19bbff72ec
Remove more spurious logging
2022-07-14 23:49:03 -04:00
3d0871306c
Temporary fix for virtual env pytorch
2022-07-14 23:40:56 -04:00
840a97e70e
Espeak phonemizer pipeline for spanish (and also english)
2022-07-14 22:14:32 -03:00
2e6be74403
change health check endpoint response to match other services
2022-07-14 20:50:02 -04:00
4524389ced
Remove logger.flush()
2022-07-14 20:29:56 -04:00
0cc56be156
Merge branch 'master' of github.com:storytold/storyteller-ml
2022-07-14 16:58:08 -04:00
89fed52526
Fix logger flush
2022-07-14 16:57:43 -04:00
7fe0699f86
Merge pull request #3 from storytold/echelon-run-tests
...
Script to run tests (eg. in CI)
2022-07-14 16:46:14 -04:00
dfa6354a4e
update docker image to fix RTX3090 issue (needs a lot of cleanup)
2022-07-14 16:35:24 -04:00
54e63f5b71
Switch to using logger instead of print()
2022-07-14 16:32:23 -04:00
d562007578
Fix script
2022-07-10 18:08:41 -04:00
e7fc2ccf31
Add script to launch workers on-prem
2022-07-10 18:06:18 -04:00
9b8298c679
Merge pull request #4 from storytold/echelon-fix-nvidia-revoked-key
...
Fix dockerfile broken by Nvidia's revoked key
2022-07-06 03:39:43 -04:00
5b215270ac
Fix dockerfile broken by Nvidia's revoked key
2022-07-06 03:38:58 -04:00
ed7e26cfd2
fix missing pipeline param (forgot to plumb through)
2022-07-06 02:56:20 -04:00
ea6dce774b
Script to run tests (eg. in CI)
...
We'll incrementally add more tests as we progress. Not everything will
need testing.
2022-07-06 02:55:04 -04:00
9ce1b5061d
Merge pull request #2 from storytold/echelon-text-pipeline-http-arg
...
Add HTTP endpoint argument for which text pipeline to use
2022-07-06 02:27:28 -04:00
b4a24ff561
Add HTTP endpoint argument for which text pipeline to use
2022-07-06 02:26:35 -04:00
551b2670bb
Merge pull request #1 from storytold/echelon-text-pipeline-system
...
Entirely reworked text handling pipeline
2022-07-05 03:50:41 -04:00
76c6dbf12b
Entirely reworked text handling pipeline
...
This breaks the text pipeline into the following pieces:
- input parsing (split known bracketed phoneme sections out first)
- normalization (fix spacing, numbers, abbreviations, etc.)
- phonetization (do CMUdict lookup or g2p prediction)
- sequencing (convert the symbolic representation to integer
representation)
Each "pipeline" can have a unique combination of operations for each of
these pipeline steps.
This will grow over time.
2022-07-05 03:45:28 -04:00
6a43c8d7ca
Add link to TTS
2022-06-18 13:13:03 -04:00
83af51d982
Add TTS readme detailing HTTP parameters
2022-06-18 13:10:01 -04:00
792e437b79
Change name of repository
2022-06-18 12:50:30 -04:00
711d5adb73
Add architecture debugging
2022-06-10 17:09:33 -04:00
44f2a1082b
update readme (to trigger build)
2022-03-31 16:10:05 -04:00
197f13fa5e
Dockerfile and Github build
...
These have a lot of cruft, but it should work. I'll switch over to using
this from the python code in the Rust monorepo
2022-03-30 23:47:53 -04:00
dd342ddec6
warning
2022-02-28 05:55:57 -05:00
e4267f607b
add architecture diagrams
2022-02-28 05:53:41 -05:00
b50205847c
readme/notes
2022-02-28 05:46:37 -05:00
ba326bfcfd
Introduce code from Rust monorepo. This is a mess
2022-02-28 05:43:54 -05:00
921fb49f9a
Initialize repo
2022-02-28 04:51:03 -05:00
Brandon Thomas