance breakdown of `large-v3` and `large-v2` models by language, using WERs (word error rates) or CER (character error rates, shown in *Italic*) evaluated on the Common Voice 15 and Fleurs datasets. Additional WER/CER metrics corresponding to the other models and datasets can be found in Appendix D.1, D.2, and D.4 of [the paper](https://arxiv.org/abs/2212.04356), as well as the BLEU (Bilingual Evaluation Understudy) scores for translation in Appendix D.3.
## Command-line usage
The following command will transcribe speech in audio files, using the `turbo` model:
whisper audio.flac audio.mp3 audio.wav --model turbo
The default setting (which selects the `small` model) works well for transcribing English. To transcribe an audio file containing non-English speech, you can specify the language using the `--language` option:
whisper japanese.wav --language Japanese
Adding `--task translate` will translate the speech into English:
whisper japanese.wav --language Japanese --task translate
Run the following to view all available options:
whisper --help
See [tokenizer.py](https://github.com/openai/whisper/blob/main/whisper/tokenizer.py) for the list of all available languages.
## Python usage
Transcription can also be performed within Python:
```python
import whisper
model = whisper.load_model("turbo")
result = model.transcribe("audio.mp3")
print(result["text"])
Internally, the `transcribe()` method reads the entire file and processes the audio with a sliding 30-second window, performing autoregressive sequence-to-sequence predictions on each window.
Below is an example usage of `whisper.detect_language()` and `whisper.decode()` which provide lower-level access to the model.
```python
import whisper
model = whisper.load_model("turbo")
# load audio and pad/trim it to fit 30 seconds
audio = whisper.load_audio("audio.mp3")
audio = whisper.pad_or_trim(audio)
# make log-Mel spectrogram and move to the same device as the model
mel = whisper.log_mel_spectrogram(audio).to(model.device)
# detect the spoken language
_, probs = model.detect_language(mel)
print(f"Detected language: {max(probs, key=probs.get)}")
# decode the audio
options = whisper.DecodingOptions()
result = whisper.decode(model, mel, options)
# print the recognized text
print(result.text)
## More examples
Please use the [
Show and tell](https://github.com/openai/whisper/discussions/categories/show-and-tell) category in Discussions for sharing more example usages of Whisper and third-party extensions such as web demos, integrations with other tools, ports for different platforms, etc.
## License
Whisper's code and model weights are released under the MIT License. See [LICENSE](https://github.com/openai/whisper/blob/main/LICENSE) for further details.
././@PaxHeader
0000000
0000000
0000000
00000000034
00000000000
011452
0000000
0000000
28 mtime=1727720481.7617333
openai-whisper-20240930/openai_whisper.egg-info/
0000755
0001751
0000177
00000000000
00000000000
022225
00runner
docker
0000000
0000000
././@PaxHeader
0000000
0000000
0000000
00000000026
00000000000
011453
0000000
0000000
22 mtime=1727720481.0
openai-whisper-20240930/openai_whisper.egg-info/PKG-INFO
0000644
0001751
0000177
00000022163
00000000000
023326
00runner
docker
0000000
0000000
Metadata-Version: 2.1
Name: openai-whisper
Version: 20240930
Summary: Robust Speech Recognition via Large-Scale Weak Supervision
Home-page: https://github.com/openai/whisper
Author: OpenAI
License: MIT
Description: # Whisper
[[Blog]](https://openai.com/blog/whisper)
[[Paper]](https://arxiv.org/abs/2212.04356)
[[Model card]](https://github.com/openai/whisper/blob/main/model-card.md)
[[Colab example]](https://colab.research.google.com/github/openai/whisper/blob/master/notebooks/LibriSpeech.ipynb)
Whisper is a general-purpose speech recognition model. It is trained on a large dataset of diverse audio and is also a multitasking model that can perform multilingual speech recognition, speech translation, and language identification.
## Approach
A Transformer sequence-to-sequence model is trained on various speech processing tasks, including multilingual speech recognition, speech translation, spoken language identification, and voice activity detection. These tasks are jointly represented as a sequence of tokens to be predicted by the decoder, allowing a single model to replace many stages of a traditional speech-processing pipeline. The multitask training format uses a set of special tokens that serve as task specifiers or classification targets.
## Setup
We used Python 3.9.9 and [PyTorch](https://pytorch.org/) 1.10.1 to train and test our models, but the codebase is expected to be compatible with Python 3.8-3.11 and recent PyTorch versions. The codebase also depends on a few Python packages, most notably [OpenAI's tiktoken](https://github.com/openai/tiktoken) for their fast tokenizer implementation. You can download and install (or update to) the latest release of Whisper with the following command:
pip install -U openai-whisper
Alternatively, the following command will pull and install the latest commit from this repository, along with its Python dependencies:
pip install git+https://github.com/openai/whisper.git
To update the package to the latest version of this repository, please run:
pip install --upgrade --no-deps --force-reinstall git+https://github.com/openai/whisper.git
It also requires the command-line tool [`ffmpeg`](https://ffmpeg.org/) to be installed on your system, which is available from most package managers:
```bash
# on Ubuntu or Debian
sudo apt update && sudo apt install ffmpeg
# on Arch Linux
sudo pacman -S ffmpeg
# on MacOS using Homebrew (https://brew.sh/)
brew install ffmpeg
# on Windows using Chocolatey (https://chocolatey.org/)
choco install ffmpeg
# on Windows using Scoop (https://scoop.sh/)
scoop install ffmpeg
```
You may need [`rust`](http://rust-lang.org) installed as well, in case [tiktoken](https://github.com/openai/tiktoken) does not provide a pre-built wheel for your platform. If you see installation errors during the `pip install` command above, please follow the [Getting started page](https://www.rust-lang.org/learn/get-started) to install Rust development environment. Additionally, you may need to configure the `PATH` environment variable, e.g. `export PATH="$HOME/.cargo/bin:$PATH"`. If the installation fails with `No module named 'setuptools_rust'`, you need to install `setuptools_rust`, e.g. by running:
```bash
pip install setuptools-rust
```
## Available models and languages
There are six model sizes, four with English-only versions, offering speed and accuracy tradeoffs.
