updates
Browse files
err2020/conformer_ctc3/asr_datamodule.py
CHANGED
@@ -31,9 +31,10 @@ from lhotse.dataset import ( # noqa F401 for PrecomputedFeatures
|
|
31 |
DynamicBucketingSampler,
|
32 |
K2SpeechRecognitionDataset,
|
33 |
PrecomputedFeatures,
|
34 |
-
|
35 |
SpecAugment,
|
36 |
)
|
|
|
37 |
from lhotse.dataset.input_strategies import ( # noqa F401 For AudioSamples
|
38 |
AudioSamples,
|
39 |
OnTheFlyFeatures,
|
@@ -307,7 +308,7 @@ class LibriSpeechAsrDataModule:
|
|
307 |
)
|
308 |
else:
|
309 |
logging.info("Using SingleCutSampler.")
|
310 |
-
train_sampler =
|
311 |
cuts_train,
|
312 |
max_duration=self.args.max_duration,
|
313 |
shuffle=self.args.shuffle,
|
|
|
31 |
DynamicBucketingSampler,
|
32 |
K2SpeechRecognitionDataset,
|
33 |
PrecomputedFeatures,
|
34 |
+
SimpleCutSampler,
|
35 |
SpecAugment,
|
36 |
)
|
37 |
+
|
38 |
from lhotse.dataset.input_strategies import ( # noqa F401 For AudioSamples
|
39 |
AudioSamples,
|
40 |
OnTheFlyFeatures,
|
|
|
308 |
)
|
309 |
else:
|
310 |
logging.info("Using SingleCutSampler.")
|
311 |
+
train_sampler = SimpleCutSampler(
|
312 |
cuts_train,
|
313 |
max_duration=self.args.max_duration,
|
314 |
shuffle=self.args.shuffle,
|
err2020/conformer_ctc3_usage.ipynb
CHANGED
@@ -2,7 +2,7 @@
|
|
2 |
"cells": [
|
3 |
{
|
4 |
"cell_type": "code",
|
5 |
-
"execution_count":
|
6 |
"id": "b6b6ded1-0a58-43cb-9065-4f4fae02a01b",
|
7 |
"metadata": {
|
8 |
"pycharm": {
|
@@ -62,7 +62,7 @@
|
|
62 |
},
|
63 |
{
|
64 |
"cell_type": "code",
|
65 |
-
"execution_count":
|
66 |
"id": "3d69d771-b421-417f-a6ff-e1d1c64ba934",
|
67 |
"metadata": {
|
68 |
"pycharm": {
|
@@ -141,7 +141,7 @@
|
|
141 |
},
|
142 |
{
|
143 |
"cell_type": "code",
|
144 |
-
"execution_count":
|
145 |
"id": "48306369-fb68-4abe-be62-0806d00059f8",
|
146 |
"metadata": {
|
147 |
"pycharm": {
|
@@ -420,7 +420,7 @@
|
|
420 |
},
|
421 |
{
|
422 |
"cell_type": "code",
|
423 |
-
"execution_count":
|
424 |
"id": "50ab7c8e-39b6-4783-8342-e79e91d2417e",
|
425 |
"metadata": {
|
426 |
"pycharm": {
|
@@ -438,7 +438,7 @@
|
|
438 |
},
|
439 |
{
|
440 |
"cell_type": "code",
|
441 |
-
"execution_count":
|
442 |
"id": "8020f371-7584-4f6c-990b-f2c023e24060",
|
443 |
"metadata": {
|
444 |
"pycharm": {
|
@@ -450,8 +450,8 @@
|
|
450 |
"name": "stdout",
|
451 |
"output_type": "stream",
|
452 |
"text": [
|
453 |
-
"CPU times: user
|
454 |
-
"Wall time:
|
455 |
]
|
456 |
},
|
457 |
{
|
@@ -473,7 +473,7 @@
|
|
473 |
" {'word': 'panna', 'start': 10.16, 'end': 10.4}]}"
|
474 |
]
|
475 |
},
|
476 |
-
"execution_count":
|
477 |
"metadata": {},
|
478 |
"output_type": "execute_result"
|
479 |
}
|
@@ -483,6 +483,296 @@
|
|
483 |
"%time transcriber.transcribe_file('audio/emt16k.wav')"
|
484 |
]
|
485 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
486 |
{
|
487 |
"cell_type": "code",
|
488 |
"execution_count": 9,
|
@@ -789,7 +1079,7 @@
|
|
789 |
"name": "python",
|
790 |
"nbconvert_exporter": "python",
|
791 |
"pygments_lexer": "ipython3",
|
792 |
-
"version": "3.9.
