Whisper Master, Please Add Support for Hugging Face Models

[lipolipo111]

Whisper Master, Please Add Support for Hugging Face Models

[lipolipo111]

Do I have to generate the tokenizer.json or can I use one that already exist?

I dunno, but tokenizer.json is same in all default models.

Just wanted to say a big thanks! Actually I download the file tokenizer.json from the model guillaumekln/faster-whisper-large-v2 and it works.

Really appreciate your help. You’re awesome!

[Purfview]

Custom models are already supported.

[lipolipo111]

Custom models are already supported.

Sure, I can help with that. Here's your text with corrected grammar, spelling, and coherence:

Custom models are already supported.

Oh, I see. The problem was that I don't know how to do it. When I tried using --model HuggingFaceModel, it didn't recognize it and showed me a list of default models. So, how should I go about it?

[Purfview]

When I tried using --model HuggingFaceModel, it didn't recognize it and showed me a list of default models. So, how should I go about it?

Have it in folder like this: faster-whisper-HuggingFaceModel

[lipolipo111]

When I tried using --model HuggingFaceModel, it didn't recognize it and showed me a list of default models. So, how should I go about it?

Have it in folder like this: faster-whisper-HuggingFaceModel

Hi,

First of all, I want to express my gratitude for all the support you've provided. It's been really helpful.

However, I'm still encountering an issue. Specifically, when I run the following command:

whisper-faster.exe "J:\Convertion\MyTestJapaneseAudio.wav" --language ja --model large-v2-japanese-5k-steps

I’ve downloaded the respective model and moved it into the _models folder as shown below:

C:\Users\lipo111\AppData\Roaming\Whisper\Purfview-Whisper-Faster_models\faster-whisper-large-v2-japanese-5k-steps

which contains the following three files:

1. config.json
2. model.bin
3. vocabulary.json

Despite this, I’m receiving an error message that the model cannot be found at the specified location:

whisper-faster.exe "J:\Convertion\MyTestJapaneseAudio.wav" --language ja --model large-v2-japanese-5k-steps

Unknown model not found at: C:\Users\lipo111\AppData\RoamingWhisper\Purfview-Whisper-Faster\_models\faster-whisper-large-v2-japanese-5k-steps

['tiny', 'tiny.en', 'base', 'base.en', 'small', 'small.en', 'medium', 'medium.en', 'large', 'large-v1', 'large-v2']

I’d appreciate any further assistance you could provide to resolve this issue.

Thanks again for your support.

[Purfview]

which contains the following three files:

1. config.json

2. model.bin

3. vocabulary.json

Copy there tokenizer.json too.

[lipolipo111]

which contains the following three files:

1. config.json

2. model.bin

3. vocabulary.json

Copy there tokenizer.json too.

Actually there is no tokenizer.json in the Hugging Face model, thats why I did't copy. The model is

zh-plus/faster-whisper-large-v2-japanese-5k-steps

Do I have to generate the tokenizer.json or can I use one that already exist?

[Purfview]

Do I have to generate the tokenizer.json or can I use one that already exist?

I dunno, but tokenizer.json is same in all default models.