Below are the names of the available models and their approximate memory requirements and inference speed relative to the large model.
The relative speeds below are measured by transcribing English speech on a A100, and the real-world speed may vary significantly depending on many factors including the language, the speaking speed, and the available hardware.
| Size | Parameters | English-only model | Multilingual model | Required VRAM | Relative speed |
|:------:|:----------:|:------------------:|:------------------:|:-------------:|:--------------:|
| tiny | 39 M | `tiny.en` | `tiny` | ~1 GB | ~10x |
| base | 74 M | `base.en` | `base` | ~1 GB | ~7x |
| small | 244 M | `small.en` | `small` | ~2 GB | ~4x |
| medium | 769 M | `medium.en` | `medium` | ~5 GB | ~2x |
| large | 1550 M | N/A | `large` | ~10 GB | 1x |
| turbo | 809 M | N/A | `turbo` | ~6 GB | ~8x |
The `.en` models for English-only applications tend to perform better, especially for the `tiny.en` and `base.en` models. We observed that the difference becomes less significant for the `small.en` and `medium.en` models.
Additionally, the `turbo` model is an optimized version of `large-v3` that offers faster transcription speed with a minimal degradation in accuracy.
Whisper's performance varies widely depending on the language. The figure below shows a performance breakdown of `large-v3` and `large-v2` models by language, using WERs (word error rates) or CER (character error rates, shown in *Italic*) evaluated on the Common Voice 15 and Fleurs datasets. Additional WER/CER metrics corresponding to the other models and datasets can be found in Appendix D.1, D.2, and D.4 of [the paper](https://arxiv.org/abs/2212.04356), as well as the BLEU (Bilingual Evaluation Understudy) scores for translation in Appendix D.3.
## Command-line usage
The following command will transcribe speech in audio files, using the `turbo` model:
whisper audio.flac audio.mp3 audio.wav --model turbo
The default setting (which selects the `small` model) works well for transcribing English. To transcribe an audio file containing non-English speech, you can specify the language using the `--language` option:
whisper japanese.wav --language Japanese
Adding `--task translate` will translate the speech into English:
whisper japanese.wav --language Japanese --task translate
Run the following to view all available options:
whisper --help
See [tokenizer.py](https://github.com/openai/whisper/blob/main/whisper/tokenizer.py) for the list of all available languages.
## Python usage
Transcription can also be performed within Python:
```python
import whisper
model = whisper.load_model("turbo")
result = model.transcribe("audio.mp3")
print(result["text"])
```
Internally, the `transcribe()` method reads the entire file and processes the audio with a sliding 30-second window, performing autoregressive sequence-to-sequence predictions on each window.
Below is an example usage of `whisper.detect_language()` and `whisper.decode()` which provide lower-level access to the model.
```python
import whisper
model = whisper.load_model("turbo")
# load audio and pad/trim it to fit 30 seconds
audio = whisper.load_audio("audio.mp3")
audio = whisper.pad_or_trim(audio)
# make log-Mel spectrogram and move to the same device as the model
mel = whisper.log_mel_spectrogram(audio).to(model.device)
# detect the spoken language
_, probs = model.detect_language(mel)
print(f"Detected language: {max(probs, key=probs.get)}")
# decode the audio
options = whisper.DecodingOptions()
result = whisper.decode(model, mel, options)
# print the recognized text
print(result.text)
```
## More examples
Please use the [
Show and tell](https://github.com/openai/whisper/discussions/categories/show-and-tell) category in Discussions for sharing more example usages of Whisper and third-party extensions such as web demos, integrations with other tools, ports for different platforms, etc.
## License
Whisper's code and model weights are released under the MIT License. See [LICENSE](https://github.com/openai/whisper/blob/main/LICENSE) for further details.