|
793 |
}
|
794 |
},
|
795 |
"nbformat": 4,
|
|
|
2 |
"cells": [
|
3 |
{
|
4 |
"cell_type": "code",
|
5 |
+
"execution_count": 1,
|
6 |
"id": "b6b6ded1-0a58-43cb-9065-4f4fae02a01b",
|
7 |
"metadata": {
|
8 |
"pycharm": {
|
|
|
62 |
},
|
63 |
{
|
64 |
"cell_type": "code",
|
65 |
+
"execution_count": 2,
|
66 |
"id": "3d69d771-b421-417f-a6ff-e1d1c64ba934",
|
67 |
"metadata": {
|
68 |
"pycharm": {
|
|
|
141 |
},
|
142 |
{
|
143 |
"cell_type": "code",
|
144 |
+
"execution_count": 3,
|
145 |
"id": "48306369-fb68-4abe-be62-0806d00059f8",
|
146 |
"metadata": {
|
147 |
"pycharm": {
|
|
|
420 |
},
|
421 |
{
|
422 |
"cell_type": "code",
|
423 |
+
"execution_count": 4,
|
424 |
"id": "50ab7c8e-39b6-4783-8342-e79e91d2417e",
|
425 |
"metadata": {
|
426 |
"pycharm": {
|
|
|
438 |
},
|
439 |
{
|
440 |
"cell_type": "code",
|
441 |
+
"execution_count": 11,
|
442 |
"id": "8020f371-7584-4f6c-990b-f2c023e24060",
|
443 |
"metadata": {
|
444 |
"pycharm": {
|
|
|
450 |
"name": "stdout",
|
451 |
"output_type": "stream",
|
452 |
"text": [
|
453 |
+
"CPU times: user 7.06 s, sys: 655 ms, total: 7.71 s\n",
|
454 |
+
"Wall time: 9.01 s\n"
|
455 |
]
|
456 |
},
|
457 |
{
|
|
|
473 |
" {'word': 'panna', 'start': 10.16, 'end': 10.4}]}"
|
474 |
]
|
475 |
},
|
476 |
+
"execution_count": 11,
|
477 |
"metadata": {},
|
478 |
"output_type": "execute_result"
|
479 |
}
|
|
|
483 |
"%time transcriber.transcribe_file('audio/emt16k.wav')"
|
484 |
]
|
485 |
},
|
486 |
+
{
|
487 |
+
"cell_type": "code",
|
488 |
+
"execution_count": 15,
|
489 |
+
"id": "dc56ed92-9e9c-4316-bc81-a736a0d560a4",
|
490 |
+
"metadata": {
|
491 |
+
"pycharm": {
|
492 |
+
"name": "#%%\n"
|
493 |
+
}
|
494 |
+
},
|
495 |
+
"outputs": [
|
496 |
+
{
|
497 |
+
"name": "stdout",
|
498 |
+
"output_type": "stream",
|
499 |
+
"text": [
|
500 |
+
"ffprobe version 4.2.7-0ubuntu0.1 Copyright (c) 2007-2022 the FFmpeg developers\n",
|
501 |
+
" built with gcc 9 (Ubuntu 9.4.0-1ubuntu1~20.04.1)\n",
|
502 |
+
" configuration: --prefix=/usr --extra-version=0ubuntu0.1 --toolchain=hardened --libdir=/usr/lib/x86_64-linux-gnu --incdir=/usr/include/x86_64-linux-gnu --arch=amd64 --enable-gpl --disable-stripping --enable-avresample --disable-filter=resample --enable-avisynth --enable-gnutls --enable-ladspa --enable-libaom --enable-libass --enable-libbluray --enable-libbs2b --enable-libcaca --enable-libcdio --enable-libcodec2 --enable-libflite --enable-libfontconfig --enable-libfreetype --enable-libfribidi --enable-libgme --enable-libgsm --enable-libjack --enable-libmp3lame --enable-libmysofa --enable-libopenjpeg --enable-libopenmpt --enable-libopus --enable-libpulse --enable-librsvg --enable-librubberband --enable-libshine --enable-libsnappy --enable-libsoxr --enable-libspeex --enable-libssh --enable-libtheora --enable-libtwolame --enable-libvidstab --enable-libvorbis --enable-libvpx --enable-libwavpack --enable-libwebp --enable-libx265 --enable-libxml2 --enable-libxvid --enable-libzmq --enable-libzvbi --enable-lv2 --enable-omx --enable-openal --enable-opencl --enable-opengl --enable-sdl2 --enable-libdc1394 --enable-libdrm --enable-libiec61883 --enable-nvenc --enable-chromaprint --enable-frei0r --enable-libx264 --enable-shared\n",
|
503 |
+
" libavutil 56. 31.100 / 56. 31.100\n",
|
504 |
+
" libavcodec 58. 54.100 / 58. 54.100\n",
|
505 |
+
" libavformat 58. 29.100 / 58. 29.100\n",
|
506 |
+
" libavdevice 58. 8.100 / 58. 8.100\n",
|
507 |
+
" libavfilter 7. 57.100 / 7. 57.100\n",
|
508 |
+
" libavresample 4. 0. 0 / 4. 0. 0\n",
|
509 |
+
" libswscale 5. 5.100 / 5. 5.100\n",
|
510 |
+
" libswresample 3. 5.100 / 3. 5.100\n",
|
511 |
+
" libpostproc 55. 5.100 / 55. 5.100\n",
|
512 |
+
"Input #0, wav, from 'audio/emt16k.wav':\n",
|
513 |
+
" Duration: 00:00:12.74, bitrate: 256 kb/s\n",
|
514 |
+
" Stream #0:0: Audio: pcm_s16le ([1][0][0][0] / 0x0001), 16000 Hz, 1 channels, s16, 256 kb/s\n"
|
515 |
+
]
|
516 |
+
}
|
517 |
+
],
|
518 |
+
"source": [
|
519 |
+
"!ffprobe -i audio/emt16k.wav"
|
520 |
+
]
|
521 |
+
},
|
522 |
+
{
|
523 |
+
"cell_type": "code",
|
524 |
+
"execution_count": 13,
|
525 |
+
"id": "e7251753-89a4-4625-8f47-71b28f73ee3d",
|
526 |
+
"metadata": {
|
527 |
+
"pycharm": {
|
528 |
+
"name": "#%%\n"
|
529 |
+
}
|
530 |
+
},
|
531 |
+
"outputs": [
|
532 |
+
{
|
533 |
+
"name": "stdout",
|
534 |
+
"output_type": "stream",
|
535 |
+
"text": [
|
536 |
+
"ffprobe version 4.2.7-0ubuntu0.1 Copyright (c) 2007-2022 the FFmpeg developers\n",
|
537 |
+
" built with gcc 9 (Ubuntu 9.4.0-1ubuntu1~20.04.1)\n",
|
538 |
+
" configuration: --prefix=/usr --extra-version=0ubuntu0.1 --toolchain=hardened --libdir=/usr/lib/x86_64-linux-gnu --incdir=/usr/include/x86_64-linux-gnu --arch=amd64 --enable-gpl --disable-stripping --enable-avresample --disable-filter=resample --enable-avisynth --enable-gnutls --enable-ladspa --enable-libaom --enable-libass --enable-libbluray --enable-libbs2b --enable-libcaca --enable-libcdio --enable-libcodec2 --enable-libflite --enable-libfontconfig --enable-libfreetype --enable-libfribidi --enable-libgme --enable-libgsm --enable-libjack --enable-libmp3lame --enable-libmysofa --enable-libopenjpeg --enable-libopenmpt --enable-libopus --enable-libpulse --enable-librsvg --enable-librubberband --enable-libshine --enable-libsnappy --enable-libsoxr --enable-libspeex --enable-libssh --enable-libtheora --enable-libtwolame --enable-libvidstab --enable-libvorbis --enable-libvpx --enable-libwavpack --enable-libwebp --enable-libx265 --enable-libxml2 --enable-libxvid --enable-libzmq --enable-libzvbi --enable-lv2 --enable-omx --enable-openal --enable-opencl --enable-opengl --enable-sdl2 --enable-libdc1394 --enable-libdrm --enable-libiec61883 --enable-nvenc --enable-chromaprint --enable-frei0r --enable-libx264 --enable-shared\n",
