diff --git a/.compute b/.compute index 8603fb87..08b8d8df 100755 --- a/.compute +++ b/.compute @@ -38,7 +38,7 @@ scratch=0 num_layers_dropped=5 -target_lang='ky' +target_lang='cy' source_model='en' input_dir="/snakepit/shared/data/mozilla/CommonVoice/v2.0-alpha1.0" #input_dir="/home/josh/Downloads/CV" diff --git a/.install b/.install index dc4bfa6b..b1c4c758 100755 --- a/.install +++ b/.install @@ -13,8 +13,8 @@ python3 util/taskcluster.py --arch gpu --target ../tmp/native_client pip3 install $(python3 util/taskcluster.py --decoder) ## GET SOURCE CHECKPOINT ## -SOURCE_MODEL='../keep/en' -mkdir $SOURCE_MODEL -wget https://github.com/mozilla/DeepSpeech/releases/download/v0.3.0/deepspeech-0.3.0-checkpoint.tar.gz --directory-prefix="${SOURCE_MODEL}" -tar xvzf ${SOURCE_MODEL}/deepspeech-0.3.0-checkpoint.tar.gz --directory "${SOURCE_MODEL}" -rm ${SOURCE_MODEL}/deepspeech-0.3.0-checkpoint.tar.gz +# SOURCE_MODEL='../keep/en' +# mkdir $SOURCE_MODEL +# wget https://github.com/mozilla/DeepSpeech/releases/download/v0.3.0/deepspeech-0.3.0-checkpoint.tar.gz --directory-prefix="${SOURCE_MODEL}" +# tar xvzf ${SOURCE_MODEL}/deepspeech-0.3.0-checkpoint.tar.gz --directory "${SOURCE_MODEL}" +# rm ${SOURCE_MODEL}/deepspeech-0.3.0-checkpoint.tar.gz diff --git a/data/alphabet.txt b/data/alphabet.txt index 326002cf..e40b5067 100644 --- a/data/alphabet.txt +++ b/data/alphabet.txt @@ -1,77 +1,65 @@ +- , ! ? . -• -— +’ +' +\\ +% +3 a +A +â +ä +b +B c +C +d +D +e +E +ê +f +F +g +G +h +H i +I +î +ï +j +J +k +K +l +L +m +M n +N +o +O +ô p +P r +R s +S t -а -А -б -Б -в -В -г -Г -д -Д -е -Е -ё -ж -Ж -з -З -и -И -й -к -К -л -Л -м -М -н -Н -ң -о -О -ө -Ө -п -П -р -Р -с -С -т -Т -у -У -ү -Ү -ф -Ф -х -Х -ц -ч -Ч -ш -Ш -ъ -ы -ь -э -Э -ю -я -Я +T +u +U +v +w +W +ŵ +y +Y +ŷ +z diff --git a/data/lm/lm.binary b/data/lm/lm.binary index 1c56d0b6..fcd25e09 100644 --- a/data/lm/lm.binary +++ b/data/lm/lm.binary @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:d804a90a78d9c5edbce735862d357891dafce4c8dfc99c31a2f52c82d78608ab -size 43081482 +oid sha256:0ac1b48a8ec930bccaa19c72d4d3eb095880fb5f6ed02f83ae0205f2ba60d448 +size 26612630 diff --git a/data/lm/trie b/data/lm/trie index 0787e7a6..7369e394 100644 --- a/data/lm/trie +++ b/data/lm/trie @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:0100fc12c74f397b646673b5ec0126839f53d19d43c1b6e6aae1bd38c05b630c -size 14494419 +oid sha256:f345667e7a205e59e29c3d26412524282bca6e418dd310cae7cd8e202dc639fc +size 9331231