Platform: UNKNOWN
Requires-Python: >=3.8
Description-Content-Type: text/markdown
Provides-Extra: dev
././@PaxHeader
0000000
0000000
0000000
00000000026
00000000000
011453
0000000
0000000
22 mtime=1727720481.0
openai-whisper-20240930/openai_whisper.egg-info/SOURCES.txt
0000644
0001751
0000177
00000001337
00000000000
024115
00runner
docker
0000000
0000000
LICENSE
MANIFEST.in
README.md
pyproject.toml
requirements.txt
setup.py
openai_whisper.egg-info/PKG-INFO
openai_whisper.egg-info/SOURCES.txt
openai_whisper.egg-info/dependency_links.txt
openai_whisper.egg-info/entry_points.txt
openai_whisper.egg-info/requires.txt
openai_whisper.egg-info/top_level.txt
whisper/__init__.py
whisper/__main__.py
whisper/audio.py
whisper/decoding.py
whisper/model.py
whisper/timing.py
whisper/tokenizer.py
whisper/transcribe.py
whisper/triton_ops.py
whisper/utils.py
whisper/version.py
whisper/assets/gpt2.tiktoken
whisper/assets/mel_filters.npz
whisper/assets/multilingual.tiktoken
whisper/normalizers/__init__.py
whisper/normalizers/basic.py
whisper/normalizers/english.json
whisper/normalizers/english.py
././@PaxHeader
0000000
0000000
0000000
00000000026
00000000000
011453
0000000
0000000
22 mtime=1727720481.0
openai-whisper-20240930/openai_whisper.egg-info/dependency_links.txt
0000644
0001751
0000177
00000000001
00000000000
026273
00runner
docker
0000000
0000000
././@PaxHeader
0000000
0000000
0000000
00000000026
00000000000
011453
0000000
0000000
22 mtime=1727720481.0
openai-whisper-20240930/openai_whisper.egg-info/entry_points.txt
0000644
0001751
0000177
00000000064
00000000000
025523
00runner
docker
0000000
0000000
[console_scripts]
whisper = whisper.transcribe:cli
././@PaxHeader
0000000
0000000
0000000
00000000026
00000000000
011453
0000000
0000000
22 mtime=1727720481.0
openai-whisper-20240930/openai_whisper.egg-info/requires.txt
0000644
0001751
0000177
00000000275
00000000000
024631
00runner
docker
0000000
0000000
more-itertools
tiktoken
[:platform_machine == "x86_64" and sys_platform == "linux" or sys_platform == "linux2"]
triton>=2.0.0
pytest
flake8
././@PaxHeader
0000000
0000000
0000000
00000000026
00000000000
011453
0000000
0000000
22 mtime=1727720481.0
openai-whisper-20240930/openai_whisper.egg-info/top_level.txt
0000644
0001751
0000177
00000000010
00000000000
024746
00runner
docker
0000000
0000000
whisper
././@PaxHeader
0000000
0000000
0000000
00000000026
00000000000
011453
0000000
0000000
22 mtime=1727720469.0
openai-whisper-20240930/pyproject.toml
0000644
0001751
0000177
00000000163
00000000000
020433
00runner
docker
0000000
0000000
[tool.black]
[tool.isort]
profile = "black"
include_trailing_comma = true
line_length = 88
multi_line_output = 3
././@PaxHeader
0000000
0000000
0000000
00000000026
00000000000
011453
0000000
0000000
22 mtime=1727720469.0
openai-whisper-20240930/requirements.txt
0000644
0001751
0000177
00000000214
00000000000
021000
00runner
docker
0000000
0000000
more-itertools
tiktoken
triton>=2.0.0;platform_machine=="x86_64" and sys_platform=="linux" or sys_platform=="linux2"
././@PaxHeader
0000000
0000000
0000000
00000000034
00000000000
011452
0000000
0000000
28 mtime=1727720481.7657332
openai-whisper-20240930/setup.cfg
0000644
0001751
0000177
00000000046
00000000000
017340
00runner
docker
0000000
0000000
[egg_info]
tag_build =
tag_date = 0
././@PaxHeader
0000000
0000000
0000000
00000000026
00000000000
011453
0000000
0000000
22 mtime=1727720469.0
openai-whisper-20240930/setup.py
0000644
0001751
0000177
00000002335
00000000000
017234
00runner
docker
0000000
0000000
import platform
import sys
from pathlib import Path
import pkg_resources
from setuptools import find_packages, setup
def read_version(fname="whisper/version.py"):
exec(compile(open(fname, encoding="utf-8").read(), fname, "exec"))
return locals()["__version__"]
requirements = []
if sys.platform.startswith("linux") and platform.machine() == "x86_64":
requirements.append("triton>=2.0.0")
setup(
name="openai-whisper",
py_modules=["whisper"],
version=read_version(),
description="Robust Speech Recognition via Large-Scale Weak Supervision",
long_description=open("README.md", encoding="utf-8").read(),
long_description_content_type="text/markdown",
readme="README.md",
python_requires=">=3.8",
author="OpenAI",
url="https://github.com/openai/whisper",
license="MIT",
packages=find_packages(exclude=["tests*"]),
install_requires=[
str(r)
for r in pkg_resources.parse_requirements(
Path(__file__).with_name("requirements.txt").open()
)
],
entry_points={
"console_scripts": ["whisper=whisper.transcribe:cli"],
},
include_package_data=True,
extras_require={"dev": ["pytest", "scipy", "black", "flake8", "isort"]},
././@PaxHeader
0000000
0000000
0000000
00000000034
00000000000
011452
0000000
0000000
28 mtime=1727720481.7627332
openai-whisper-20240930/whisper/
0000755
0001751
0000177
00000000000
00000000000
017200
00runner
docker
0000000
0000000
././@PaxHeader
0000000
0000000
0000000
00000000026
00000000000
011453
0000000
0000000
