This article explains how to manually download the model from the Hugging Face website if the automatic download fails when using the faster-whisper[local] speech recognition channel. If you don't want to download manually, click here to view alternative solutions, including proxy settings, dedicated download tools, and downloading model archives from GitHub.

When using the faster-whisper[local] speech recognition channel, the model needs to be downloaded from the external site https://huggingface.co or the domestic mirror https://hf-mirror.com. The former requires a stable and reliable proxy tool for access, otherwise it may fail frequently. The latter mirror site itself seems to have unstable network connections, often resulting in download timeouts and failures.

Step 1: First Attempt a Speech Recognition to Let the Software Generate the Required Directory Structure

Select Audio/Video to Subtitles on the left, upload a test audio file, choose the desired model, and start transcription.
Wait until it fails, then confirm that the corresponding model directory structure has been generated in the software directory/models folder.

The folder names for each model are as follows:

• tiny.en: models--Systran--faster-whisper-tiny.en
• tiny: models--Systran--faster-whisper-tiny
• base.en: models--Systran--faster-whisper-base.en
• base: models--Systran--faster-whisper-base
• small.en: models--Systran--faster-whisper-small.en
• small: models--Systran--faster-whisper-small
• medium.en: models--Systran--faster-whisper-medium.en
• medium: models--Systran--faster-whisper-medium
• large-v1: models--Systran--faster-whisper-large-v1
• large-v2: models--Systran--faster-whisper-large-v2
• large-v3: models--Systran--faster-whisper-large-v3
• large: models--Systran--faster-whisper-large-v3
• distil-large-v2: models--Systran--faster-distil-whisper-large-v2
• distil-medium.en: models--Systran--faster-distil-whisper-medium.en
• distil-small.en: models--Systran--faster-distil-whisper-small.en
• distil-large-v3: models--Systran--faster-distil-whisper-large-v3
• large-v3-turbo: models--mobiuslabsgmbh--faster-whisper-large-v3-turbo
• turbo: models--mobiuslabsgmbh--faster-whisper-large-v3-turbo

Step 2: Delete the Incomplete Downloaded Files

Note: Only delete the files, do not delete the folders.

Then, enter the model folder. Taking large-v3-turbo as an example:

After entering software directory/models/models--mobiuslabsgmbh--faster-whisper-large-v3-turbo, you will see three subfolders: snapshots, blobs, and refs.

1. Enter the blobs subfolder and delete all downloaded files.

2. Go back and enter the snapshots folder. You will see a folder with a very long name composed of numbers and letters. Enter it, and you will find approximately 4-5 files, possibly all with a size of 0. Delete all of them. Wait in this folder; the manually downloaded model files will be copied directly here later.

Step 3: Manually Download the Model from the Repository

Go to the model's address on huggingface.co, manually download the model files, and place them in the snapshots\xxxxxxxxxxx folder.

For example, the download address for large-v3-turbo is https://huggingface.co/mobiuslabsgmbh/faster-whisper-large-v3-turbo/tree/main.

After opening the URL, click the download icon next to each file to download it directly in your browser. After downloading, copy the files to snapshots\xxxxxxxxxx.

Then, when you attempt speech recognition again, it will prompt Model already exists.

Download Addresses for Each Model

• tiny.en: https://huggingface.co/Systran/faster-whisper-tiny.en/tree/main
• tiny: https://huggingface.co/Systran/faster-whisper-tiny/tree/main
• base.en: https://huggingface.co/Systran/faster-whisper-base.en/tree/main
• base: https://huggingface.co/Systran/faster-whisper-base/tree/main
• small.en: https://huggingface.co/Systran/faster-whisper-small.en/tree/main
• small: https://huggingface.co/Systran/faster-whisper-small/tree/main
• medium.en: https://huggingface.co/Systran/faster-whisper-medium.en/tree/main
• medium: https://huggingface.co/Systran/faster-whisper-medium/tree/main
• large-v1: https://huggingface.co/Systran/faster-whisper-large-v1/tree/main
• large-v2: https://huggingface.co/Systran/faster-whisper-large-v2/tree/main
• large-v3: https://huggingface.co/Systran/faster-whisper-large-v3/tree/main
• large: https://huggingface.co/Systran/faster-whisper-large-v3/tree/main
• distil-large-v2: https://huggingface.co/Systran/faster-distil-whisper-large-v2/tree/main
• distil-medium.en: https://huggingface.co/Systran/faster-distil-whisper-medium.en/tree/main
• distil-small.en: https://huggingface.co/Systran/faster-distil-whisper-small.en/tree/main
• distil-large-v3: https://huggingface.co/Systran/faster-distil-whisper-large-v3/tree/main
• large-v3-turbo: https://huggingface.co/mobiuslabsgmbh/faster-whisper-large-v3-turbo/tree/main
• turbo: https://huggingface.co/mobiuslabsgmbh/faster-whisper-large-v3-turbo/tree/main