|
539 |
+
" libavutil 56. 31.100 / 56. 31.100\n",
|
540 |
+
" libavcodec 58. 54.100 / 58. 54.100\n",
|
541 |
+
" libavformat 58. 29.100 / 58. 29.100\n",
|
542 |
+
" libavdevice 58. 8.100 / 58. 8.100\n",
|
543 |
+
" libavfilter 7. 57.100 / 7. 57.100\n",
|
544 |
+
" libavresample 4. 0. 0 / 4. 0. 0\n",
|
545 |
+
" libswscale 5. 5.100 / 5. 5.100\n",
|
546 |
+
" libswresample 3. 5.100 / 3. 5.100\n",
|
547 |
+
" libpostproc 55. 5.100 / 55. 5.100\n",
|
548 |
+
"Input #0, wav, from 'unejutt_wav/New Recording 13.wav':\n",
|
549 |
+
" Metadata:\n",
|
550 |
+
" title : New Recording 13\n",
|
551 |
+
" encoder : Lavf58.29.100\n",
|
552 |
+
" Duration: 00:07:40.65, bitrate: 256 kb/s\n",
|
553 |
+
" Stream #0:0: Audio: pcm_s16le ([1][0][0][0] / 0x0001), 16000 Hz, 1 channels, s16, 256 kb/s\n"
|
554 |
+
]
|
555 |
+
}
|
556 |
+
],
|
557 |
+
"source": [
|
558 |
+
"!ffprobe -i 'unejutt_wav/New Recording 13.wav'"
|
559 |
+
]
|
560 |
+
},
|
561 |
+
{
|
562 |
+
"cell_type": "markdown",
|
563 |
+
"id": "cc66b296-6307-4ba3-93d4-a2b87cca32bf",
|
564 |
+
"metadata": {
|
565 |
+
"pycharm": {
|
566 |
+
"name": "#%% md\n"
|
567 |
+
}
|
568 |
+
},
|
569 |
+
"source": [
|
570 |
+
"## convert m4a to wav"
|
571 |
+
]
|
572 |
+
},
|
573 |
+
{
|
574 |
+
"cell_type": "code",
|
575 |
+
"execution_count": 17,
|
576 |
+
"id": "489ed4d2-58af-44d8-9031-72c98c8c0798",
|
577 |
+
"metadata": {
|
578 |
+
"pycharm": {
|
579 |
+
"name": "#%%\n"
|
580 |
+
}
|
581 |
+
},
|
582 |
+
"outputs": [
|
583 |
+
{
|
584 |
+
"name": "stdout",
|
585 |
+
"output_type": "stream",
|
586 |
+
"text": [
|
587 |
+
"CPU times: user 31 s, sys: 1.94 s, total: 33 s\n",
|
588 |
+
"Wall time: 32.7 s\n"
|
589 |
+
]
|
590 |
+
}
|
591 |
+
],
|
592 |
+
"source": [
|
593 |
+
"#transribe audiofile (NB! model assumes sample rate of 16000)\n",
|
594 |
+
"%time trs=transcriber.transcribe_file('unejutt_wav/sub.wav')"
|
595 |
+
]
|
596 |
+
},
|
597 |
+
{
|
598 |
+
"cell_type": "code",
|
599 |
+
"execution_count": 18,
|
600 |
+
"id": "11962120-708f-462d-97ac-7ef36ace464d",
|
601 |
+
"metadata": {
|
602 |
+
"pycharm": {
|
603 |
+
"name": "#%%\n"
|
604 |
+
}
|
605 |
+
},
|
606 |
+
"outputs": [
|
607 |
+
{
|
608 |
+
"data": {
|
609 |
+
"text/plain": [
|
610 |
+