22 mtime=1727720469.0
openai-whisper-20240930/whisper/__init__.py
0000644
0001751
0000177
00000016257
00000000000
021324
00runner
docker
0000000
0000000
import hashlib
import io
import os
import urllib
import warnings
from typing import List, Optional, Union
import torch
from tqdm import tqdm
from .audio import load_audio, log_mel_spectrogram, pad_or_trim
from .decoding import DecodingOptions, DecodingResult, decode, detect_language
from .model import ModelDimensions, Whisper
from .transcribe import transcribe
from .version import __version__
_MODELS = {
"tiny.en": "https://openaipublic.azureedge.net/main/whisper/models/d3dd57d32accea0b295c96e26691aa14d8822fac7d9d27d5dc00b4ca2826dd03/tiny.en.pt",
"tiny": "https://openaipublic.azureedge.net/main/whisper/models/65147644a518d12f04e32d6f3b26facc3f8dd46e5390956a9424a650c0ce22b9/tiny.pt",
"base.en": "https://openaipublic.azureedge.net/main/whisper/models/25a8566e1d0c1e2231d1c762132cd20e0f96a85d16145c3a00adf5d1ac670ead/base.en.pt",
"base": "https://openaipublic.azureedge.net/main/whisper/models/ed3a0b6b1c0edf879ad9b11b1af5a0e6ab5db9205f891f668f8b0e6c6326e34e/base.pt",
"small.en": "https://openaipublic.azureedge.net/main/whisper/models/f953ad0fd29cacd07d5a9eda5624af0f6bcf2258be67c92b79389873d91e0872/small.en.pt",
"small": "https://openaipublic.azureedge.net/main/whisper/models/9ecf779972d90ba49c06d968637d720dd632c55bbf19d441fb42bf17a411e794/small.pt",
"medium.en": "https://openaipublic.azureedge.net/main/whisper/models/d7440d1dc186f76616474e0ff0b3b6b879abc9d1a4926b7adfa41db2d497ab4f/medium.en.pt",
"medium": "https://openaipublic.azureedge.net/main/whisper/models/345ae4da62f9b3d59415adc60127b97c714f32e89e936602e85993674d08dcb1/medium.pt",
"large-v1": "https://openaipublic.azureedge.net/main/whisper/models/e4b87e7e0bf463eb8e6956e646f1e277e901512310def2c24bf0e11bd3c28e9a/large-v1.pt",
"large-v2": "https://openaipublic.azureedge.net/main/whisper/models/81f7c96c852ee8fc832187b0132e569d6c3065a3252ed18e56effd0b6a73e524/large-v2.pt",
"large-v3": "https://openaipublic.azureedge.net/main/whisper/models/e5b1a55b89c1367dacf97e3e19bfd829a01529dbfdeefa8caeb59b3f1b81dadb/large-v3.pt",
"large": "https://openaipublic.azureedge.net/main/whisper/models/e5b1a55b89c1367dacf97e3e19bfd829a01529dbfdeefa8caeb59b3f1b81dadb/large-v3.pt",
"large-v3-turbo": "https://openaipublic.azureedge.net/main/whisper/models/aff26ae408abcba5fbf8813c21e62b0941638c5f6eebfb145be0c9839262a19a/large-v3-turbo.pt",
"turbo": "https://openaipublic.azureedge.net/main/whisper/models/aff26ae408abcba5fbf8813c21e62b0941638c5f6eebfb145be0c9839262a19a/large-v3-turbo.pt",
# base85-encoded (n_layers, n_heads) boolean arrays indicating the cross-attention heads that are
# highly correlated to the word-level timing, i.e. the alignment between audio and text tokens.
_ALIGNMENT_HEADS = {
"tiny.en": b"ABzY8J1N>@0{>%R00Bk>$p{7v037`oCl~+#00",
"tiny": b"ABzY8bu8Lr0{>%RKn9Fp%m@SkK7Kt=7ytkO",
"base.en": b"ABzY8;40c<0{>%RzzG;p*o+Vo09|#PsxSZm00",
"base": b"ABzY8KQ!870{>%RzyTQH3`Q^yNP!>##QT-<FaQ7m",
"small.en": b"ABzY8>?_)10{>%RpeA61k&I|OI3I$65C{;;pbCHh0B{qLQ;+}v00",
"small": b"ABzY8DmU6=0{>%Rpa?J`kvJ6qF(V^F86#Xh7JUGMK}P<N0000",
"medium.en": b"ABzY8usPae0{>%R7<zz_OvQ{)4kMa0BMw6u5rT}kRKX;$NfYBv00*Hl@qhsU00",
"medium": b"ABzY8B0Jh+0{>%R7}kK1fFL7w6%<-Pf*t^=N)Qr&0RR9",
"large-v1": b"ABzY8r9j$a0{>%R7#4sLmoOs{s)o3~84-RPdcFk!JR<kSfC2yj",
"large-v2": b"ABzY8zd+h!0{>%R7=D0pU<_bnWW*tkYAhobTNnu$jnkEkXqp)j;w1Tzk)UH3X%SZd&fFZ2fC2yj",
"large-v3": b"ABzY8gWO1E0{>%R7(9S+Kn!D~%ngiGaR?*L!iJG9p-nab0JQ=-{D1-g00",
"large": b"ABzY8gWO1E0{>%R7(9S+Kn!D~%ngiGaR?*L!iJG9p-nab0JQ=-{D1-g00",
"large-v3-turbo": b"ABzY8j^C+e0{>%RARaKHP%t(lGR*)0g!tONPyhe`",
"turbo": b"ABzY8j^C+e0{>%RARaKHP%t(lGR*)0g!tONPyhe`",
def _download(url: str, root: str, in_memory: bool) -> Union[bytes, str]:
os.makedirs(root, exist_ok=True)
expected_sha256 = url.split("/")[-2]
download_target = os.path.join(root, os.path.basename(url))
if os.path.exists(download_target) and not os.path.isfile(download_target):
raise RuntimeError(f"{download_target} exists and is not a regular file")
if os.path.isfile(download_target):
with open(download_target, "rb") as f:
model_bytes = f.read()
if hashlib.sha256(model_bytes).hexdigest() == expected_sha256:
return model_bytes if in_memory else download_target
else:
warnings.warn(
f"{download_target} exists, but the SHA256 checksum does not match; re-downloading the file"
)
with urllib.request.urlopen(url) as source, open(download_target, "wb") as output:
with tqdm(
total=int(source.info().get("Content-Length")),
ncols=80,
unit="iB",
unit_scale=True,
unit_divisor=1024,
) as loop:
while True:
buffer = source.read(8192)
if not buffer:
break
output.write(buffer)
loop.update(len(buffer))
model_bytes = open(download_target, "rb").read()
if hashlib.sha256(model_bytes).hexdigest() != expected_sha256:
raise RuntimeError(
"Model has been downloaded but the SHA256 checksum does not not match. Please retry loading the model."
