Free self-hosted STT for VRChat. https://git.yummers.dev/TaSTT.git/atom?h=v0.3 2023-01-04T18:03:39+00:00 2023-01-04T18:03:39+00:00 yum yum.food.vr@gmail.com 2023-01-04T17:52:02+00:00 urn:sha1:66d311b3267620995e5c35b16f3fba18ed0c48f3 Previously, paths containing spaces would be interpreted by python's argument parser as multiple separate arguments, causing it to fail. Now we escape paths inside PythonWrapper using std::quoted(). * Improve PII filtering. Python output would contain multiple path separators (like C:\\Users\\foo\\), defeating the PII regex. * Silence compiler warning in PII filter. * Document usability improvements. * Transcription layer exponential backoff goes to ~infinity when paused. This is a hack, since we really don't need to transcribe at all when paused, but it lets us keep the code simple. Good enough until the next rewrite. * Shader only samples background when necessary. * Limit matchStrings() print()s to DEBUG mode 2023-01-01T23:06:57+00:00 yum yum.food.vr@gmail.com 2023-01-01T23:06:15+00:00 urn:sha1:edc199f99cb7eaa9d4972dd40a8aafd1ebbf7c90 This seems like the best way to support users. 2022-12-25T22:22:36+00:00 yum yum.food.vr@gmail.com 2022-12-21T23:54:33+00:00 urn:sha1:4177ff5acb6fb583339f90374f31892144f3354b Add a new shader to make the box a little prettier. * Reduce material slots required from 2 to 1 * Add rounding to edge of box 2022-12-22T22:54:03+00:00 yum yum.food.vr@gmail.com 2022-12-22T22:49:44+00:00 urn:sha1:6f2c1dace46a68620bc61a732a2f43252bd5d3ba By sending encoded words rather than letters, we could speed up English paging rate by 2.5x over an optimized implementation Word-encoded implementation: 16 bits per word (capped at 64k possible words). Optimized char-based imlementation: (5.7 chars per word) * (7 bits per char) == 39.9 bits per word 2.5x slower than word encoding. Today's char-based implementation: (5.7 chars per word) * (16 bits per char) == 91.2 bits per word 5.7x slower than word encoding. 2022-12-22T21:47:40+00:00 yum yum.food.vr@gmail.com 2022-12-22T21:47:40+00:00 urn:sha1:9c6b3c43ddecca315cdba11ebd14f6926527fa8b 2022-12-01T21:39:49+00:00 yum yum.food.vr@gmail.com 2022-12-01T21:39:49+00:00 urn:sha1:f9bcee6acf6f8a418d2f96368941fc05e4b72252 Also decrease sync params & add a few more emotes. 2022-11-15T05:36:13+00:00 yum yum.food.vr@gmail.com 2022-11-15T05:30:50+00:00 urn:sha1:2505a5cc486cd913db50a475e45c3701b9710282 After re-reading the paper, I noticed that they apply a couple optimizations I wasn't using. Use the top-level `whisper.transcribe` method, which is a little slower, but more accurate than the one I was using. Although this method is slower, it has better temporal stability due to the increased quality, which I think should make for an overall more responsive UX. Lower transcription quality means the paging layer has to waste time updating earlier cells. Also, drop the auto-commit stuff and go back to string stitching. I think it's better to let the user manually commit. A rework of the hand controls is probably coming soon. Finally, update README. 2022-11-07T05:22:54+00:00 yum yum.food.vr@gmail.com 2022-11-07T05:22:54+00:00 urn:sha1:629c0f611de1622131bb0fa364c170219f6252ed 2022-11-06T20:50:38+00:00 yum yum.food.vr@gmail.com 2022-11-06T20:50:38+00:00 urn:sha1:7146acb9d4ad751fc5ced411a2990d0aad17d08f Add a `matchStrings` which does basically the same thing as `matchStringList` except it doesn't split the input at space boundaries. I think this should work better for Japanese and Chinese, since they don't use spaces. Doesn't seem to cause any accuracy regressions for English. Also update the README. 2022-11-05T19:14:12+00:00 yum yum.food.vr@gmail.com 2022-11-05T19:11:15+00:00 urn:sha1:7f930340e3aa94d6aca120d28436594427469373 Instead of generating one animation for every single character in our character set, we just generate 2: the lowest and the highest. We use blend trees to interpolate between these two extremes. This reduces the number of animations we have to generate by a factor of 80. It also clears the way for multi-language support (coming soon). It also means we don't have to reopen unity every time we generate a new animator.