"{'text': 'see muttikas oli ju seal sees kinni koju ei saanud ikkagi sellega minna aga ei olnud midagi altperevellu oli ka val mees tema käed oli kultsed hakkas mõtlema et kuidas kuidas ta nüüd ikkagi õhtul koju saab siis ta hommikul tuli selle muttikaga sinna sõitis sisse kiiveri peas jäi terveks ei juhtunud midagi kuidas sa nüüd õhtul koju lähed kui see muttikas on juba seal kuivati sisse sõidetud ja käite plevaatorit saa elevaatrika seisma panna kuivati põlemad vilja kõikuvad teise põ kuivati maha seal küla nälga ja ja noh see oli sügaval nõukaal siis partei poolt pandud need plaanid ja normid jäävad täitmata ja siis noh ontitsa nagu pekkis kohahest et seal tulevad seal meil mingid sõdurid ja värgid ka peale et hakka seda jama seal tegema siis ta juba tükk aega mõtles et et ega see see kiiravitsmootor see on ju niisugune et uimass võimas moootor sellegaetakse tankegi siin minema põllu peaelt mõeldud',\n",
|
611 |
+
" 'words': [{'word': 'see', 'start': 0.48, 'end': 0.52},\n",
|
612 |
+
" {'word': 'muttikas', 'start': 0.6, 'end': 1.0},\n",
|
613 |
+
" {'word': 'oli', 'start': 1.08, 'end': 1.12},\n",
|
614 |
+
" {'word': 'ju', 'start': 1.2, 'end': 1.24},\n",
|
615 |
+
" {'word': 'seal', 'start': 1.32, 'end': 1.36},\n",
|
616 |
+
" {'word': 'sees', 'start': 1.48, 'end': 1.64},\n",
|
617 |
+
" {'word': 'kinni', 'start': 1.68, 'end': 1.72},\n",
|
618 |
+
" {'word': 'koju', 'start': 3.36, 'end': 3.48},\n",
|
619 |
+
" {'word': 'ei', 'start': 3.56, 'end': 3.64},\n",
|
620 |
+
" {'word': 'saanud', 'start': 3.64, 'end': 3.8},\n",
|
621 |
+
" {'word': 'ikkagi', 'start': 3.96, 'end': 4.0},\n",
|
622 |
+
" {'word': 'sellega', 'start': 4.16, 'end': 4.36},\n",
|
623 |
+
" {'word': 'minna', 'start': 4.56, 'end': 4.6},\n",
|
624 |
+
" {'word': 'aga', 'start': 6.72, 'end': 6.76},\n",
|
625 |
+
" {'word': 'ei', 'start': 6.88, 'end': 6.96},\n",
|
626 |
+
" {'word': 'olnud', 'start': 7.04, 'end': 7.08},\n",
|
627 |
+
" {'word': 'midagi', 'start': 7.28, 'end': 7.32},\n",
|
628 |
+
" {'word': 'altperevellu', 'start': 8.6, 'end': 9.24},\n",
|
629 |
+
" {'word': 'oli', 'start': 9.72, 'end': 9.76},\n",
|
630 |
+
" {'word': 'ka', 'start': 9.8, 'end': 9.84},\n",
|
631 |
+
" {'word': 'val', 'start': 9.88, 'end': 10.0},\n",
|
632 |
+
" {'word': 'mees', 'start': 10.04, 'end': 10.28},\n",
|
633 |
+
" {'word': 'tema', 'start': 11.12, 'end': 11.24},\n",
|
634 |
+
" {'word': 'käed', 'start': 11.56, 'end': 11.8},\n",
|
635 |
+
" {'word': 'oli', 'start': 11.84, 'end': 11.88},\n",
|
636 |
+
" {'word': 'kultsed', 'start': 11.92, 'end': 12.2},\n",
|
637 |
+
" {'word': 'hakkas', 'start': 13.48, 'end': 13.6},\n",
|
638 |
+
" {'word': 'mõtlema', 'start': 13.64, 'end': 13.92},\n",
|
639 |
+
" {'word': 'et', 'start': 13.96, 'end': 14.0},\n",
|
640 |
+
" {'word': 'kuidas', 'start': 14.08, 'end': 14.12},\n",
|
641 |
+
" {'word': 'kuidas', 'start': 15.28, 'end': 15.32},\n",
|
642 |
+
" {'word': 'ta', 'start': 15.52, 'end': 15.56},\n",
|
643 |
+
" {'word': 'nüüd', 'start': 15.64, 'end': 15.68},\n",
|
644 |
+
