From e5d308a4d4d584e7e56995b17f4c2af3df675a4f Mon Sep 17 00:00:00 2001 From: Ed Summers Date: Tue, 15 Oct 2024 14:32:16 -0400 Subject: [PATCH] Add new Turbo model See: https://github.com/openai/whisper/discussions/2363 Also, don't redundantly download large-v3.pt. --- README.md | 4 ++-- whisper_models/urls.txt | 2 +- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/README.md b/README.md index 38483cd..cbb5617 100644 --- a/README.md +++ b/README.md @@ -56,10 +56,10 @@ docker run --rm --tty --volume .:/app --env-file .env sul-speech-to-text --creat ### Run the Job -Now you can run the container and have it pick up the job you placed into the queue: +Now you can run the container and have it pick up the job you placed into the queue. You can drop the `--gpus all` if you don't have a GPU. ```shell -docker run --rm --tty --env-file .env sul-speech-to-text --no-daemon +docker run --rm --tty --env-file .env --gpus all sul-speech-to-text --no-daemon ``` Wait for the results to appear: diff --git a/whisper_models/urls.txt b/whisper_models/urls.txt index d7e90af..4bd79d3 100644 --- a/whisper_models/urls.txt +++ b/whisper_models/urls.txt @@ -9,4 +9,4 @@ https://openaipublic.azureedge.net/main/whisper/models/345ae4da62f9b3d59415adc60 https://openaipublic.azureedge.net/main/whisper/models/e4b87e7e0bf463eb8e6956e646f1e277e901512310def2c24bf0e11bd3c28e9a/large-v1.pt https://openaipublic.azureedge.net/main/whisper/models/81f7c96c852ee8fc832187b0132e569d6c3065a3252ed18e56effd0b6a73e524/large-v2.pt https://openaipublic.azureedge.net/main/whisper/models/e5b1a55b89c1367dacf97e3e19bfd829a01529dbfdeefa8caeb59b3f1b81dadb/large-v3.pt -https://openaipublic.azureedge.net/main/whisper/models/e5b1a55b89c1367dacf97e3e19bfd829a01529dbfdeefa8caeb59b3f1b81dadb/large-v3.pt +https://openaipublic.azureedge.net/main/whisper/models/aff26ae408abcba5fbf8813c21e62b0941638c5f6eebfb145be0c9839262a19a/large-v3-turbo.pt