)
return model_bytes if in_memory else download_target
def available_models() -> List[str]:
"""Returns the names of available models"""
return list(_MODELS.keys())
def load_model(
name: str,
device: Optional[Union[str, torch.device]] = None,
download_root: str = None,
in_memory: bool = False,
) -> Whisper:
"""
Load a Whisper ASR model
Parameters
----------
name : str
one of the official model names listed by `whisper.available_models()`, or
path to a model checkpoint containing the model dimensions and the model state_dict.
device : Union[str, torch.device]
the PyTorch device to put the model into
download_root: str
path to download the model files; by default, it uses "~/.cache/whisper"
in_memory: bool
whether to preload the model weights into host memory
Returns
-------
model : Whisper
The Whisper ASR model instance
"""
if device is None:
device = "cuda" if torch.cuda.is_available() else "cpu"
if download_root is None:
default = os.path.join(os.path.expanduser("~"), ".cache")
download_root = os.path.join(os.getenv("XDG_CACHE_HOME", default), "whisper")
if name in _MODELS:
checkpoint_file = _download(_MODELS[name], download_root, in_memory)
alignment_heads = _ALIGNMENT_HEADS[name]
elif os.path.isfile(name):
checkpoint_file = open(name, "rb").read() if in_memory else name
alignment_heads = None
else:
raise RuntimeError(
f"Model {name} not found; available models = {available_models()}"
)
with (
io.BytesIO(checkpoint_file) if in_memory else open(checkpoint_file, "rb")
) as fp:
checkpoint = torch.load(fp, map_location=device)
del checkpoint_file
dims = ModelDimensions(**checkpoint["dims"])
model = Whisper(dims)
model.load_state_dict(checkpoint["model_state_dict"])
if alignment_heads is not None:
model.set_alignment_heads(alignment_heads)
return model.to(device)
././@PaxHeader
0000000
0000000
0000000
00000000026
00000000000
011453
0000000
0000000
22 mtime=1727720469.0
openai-whisper-20240930/whisper/__main__.py
0000644
0001751
0000177
00000000043
00000000000
021267
00runner
docker
0000000
0000000
from .transcribe import cli
././@PaxHeader
0000000
0000000
0000000
00000000034
00000000000
011452
0000000
0000000
28 mtime=1727720481.7637334
openai-whisper-20240930/whisper/assets/
0000755
0001751
0000177
00000000000
00000000000
020502
00runner
docker
0000000
0000000
././@PaxHeader
0000000
0000000
0000000
00000000026
00000000000
011453
0000000
0000000
22 mtime=1727720469.0
openai-whisper-20240930/whisper/assets/gpt2.tiktoken
0000644
0001751
0000177
00003137742
00000000000
023151
00runner
docker
0000000
0000000
IQ== 0
Ig== 1
Iw== 2
JA== 3
JQ== 4
Jg== 5
Jw== 6
KA== 7
KQ== 8
Kg== 9
Kw== 10
LA== 11
LQ== 12
Lg== 13
Lw== 14
MA== 15
MQ== 16
Mg== 17
Mw== 18
NA== 19
NQ== 20
Ng== 21
Nw== 22
OA== 23
OQ== 24
Og== 25
Ow== 26
PA== 27
PQ== 28
Pg== 29
Pw== 30
QA== 31
QQ== 32
Qg== 33
Qw== 34
RA== 35
RQ== 36
Rg== 37
Rw== 38
SA== 39
SQ== 40
Sg== 41
Sw== 42
TA== 43
TQ== 44
Tg== 45
Tw== 46
UA== 47
UQ== 48
Ug== 49
Uw== 50
VA== 51
VQ== 52
Vg== 53
Vw== 54
WA== 55
WQ== 56
Wg== 57
Ww== 58