" {'word': 'ikkagi', 'start': 15.84, 'end': 15.88},\n",
|
645 |
+
" {'word': 'õhtul', 'start': 16.0, 'end': 16.24},\n",
|
646 |
+
" {'word': 'koju', 'start': 16.24, 'end': 16.4},\n",
|
647 |
+
" {'word': 'saab', 'start': 16.44, 'end': 16.48},\n",
|
648 |
+
" {'word': 'siis', 'start': 16.68, 'end': 16.72},\n",
|
649 |
+
" {'word': 'ta', 'start': 16.84, 'end': 16.88},\n",
|
650 |
+
" {'word': 'hommikul', 'start': 17.0, 'end': 17.36},\n",
|
651 |
+
" {'word': 'tuli', 'start': 17.44, 'end': 17.6},\n",
|
652 |
+
" {'word': 'selle', 'start': 17.64, 'end': 17.68},\n",
|
653 |
+
" {'word': 'muttikaga', 'start': 18.28, 'end': 18.68},\n",
|
654 |
+
" {'word': 'sinna', 'start': 18.72, 'end': 18.76},\n",
|
655 |
+
" {'word': 'sõitis', 'start': 18.96, 'end': 19.24},\n",
|
656 |
+
" {'word': 'sisse', 'start': 19.28, 'end': 19.32},\n",
|
657 |
+
" {'word': 'kiiveri', 'start': 20.52, 'end': 20.8},\n",
|
658 |
+
" {'word': 'peas', 'start': 20.88, 'end': 21.12},\n",
|
659 |
+
" {'word': 'jäi', 'start': 21.16, 'end': 21.36},\n",
|
660 |
+
" {'word': 'terveks', 'start': 21.36, 'end': 21.72},\n",
|
661 |
+
" {'word': 'ei', 'start': 21.76, 'end': 21.8},\n",
|
662 |
+
" {'word': 'juhtunud', 'start': 21.96, 'end': 22.08},\n",
|
663 |
+
" {'word': 'midagi', 'start': 22.2, 'end': 22.24},\n",
|
664 |
+
" {'word': 'kuidas', 'start': 23.04, 'end': 23.08},\n",
|
665 |
+
" {'word': 'sa', 'start': 23.28, 'end': 23.32},\n",
|
666 |
+
" {'word': 'nüüd', 'start': 23.36, 'end': 23.4},\n",
|
667 |
+
" {'word': 'õhtul', 'start': 23.48, 'end': 23.72},\n",
|
668 |
+
" {'word': 'koju', 'start': 23.72, 'end': 23.84},\n",
|
669 |
+
" {'word': 'lähed', 'start': 23.92, 'end': 24.12},\n",
|
670 |
+
" {'word': 'kui', 'start': 24.16, 'end': 24.2},\n",
|
671 |
+
" {'word': 'see', 'start': 24.24, 'end': 24.28},\n",
|
672 |
+
" {'word': 'muttikas', 'start': 24.4, 'end': 24.76},\n",
|
673 |
+
" {'word': 'on', 'start': 24.8, 'end': 24.88},\n",
|
674 |
+
" {'word': 'juba', 'start': 24.96, 'end': 25.0},\n",
|
675 |
+
" {'word': 'seal', 'start': 25.08, 'end': 25.12},\n",
|
676 |
+
" {'word': 'kuivati', 'start': 25.52, 'end': 25.8},\n",
|
677 |
+
" {'word': 'sisse', 'start': 25.84, 'end': 25.88},\n",
|
678 |
+
" {'word': 'sõidetud', 'start': 26.12, 'end': 26.4},\n",
|
679 |
+
" {'word': 'ja', 'start': 26.52, 'end': 26.56},\n",
|
680 |
+
" {'word': 'käite', 'start': 27.56, 'end': 27.8},\n",
|
681 |
+
" {'word': 'plevaatorit', 'start': 27.84, 'end': 28.44},\n",
|
682 |
+
" {'word': 'saa', 'start': 28.8, 'end': 28.84},\n",
|
683 |
+
" {'word': 'elevaatrika', 'start': 29.0, 'end': 29.52},\n",
|
684 |
+
" {'word': 'seisma', 'start': 29.56, 'end': 29.8},\n",
|
685 |
+
" {'word': 'panna', 'start': 29.84, 'end': 30.04},\n",
|
686 |
+
" {'word': 'kuivati', 'start': 30.44, 'end': 30.72},\n",
