TaSTT.git/Scripts, branch v0.11.1Free self-hosted STT for VRChat.https://git.yummers.dev/TaSTT.git/atom?h=v0.11.12023-04-25T21:27:23+00:00Restore string matching, remove affinity mask2023-04-25T21:27:23+00:00yumyum.food.vr@gmail.com2023-04-25T21:27:23+00:00urn:sha1:8534d17ab00de729867bb482ab204854127ed2a1
Affinity mask no longer affects performance. String matching is still
needed for temporal stability in fast-paced long-form transcription
tasks.
~Finish integrating faster-whisper2023-04-25T01:14:22+00:00yumyum.food.vr@gmail.com2023-04-25T01:08:16+00:00urn:sha1:89c4fa29e5810fac7c75cb5edda1565320a5b8a9
I'm able to use the new code to show text in game. Not yet play-tested.
Begin integrating faster-whisper2023-04-24T03:52:36+00:00yumyum.food.vr@gmail.com2023-04-24T03:52:36+00:00urn:sha1:b4bb6524652e0f76834ca26a4afa232855ca1348
This is a much faster, lower-VRAM reimplementation of Whisper in Python.
Early testing is extremely promising: fast transcription speed,
extremely low resource usage (CPU/RAM/VRAM), high accuracy.
Reduce texture memory usage for English speakers2023-03-22T00:13:32+00:00yumyum.food.vr@gmail.com2023-03-22T00:13:32+00:00urn:sha1:1f15133dd985442af20d42a96fbcd0007f03bd2b
We used to populate 7 4k textures + 1 2k texture for all users.
Now if the user has configured `bytes_per_char=1` in the Unity
panel, we just populate a single 512x512 texture containing the
first 128 ASCII characters.
This reduces texture memory usage by 99.74%, from 134.67 MB to
340 KB.
Fix _socket module not found issue2023-03-21T22:02:29+00:00yumyum.food.vr@gmail.com2023-03-21T21:28:46+00:00urn:sha1:656d7c2092545b18d981acfac000c73fb2128e4a
Need python310._pth, specifically 'import site' line, for
embedded python + pip to get along.
Set PYTHONPATH in synchronous multiprocessing layer2023-03-08T23:36:46+00:00yumyum.food.vr@gmail.com2023-03-08T23:36:46+00:00urn:sha1:12b6447a87da8077c7dd12b92eefc27dcf7f0818
A user saw an error like `ModuleNotFoundError: No module named _socket`.
StackOverflow blames this on PYTHONPATH, so let's try setting it.
* Fix latent bug in Scripts/transcribe.py. PyAudio.open() positional
parameters must be specified in correct order, even when telling it
which parameter is which. *shrug*
Bugfix: C++ transcription engine should not launch OSC layer2023-02-27T00:24:10+00:00yumyum.food.vr@gmail.com2023-02-27T00:24:10+00:00urn:sha1:d0d5eedb4e6c56d81ae2135a50212f2091ee65d7
Not ready yet.
Begin work on C++ custom chatbox2023-02-26T22:27:22+00:00yumyum.food.vr@gmail.com2023-02-26T22:21:18+00:00urn:sha1:f7d7858a9ff270380f5407e48d6afaf6a3a97de3
Sort of a misnomer. The idea is to use C++ for transcription and Python
for steamvr and OSC.
Having issues getting output from multithreaded Python code. Not in the
mood to figure this out today.
* Hide unimplemented parts of C++ panel.
Revert "Apply previous window conditioning to decoding layer"2023-02-23T05:54:39+00:00yumyum.food.vr@gmail.com2023-02-23T05:49:58+00:00urn:sha1:718319ee7b79d7cdbead5d765769b50c25e968f4
This reverts commit cece1ee8f1b985c2a89adb661dd02c6d44787f67.
This does *not* in fact result in improved temporal stability. It makes
makes things so unstable that even single-sentence messages fail to
ever stabilize.
Begin work on C++ implementation2023-02-23T05:49:29+00:00yumyum.food.vr@gmail.com2023-02-21T21:19:43+00:00urn:sha1:9a97fbc3c583ccd518d838faaaa36ed9aa5558e1
Use Const-me/Whisper to perform transcription. This implementation is
vastly more efficient: CPU usage, memory usage, and VRAM usage are all
dramatically reduced. It's slightly less accurate when comparing the
same model (due to the lack of beam search decoding), but since you can
use larger models, the impact is largely a wash.