kdavis-mozilla
f0dbdf7855
Renamed optimizer
2020-03-17 09:56:33 +01:00
kdavis-mozilla
561131a05c
Added optimizer for lm_alpha + lm_beta
2020-02-26 11:41:22 +01:00
lissyx
84ac39769c
Merge pull request #2787 from lissyx/gradle-android-prebuilt
...
Cache gradle deps and Android emulator setup
2020-02-26 09:29:31 +01:00
Alexandre Lissy
c5ded5adfe
Cache gradle deps and Android emulator setup
2020-02-25 19:17:13 +01:00
Reuben Morais
377f0bc4b8
Merge pull request #2789 from reuben/issue2786
...
Make const functions receive const ModelState pointers (Fixes #2786 )
2020-02-25 11:57:24 +01:00
Reuben Morais
b74738a405
Make const functions receive const ModelState pointers
2020-02-25 11:15:45 +01:00
Reuben Morais
1f1f5a98e4
Merge pull request #2781 from rhamnett/patch-2
...
Add flag to force reinitialisation of learning rate after lr_plateau
2020-02-24 16:08:13 +01:00
lissyx
b3b357cdac
Merge pull request #2776 from lissyx/pyenv-prebuilt
...
Produce pyenv ready-to-use
2020-02-24 13:09:33 +01:00
lissyx
4059da2869
Merge pull request #2785 from imskr/sk-fix
...
added ctc decoder builds tasks
2020-02-24 12:34:21 +01:00
Shubham Kumar
b03d7fe4ee
added against dependencies
2020-02-24 17:01:32 +05:30
Alexandre Lissy
e9f530f7c7
Make webrtcvad really optional
2020-02-24 12:08:12 +01:00
Alexandre Lissy
bd5044fe31
Ensure proper python ABI
2020-02-24 12:02:30 +01:00
Alexandre Lissy
8029f3d7dd
Produce pyenv ready-to-use
2020-02-24 12:01:36 +01:00
Shubham Kumar
279dc947f1
added ctc decoder builds tasks
2020-02-23 15:01:57 +05:30
Richard Hamnett
a3268545ab
Update flags.py
...
change flag datatype to boolean
2020-02-21 19:24:13 +00:00
Reuben Morais
46e7993075
Merge pull request #2771 from reuben/warn-sample-rate
...
Warn if --audio_sample_rate does not match training sample
2020-02-21 19:46:05 +01:00
Richard Hamnett
5e1f54ae4f
Reset learning rate if force set
2020-02-21 18:33:43 +00:00
Richard Hamnett
0de9e4bf80
Add force_initialize_learning_rate
...
Ability to reset learning rate which has been reduced by reduce_lr_on_plateau
2020-02-21 18:32:03 +00:00
Reuben Morais
aff310d73a
Merge pull request #2780 from rhamnett/patch-1
...
Fix transcribe.py - use new checkpoint load method
2020-02-21 18:04:24 +01:00
Reuben Morais
f264134a61
Merge pull request #2778 from mozilla/ftyers-patch-1
...
Create BIBLIOGRAPHY.md
2020-02-21 18:01:07 +01:00
Francis Tyers
965927d91b
Update BIBLIOGRAPHY.md
2020-02-21 16:59:23 +00:00
Francis Tyers
ba21d4434b
Update BIBLIOGRAPHY.md
2020-02-21 16:56:25 +00:00
Richard Hamnett
e101cb8cc5
Fix transcribe.py - use new checkpoint load method
...
Replaced non existing try_loading() method with the saver method and respect load flag
Removed tf.train.Saver()
2020-02-21 15:56:58 +00:00
Francis Tyers
943f19e1d5
Create BIBLIOGRAPHY.md
2020-02-20 23:12:46 +00:00
Reuben Morais
4291db7309
Handle graph without learning rate variable for export case
2020-02-20 15:40:35 +01:00
lissyx
234a64c6ea
Merge pull request #2766 from lissyx/homebrew-prebuilt-caches
...