|
687 |
+
" {'word': 'põlemad', 'start': 30.76, 'end': 31.12},\n",
|
688 |
+
" {'word': 'vilja', 'start': 31.36, 'end': 31.6},\n",
|
689 |
+
" {'word': 'kõikuvad', 'start': 31.64, 'end': 31.92},\n",
|
690 |
+
" {'word': 'teise', 'start': 32.2, 'end': 32.24},\n",
|
691 |
+
" {'word': 'põ', 'start': 33.04, 'end': 33.08},\n",
|
692 |
+
" {'word': 'kuivati', 'start': 33.28, 'end': 33.56},\n",
|
693 |
+
" {'word': 'maha', 'start': 33.6, 'end': 33.76},\n",
|
694 |
+
" {'word': 'seal', 'start': 33.88, 'end': 33.92},\n",
|
695 |
+
" {'word': 'küla', 'start': 34.2, 'end': 34.32},\n",
|
696 |
+
" {'word': 'nälga', 'start': 34.44, 'end': 34.68},\n",
|
697 |
+
" {'word': 'ja', 'start': 34.76, 'end': 34.8},\n",
|
698 |
+
" {'word': 'ja', 'start': 35.88, 'end': 35.92},\n",
|
699 |
+
" {'word': 'noh', 'start': 35.96, 'end': 36.0},\n",
|
700 |
+
" {'word': 'see', 'start': 36.24, 'end': 36.28},\n",
|
701 |
+
" {'word': 'oli', 'start': 36.36, 'end': 36.4},\n",
|
702 |
+
" {'word': 'sügaval', 'start': 36.92, 'end': 37.24},\n",
|
703 |
+
" {'word': 'nõukaal', 'start': 37.28, 'end': 37.64},\n",
|
704 |
+
" {'word': 'siis', 'start': 37.72, 'end': 37.76},\n",
|
705 |
+
" {'word': 'partei', 'start': 39.12, 'end': 39.52},\n",
|
706 |
+
" {'word': 'poolt', 'start': 39.52, 'end': 39.76},\n",
|
707 |
+
" {'word': 'pandud', 'start': 39.8, 'end': 40.12},\n",
|
708 |
+
" {'word': 'need', 'start': 40.16, 'end': 40.2},\n",
|
709 |
+
" {'word': 'plaanid', 'start': 40.56, 'end': 40.88},\n",
|
710 |
+
" {'word': 'ja', 'start': 40.88, 'end': 40.96},\n",
|
711 |
+
" {'word': 'normid', 'start': 40.96, 'end': 41.24},\n",
|
712 |
+
" {'word': 'jäävad', 'start': 41.24, 'end': 41.32},\n",
|
713 |
+
" {'word': 'täitmata', 'start': 41.4, 'end': 41.8},\n",
|
714 |
+
" {'word': 'ja', 'start': 41.84, 'end': 41.88},\n",
|
715 |
+
" {'word': 'siis', 'start': 41.92, 'end': 41.96},\n",
|
716 |
+
" {'word': 'noh', 'start': 42.24, 'end': 42.28},\n",
|
717 |
+
" {'word': 'ontitsa', 'start': 42.6, 'end': 43.0},\n",
|
718 |
+
" {'word': 'nagu', 'start': 43.12, 'end': 43.2},\n",
|
719 |
+
" {'word': 'pekkis', 'start': 43.28, 'end': 43.56},\n",
|
720 |
+
" {'word': 'kohahest', 'start': 43.64, 'end': 43.88},\n",
|
721 |
+
" {'word': 'et', 'start': 43.96, 'end': 44.0},\n",
|
722 |
+
" {'word': 'seal', 'start': 44.08, 'end': 44.12},\n",
|
723 |
+
" {'word': 'tulevad', 'start': 45.16, 'end': 45.44},\n",
|
724 |
+
" {'word': 'seal', 'start': 45.56, 'end': 45.6},\n",
|
725 |
+
" {'word': 'meil', 'start': 45.72, 'end': 45.76},\n",
|
726 |
+
" {'word': 'mingid', 'start': 45.96, 'end': 46.08},\n",
|
727 |
+
" {'word': 'sõdurid', 'start': 46.12, 'end': 46.44},\n",
|
728 |
+
" {'word': 'ja', 'start': 46.44, 'end': 46.48},\n",
|
729 |
+
" {'word': 'värgid', 'start': 46.48, 'end': 46.76},\n",
