Skip to content
This repository has been archived by the owner on May 1, 2023. It is now read-only.

Discard too old buffers #6

Open
shirayu opened this issue Sep 24, 2022 · 1 comment
Open

Discard too old buffers #6

shirayu opened this issue Sep 24, 2022 · 1 comment
Labels
Type: Feature New Feature

Comments

@shirayu
Copy link
Owner

shirayu commented Sep 24, 2022 •
edited
Loading

Currently, the buffer_tokens and buffer_mel is being used again in a later process.
However, when they become too large, the processing will take too long and exhaust VRAM.
Therefore, it is necessary to discard old buffers in them at the appropriate time.
@shirayu shirayu added the Type: Feature New Feature label Sep 24, 2022
@shirayu shirayu changed the title Discard too old buffer Discard too old buffers Sep 24, 2022
@shirayu
Copy link
Owner Author

shirayu commented Sep 25, 2022

--condition_on_previous_text is added in the original Whisper

openai/whisper@5d8d3e7

That is related to here.

https://github.com/shirayu/whisper_streaming/blob/11a895606af52953a73e9acd903875d317eba6f5/whisper_streaming/transcriber.py#L223-L226

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Assignees
No one assigned
Labels
Type: Feature New Feature
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
1 participant
@shirayu