Generate one-time Homebrew tarball
2020-02-20 12:47:05 +01:00
Alexandre Lissy
7d1663b1c5
Generate one-time Homebrew tarball
2020-02-20 11:46:43 +01:00
lissyx
fc39433f9b
Merge pull request #2775 from lissyx/remove-irc
...
Remove IRC notifications
2020-02-20 10:47:33 +01:00
Alexandre Lissy
2c69273a49
Remove IRC notifications
2020-02-20 10:09:11 +01:00
Reuben Morais
536b821d24
Merge pull request #2772 from reuben/error-code-hex
...
Report error code as hexadecimal numbers for easier lookup
2020-02-19 15:19:21 +01:00
Reuben Morais
f47c7f8421
Report error code as hexadecimal numbers for easier lookup
2020-02-19 14:11:01 +01:00
Reuben Morais
0b82c751db
Merge pull request #2770 from reuben/lr_reduction_rebased
...
Reduce learning rate on plateau
2020-02-18 22:20:54 +01:00
Reuben Morais
1178215423
Warn if --audio_sample_rate does not match training sample
...
In PR #2688 , we started specifying the upper frequency limit when computing Mfccs.
This value was computed as half of the --audio_sample_rate value. Despite accepting
a variable sample rate input for the Mfcc computation, the TensorFlow OP only takes
a constant upper frequency limit, so we can't pass a dynamic value computed from each
sample to the op.
This means we lost the ability to transparently train on data with multiple sample
rates. This commit adds a warning message in case a training sample does not match
the --audio_sample_rate flag.
2020-02-18 18:15:01 +01:00
Reuben Morais
559042a218
Increase epoch count in train tests to guarantee outputs in 8kHz mode
2020-02-18 18:14:15 +01:00
Reuben Morais
78e8dfdf38
Disable early stopping and LR reduction on plateau by default
2020-02-18 16:18:37 +01:00
Reuben Morais
6e12b7caed
Allow missing learning rate variable in older checkpoints
2020-02-18 16:18:31 +01:00
Daniel
17ddc5600e
Reduce learning rate on plateau.
2020-02-18 16:17:51 +01:00
lissyx
47f702bf37
Merge pull request #2768 from lissyx/auto-decompress-artifacts
...
Automagically decompress GZipped artifacts
2020-02-18 15:29:46 +01:00
Alexandre Lissy
020619fa97
Automagically decompress GZipped artifacts
...
Fixes #2760
2020-02-18 13:59:30 +01:00
Reuben Morais
44ff4c54b9
Merge pull request #2767 from reuben/readable_repr_bindings
...
Add a better __repr__ for Metadata objects in Python bindings
2020-02-18 13:28:57 +01:00
Reuben Morais
ac26a785df
Add a better __repr__ for Metadata objects in Python bindings
2020-02-18 12:19:10 +01:00
Reuben Morais
1c69d93b4e
Update training docs to mention new CuDNN flags and checkpoint dir flags
...
X-DeepSpeech: NOBUILD
2020-02-17 17:40:44 +01:00
Reuben Morais
200a46711a
Merge pull request #2763 from reuben/transfer-learning-rebase
...
Transfer learning support
2020-02-17 17:29:03 +01:00
Reuben Morais
bd8b96c19d
Remove unneeded Saver instances
2020-02-17 15:55:16 +01:00
Reuben Morais
0e2f34f8cf
Synchronize TensorFlow logging with --log_level flag
2020-02-17 12:46:00 +01:00
Reuben Morais
c46d8396bc
Respect --load when exporting
2020-02-17 12:45:37 +01:00
lissyx
9dd63d5c87
Merge pull request #2765 from lissyx/bump-v0.7.0-alpha.2
...
Bump VERSION to 0.7.0-alpha.2
2020-02-17 12:13:38 +01:00
Alexandre Lissy
685fb1cc9b
Bump VERSION to 0.7.0-alpha.2
2020-02-17 12:12:57 +01:00
lissyx
8fc6b7b216
Merge pull request #2764 from lissyx/electronjsv8.0
...
Support ElectronJS v8.0
2020-02-17 12:11:44 +01:00
Reuben Morais
cedd72da9b
Force UTF-8 IO encoding
2020-02-17 10:46:40 +01:00