|
730 |
+
" {'word': 'ka', 'start': 46.76, 'end': 46.8},\n",
|
731 |
+
" {'word': 'peale', 'start': 46.88, 'end': 46.92},\n",
|
732 |
+
" {'word': 'et', 'start': 47.16, 'end': 47.2},\n",
|
733 |
+
" {'word': 'hakka', 'start': 47.4, 'end': 47.44},\n",
|
734 |
+
" {'word': 'seda', 'start': 47.52, 'end': 47.56},\n",
|
735 |
+
" {'word': 'jama', 'start': 47.72, 'end': 47.84},\n",
|
736 |
+
" {'word': 'seal', 'start': 47.92, 'end': 47.96},\n",
|
737 |
+
" {'word': 'tegema', 'start': 48.16, 'end': 48.2},\n",
|
738 |
+
" {'word': 'siis', 'start': 50.28, 'end': 50.32},\n",
|
739 |
+
" {'word': 'ta', 'start': 50.36, 'end': 50.4},\n",
|
740 |
+
" {'word': 'juba', 'start': 50.56, 'end': 50.6},\n",
|
741 |
+
" {'word': 'tükk', 'start': 50.68, 'end': 50.92},\n",
|
742 |
+
" {'word': 'aega', 'start': 50.92, 'end': 51.04},\n",
|
743 |
+
" {'word': 'mõtles', 'start': 51.08, 'end': 51.4},\n",
|
744 |
+
" {'word': 'et', 'start': 51.44, 'end': 51.48},\n",
|
745 |
+
" {'word': 'et', 'start': 52.72, 'end': 52.76},\n",
|
746 |
+
" {'word': 'ega', 'start': 52.84, 'end': 52.96},\n",
|
747 |
+
" {'word': 'see', 'start': 53.04, 'end': 53.08},\n",
|
748 |
+
" {'word': 'see', 'start': 53.64, 'end': 53.68},\n",
|
749 |
+
" {'word': 'kiiravitsmootor', 'start': 53.8, 'end': 54.6},\n",
|
750 |
+
" {'word': 'see', 'start': 54.6, 'end': 54.64},\n",
|
751 |
+
" {'word': 'on', 'start': 54.76, 'end': 54.8},\n",
|
752 |
+
" {'word': 'ju', 'start': 54.84, 'end': 54.88},\n",
|
753 |
+
" {'word': 'niisugune', 'start': 54.96, 'end': 55.0},\n",
|
754 |
+
" {'word': 'et', 'start': 55.64, 'end': 55.68},\n",
|
755 |
+
" {'word': 'uimass', 'start': 55.84, 'end': 56.24},\n",
|
756 |
+
" {'word': 'võimas', 'start': 56.32, 'end': 56.56},\n",
|
757 |
+
" {'word': 'moootor', 'start': 56.6, 'end': 56.92},\n",
|
758 |
+
" {'word': 'sellegaetakse', 'start': 57.76, 'end': 58.88},\n",
|
759 |
+
" {'word': 'tankegi', 'start': 59.04, 'end': 59.44},\n",
|
760 |
+
" {'word': 'siin', 'start': 59.48, 'end': 59.52},\n",
|
761 |
+
" {'word': 'minema', 'start': 59.64, 'end': 59.88},\n",
|
762 |
+
" {'word': 'põllu', 'start': 59.96, 'end': 60.2},\n",
|
763 |
+
" {'word': 'peaelt', 'start': 60.2, 'end': 60.36},\n",
|
764 |
+
" {'word': 'mõeldud', 'start': 63.28, 'end': 63.6}]}"
|
765 |
+
]
|
766 |
+
},
|
767 |
+
"execution_count": 18,
|
768 |
+
"metadata": {},
|
769 |
+
"output_type": "execute_result"
|
770 |
+
}
|
771 |
+
],
|
772 |
+
"source": [
|
773 |
+
"trs"
|
774 |
+
]
|
775 |
+
},
|
776 |
{
|
777 |
"cell_type": "code",
|
778 |
"execution_count": 9,
|
|
|
1079 |
"name": "python",
|
1080 |
"nbconvert_exporter": "python",
|
1081 |
"pygments_lexer": "ipython3",
|
1082 |
+
"version": "3.9.18"
|
1083 |
}
|
1084 |
},
|
1085 |
"nbformat": 4,
|