XA== 59
XQ== 60
Xg== 61
Xw== 62
YA== 63
YQ== 64
Yg== 65
Yw== 66
ZA== 67
ZQ== 68
Zg== 69
Zw== 70
aA== 71
aQ== 72
ag== 73
aw== 74
bA== 75
bQ== 76
bg== 77
bw== 78
cA== 79
cQ== 80
cg== 81
cw== 82
dA== 83
dQ== 84
dg== 85
dw== 86
eA== 87
eQ== 88
eg== 89
ew== 90
fA== 91
fQ== 92
fg== 93
oQ== 94
og== 95
ow== 96
pA== 97
pQ== 98
pg== 99
pw== 100
qA== 101
qQ== 102
qg== 103
qw== 104
rA== 105
rg== 106
rw== 107
sA== 108
sQ== 109
sg== 110
sw== 111
tA== 112
tQ== 113
tg== 114
tw== 115
uA== 116
uQ== 117
ug== 118
uw== 119
vA== 120
vQ== 121
vg== 122
vw== 123
wA== 124
wQ== 125
wg== 126
ww== 127
xA== 128
xQ== 129
xg== 130
xw== 131
yA== 132
yQ== 133
yg== 134
yw== 135
zA== 136
zQ== 137
zg== 138
zw== 139
0A== 140
0Q== 141
0g== 142
0w== 143
1A== 144
1Q== 145
1g== 146
1w== 147
2A== 148
2Q== 149
2g== 150
2w== 151
3A== 152
3Q== 153
3g== 154
3w== 155
4A== 156
4Q== 157
4g== 158
4w== 159
5A== 160
5Q== 161
5g== 162
5w== 163
6A== 164
6Q== 165
6g== 166
6w== 167
7A== 168
7Q== 169
7g== 170
7w== 171
8A== 172
8Q== 173
8g== 174
8w== 175
9A== 176
9Q== 177
9g== 178
9w== 179
+A== 180
+Q== 181
+g== 182
+w== 183
/A== 184
/Q== 185
/g== 186
/w== 187
AA== 188
AQ== 189
Ag== 190
Aw== 191
BA== 192
BQ== 193
Bg== 194
Bw== 195
CA== 196
CQ== 197
Cg== 198
Cw== 199
DA== 200
DQ== 201
Dg== 202
Dw== 203
EA== 204
EQ== 205
Eg== 206
Ew== 207
FA== 208
FQ== 209
Fg== 210
Fw== 211
GA== 212
GQ== 213
Gg== 214
Gw== 215
HA== 216
HQ== 217
Hg== 218
Hw== 219
IA== 220
fw== 221
gA== 222
gQ== 223
gg== 224
gw== 225
hA== 226
hQ== 227
hg== 228
hw== 229
iA== 230
iQ== 231
ig== 232
iw== 233
jA== 234
jQ== 235
jg== 236
jw== 237
kA== 238
kQ== 239
kg== 240
kw== 241
lA== 242
lQ== 243
lg== 244
lw== 245
mA== 246
mQ== 247
mg== 248
mw== 249
nA== 250
nQ== 251
ng== 252
nw== 253
oA== 254
rQ== 255
IHQ= 256
IGE= 257
aGU= 258
aW4= 259
cmU= 260
b24= 261
IHRoZQ== 262
ZXI= 263
IHM= 264
YXQ= 265
IHc= 266
IG8= 267
ZW4= 268
IGM= 269
aXQ= 270
aXM= 271
YW4= 272
b3I= 273
ZXM= 274
IGI= 275
ZWQ= 276
IGY= 277
aW5n 278
IHA= 279
b3U= 280
IGFu 281
YWw= 282
YXI= 283
IHRv 284
IG0= 285
IG9m 286
IGlu 287
IGQ= 288
IGg= 289
IGFuZA== 290
aWM= 291
YXM= 292
bGU= 293
IHRo 294
aW9u 295
b20= 296
bGw= 297
ZW50 298
IG4= 299
IGw= 300
c3Q= 301
IHJl 302
dmU= 303
IGU= 304
cm8= 305
bHk= 306
IGJl 307
IGc= 308
IFQ= 309
Y3Q= 310
IFM= 311
aWQ= 312
b3Q= 313
IEk= 314
dXQ= 315
ZXQ= 316
IEE= 317
IGlz 318
IG9u 319
aW0= 320
YW0= 321
b3c= 322
YXk= 323
YWQ= 324
c2U= 325
IHRoYXQ= 326
IEM= 327
aWc= 328
IGZvcg== 329
YWM= 330
IHk= 331
dmVy 332
dXI= 333
IHU= 334
bGQ= 335
IHN0 336
IE0= 337
J3M= 338
IGhl 339
IGl0 340
YXRpb24= 341
aXRo 342
aXI= 343
Y2U= 344
IHlvdQ== 345
aWw= 346
IEI= 347
IHdo 348
b2w= 349
IFA= 350
IHdpdGg= 351
IDE= 352
dGVy 353
Y2g= 354
IGFz 355
IHdl 356
ICg= 357
bmQ= 358
aWxs 359
IEQ= 360
aWY= 361
IDI= 362
YWc= 363
ZXJz 364
a2U= 365
ICI= 366
IEg= 367
ZW0= 368
IGNvbg== 369
IFc= 370
IFI= 371
aGVy 372
IHdhcw== 373
IHI= 374
b2Q= 375
IEY= 376
dWw= 377
YXRl 378
IGF0 379
cmk= 380
cHA= 381
b3Jl 382
IFRoZQ== 383
IHNl 384
dXM= 385
IHBybw== 386
IGhh 387
dW0= 388
IGFyZQ== 389
IGRl 390
YWlu 391
YW5k 392
IG9y 393
aWdo 394
ZXN0 395
aXN0 396
YWI= 397
cm9t 398
IE4= 399
dGg= 400
IGNvbQ== 401
IEc= 402
dW4= 403
b3A= 404
MDA= 405
IEw= 406
IG5vdA== 407
ZXNz 408
IGV4 409
IHY= 410
cmVz 411
IEU= 412
ZXc= 413
aXR5 414
YW50 415
IGJ5 416
ZWw= 417
b3M= 418
b3J0 419
b2M= 420
cXU= 421
IGZyb20= 422
IGhhdmU= 423
IHN1 424
aXZl 425
b3VsZA== 426
IHNo 427
IHRoaXM= 428
bnQ= 429
cmE= 430
cGU= 431
aWdodA== 432
YXJ0 433
bWVudA== 434
IGFs 435
dXN0 436
ZW5k 437
LS0= 438
YWxs 439
IE8= 440
YWNr 441
IGNo 442
IGxl 443
aWVz 444
cmVk 445
YXJk 446
4oA= 447
b3V0 448
IEo= 449
IGFi 450
ZWFy 451
aXY= 452
YWxseQ== 453
b3Vy 454
b3N0 455
Z2g= 456
cHQ= 457
IHBs 458
YXN0 459
IGNhbg== 460
YWs= 461
b21l 462
dWQ= 463
VGhl 464
IGhpcw== 465
IGRv 466
IGdv 467
IGhhcw== 468
Z2U= 469
J3Q= 470
IFU= 471
cm91 472
IHNh 473
IGo= 474
IGJ1dA== 475
IHdvcg== 476
IGFsbA== 477
ZWN0 478
IGs= 479
YW1l 480
IHdpbGw= 481
b2s= 482
IHdoZQ== 483
IHRoZXk= 484
aWRl 485
MDE= 486
ZmY= 487
aWNo 488
cGw= 489
dGhlcg== 490
IHRy 491
Li4= 492
IGludA== 493
aWU= 494
dXJl 495
YWdl 496
IG5l 497
aWFs 498
YXA= 499
aW5l 500
aWNl 501
IG1l 502
IG91dA== 503
YW5z 504
b25l 505
b25n 506
aW9ucw== 507
IHdobw== 508
IEs= 509
IHVw 510
IHRoZWly 511
IGFk 512
IDM= 513
IHVz 514
YXRlZA== 515
b3Vz 516
IG1vcmU= 517
dWU= 518
b2c= 519
IFN0 520
aW5k 521
aWtl 522
IHNv 523
aW1l 524
cGVy 525
LiI= 526
YmVy 527
aXo= 528
YWN0 529
IG9uZQ== 530
IHNhaWQ= 531
IC0= 532
YXJl 533
IHlvdXI= 534
Y2M= 535
IFRo 536
IGNs 537
ZXA= 538
YWtl 539
YWJsZQ== 540
aXA= 541
IGNvbnQ= 542
IHdoaWNo 543
aWE= 544
IGlt 545
IGFib3V0 546
IHdlcmU= 547
dmVyeQ== 548
dWI= 549
IGhhZA== 550
IGVu 551
IGNvbXA= 552
LCI= 553
IElu 554
IHVu 555
IGFn 556
aXJl 557
YWNl 558
YXU= 559
YXJ5 560
IHdvdWxk 561
YXNz 562
cnk= 563
IOKA 564
Y2w= 565
b29r 566
ZXJl 567
c28= 568
IFY= 569
aWdu 570
aWI= 571
IG9mZg== 572
IHRl 573
dmVu 574
IFk= 575
aWxl 576
b3Nl 577
aXRl 578
b3Jt 579
IDIwMQ== 580
IHJlcw== 581
IG1hbg== 582
IHBlcg== 583
IG90aGVy 584
b3Jk 585
dWx0 586
IGJlZW4= 587
IGxpa2U= 588
YXNl 589
YW5jZQ== 590
a3M= 591
YXlz 592
b3du 593
ZW5jZQ== 594
IGRpcw== 595
Y3Rpb24= 596
IGFueQ== 597
IGFwcA== 598
IHNw 599
aW50 600
cmVzcw== 601
YXRpb25z 602
YWls 603
IDQ= 604
aWNhbA== 605
IHRoZW0= 606
IGhlcg== 607
b3VudA== 608
IENo 609
IGFy 610
IGlm 611
IHRoZXJl 612
IHBl 613
IHllYXI= 614
YXY= 615
IG15 616
IHNvbWU= 617
IHdoZW4= 618
b3VnaA== 619
YWNo 620
IHRoYW4= 621
cnU= 622
b25k 623
aWNr 624
IG92ZXI= 625
dmVs 626
IHF1 627
Cgo= 628
IHNj 629
cmVhdA== 630
cmVl 631
IEl0 632
b3VuZA== 633
cG9ydA== 634
IGFsc28= 635
IHBhcnQ= 636
ZnRlcg== 637
IGtu 638
IGJlYw== 639
IHRpbWU= 640
ZW5z 641
IDU= 642
b3BsZQ== 643
IHdoYXQ= 644
IG5v 645
ZHU= 646
bWVy 647
YW5n 648
IG5ldw== 649
LS0tLQ== 650
IGdldA== 651
b3J5 652
aXRpb24= 653
aW5ncw== 654
IGp1c3Q= 655
IGludG8= 656
IDA= 657
ZW50cw== 658
b3Zl 659
dGU= 660
IHBlb3BsZQ== 661
IHByZQ== 662
IGl0cw== 663
IHJlYw== 664
IHR3 665
aWFu 666
aXJzdA== 667
YXJr 668
b3Jz 669
IHdvcms= 670
YWRl 671
b2I= 672
IHNoZQ== 673
IG91cg== 674
d24= 675
aW5r 676
bGlj 677
IDE5 678
IEhl 679
aXNo 680
bmRlcg== 681
YXVzZQ== 682
IGhpbQ== 683
b25z 684
IFs= 685
IHJv 686
Zm9ybQ== 687
aWxk 688
YXRlcw== 689
dmVycw== 690
IG9ubHk= 691
b2xs 692
IHNwZQ== 693
Y2s= 694
ZWxs 695
YW1w 696
IGFjYw== 697
IGJs 698
aW91cw== 699
dXJu 700
ZnQ= 701
b29k 702
IGhvdw== 703
aGVk 704
ICc= 705
IGFmdGVy 706
YXc= 707
IGF0dA== 708
b3Y= 709
bmU= 710
IHBsYXk= 711
ZXJ2 712
aWN0 713
IGNvdWxk 714
aXR0 715
IGFt 716
IGZpcnN0 717
IDY= 718
IGFjdA== 719
ICQ= 720
ZWM= 721
aGluZw== 722
dWFs 723
dWxs 724
IGNvbW0= 725
b3k= 726
b2xk 727
Y2Vz 728
YXRlcg== 729
IGZl 730
IGJldA== 731
d2U= 732
aWZm 733
IHR3bw== 734
b2Nr 735
IGJhY2s= 736
KS4= 737
aWRlbnQ= 738
IHVuZGVy 739
cm91Z2g= 740
c2Vs 741
eHQ= 742
IG1heQ== 743
cm91bmQ= 744
IHBv 745
cGg= 746
aXNz 747
IGRlcw== 748
IG1vc3Q= 749
IGRpZA== 750
IGFkZA== 751
amVjdA== 752
IGluYw== 753
Zm9yZQ== 754
IHBvbA== 755
b250 756
IGFnYWlu 757
Y2x1ZA== 758
dGVybg== 759
IGtub3c= 760
IG5lZWQ= 761
IGNvbnM= 762
IGNv 763
IC4= 764
IHdhbnQ= 765
IHNlZQ== 766
IDc= 767
bmluZw== 768
aWV3 769
IFRoaXM= 770
Y2Vk 771
IGV2ZW4= 772
IGluZA== 773
dHk= 774
IFdl 775
YXRo 776
IHRoZXNl 777
IHBy 778
IHVzZQ== 779
IGJlY2F1c2U= 780
IGZs 781
bmc= 782
IG5vdw== 783
IOKAkw== 784
Y29t 785
aXNl 786
IG1ha2U= 787
IHRoZW4= 788
b3dlcg== 789
IGV2ZXJ5 790
IFVu 791
IHNlYw== 792
b3Nz 793
dWNo 794
IGVt 795
ID0= 796
IFJl 797
aWVk 798
cml0 799
IGludg== 800
bGVjdA== 801
IHN1cHA= 802
YXRpbmc= 803
IGxvb2s= 804
bWFu 805
cGVjdA== 806
IDg= 807
cm93 808
IGJ1 809
IHdoZXJl 810
aWZpYw== 811
IHllYXJz 812
aWx5 813
IGRpZmY= 814
IHNob3VsZA== 815
IHJlbQ== 816
VGg= 817
SW4= 818
IGV2 819
ZGF5 820
J3Jl 821
cmli 822
IHJlbA== 823
c3M= 824
IGRlZg== 825
IHJpZ2h0 826
IHN5 827
KSw= 828
bGVz 829
MDAw 830
aGVu 831
IHRocm91Z2g= 832
IFRy 833
X18= 834
IHdheQ== 835
IGRvbg== 836
ICw= 837
IDEw 838
YXNlZA== 839
IGFzcw== 840
dWJsaWM= 841
IHJlZw== 842
IEFuZA== 843
aXg= 844
IHZlcnk= 845
IGluY2x1ZA== 846
b3RoZXI= 847
IGltcA== 848
b3Ro 849
IHN1Yg== 850
IOKAlA== 851
IGJlaW5n 852
YXJn 853
IFdo 854
PT0= 855
aWJsZQ== 856
IGRvZXM= 857
YW5nZQ== 858
cmFt 859
IDk= 860
ZXJ0 861
cHM= 862
aXRlZA== 863
YXRpb25hbA== 864
IGJy 865
IGRvd24= 866
IG1hbnk= 867
YWtpbmc= 868
IGNhbGw= 869
dXJpbmc= 870
aXRpZXM= 871
IHBo 872
aWNz 873
YWxz 874
IGRlYw== 875
YXRpdmU= 876
ZW5lcg== 877
IGJlZm9yZQ== 878
aWxpdHk= 879
IHdlbGw= 880
IG11Y2g= 881
ZXJzb24= 882
IHRob3Nl 883
IHN1Y2g= 884
IGtl 885
IGVuZA== 886
IEJ1dA== 887
YXNvbg== 888
dGluZw== 889
IGxvbmc= 890
ZWY= 891
IHRoaW5r 892
eXM= 893
IGJlbA== 894
IHNt 895
aXRz 896
YXg= 897
IG93bg== 898
IHByb3Y= 899
IHNldA== 900
aWZl 901
bWVudHM= 902
Ymxl 903
d2FyZA== 904
IHNob3c= 905
IHByZXM= 906
bXM= 907
b21ldA== 908
IG9i 909
IHNheQ== 910
IFNo 911
dHM= 912
ZnVs 913
IGVmZg== 914
IGd1 915
IGluc3Q= 916
dW5k 917
cmVu 918
Y2Vzcw== 919
IGVudA== 920
IFlvdQ== 921
IGdvb2Q= 922
IHN0YXJ0 923
aW5jZQ== 924
IG1hZGU= 925
dHQ= 926
c3RlbQ== 927
b2xvZw== 928
dXA= 929
IHw= 930
dW1w 931
IGhlbA== 932
dmVybg== 933
dWxhcg== 934
dWFsbHk= 935
IGFj 936
IG1vbg== 937
IGxhc3Q= 938
IDIwMA== 939
MTA= 940
IHN0dWQ= 941
dXJlcw== 942
IEFy 943
c2VsZg== 944
YXJz 945
bWVyaWM= 946
dWVz 947
Y3k= 948
IG1pbg== 949
b2xsb3c= 950
IGNvbA== 951
