Closed riyijiye closed 5 years ago
What librosa version do you use? Can you please try the latest version (0.6.3)? I've tested 0.6.1, 0.6.2, 0.6.3 with NVIDIA TensorFlow container: all versions work fine.
the issue was with 0.6.3 and 0.6.2, not sure why, but is ok now with 0.6.1. thanks!
requirements.txt should install the right version..this is a pain :-)
David, Can you open issue on Github with more details on which version worked please? This would help us to track and add corresponding test, Thanks, Boris
On Thu, Jun 6, 2019 at 12:43 PM David Levinthal Ph.D. < notifications@github.com> wrote:
requirements.txt should install the right version..this is a pain :-)
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/NVIDIA/OpenSeq2Seq/issues/408?email_source=notifications&email_token=ABZRNKLTILVBPIOXLTYRFRDPZFSFNA5CNFSM4HG55S3KYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODXD6JNQ#issuecomment-499639478, or mute the thread https://github.com/notifications/unsubscribe-auth/ABZRNKPRXRELQXFYUTG2BUTPZFSFNANCNFSM4HG55S3A .
should 408 just be reopened with the comment that version 0.6.1 works and the current release 0.6.3 does not?
On Fri, Jun 7, 2019 at 8:33 AM Boris Ginsburg notifications@github.com wrote:
David, Can you open issue on Github with more details on which version worked please? This would help us to track and add corresponding test, Thanks, Boris
On Thu, Jun 6, 2019 at 12:43 PM David Levinthal Ph.D. < notifications@github.com> wrote:
requirements.txt should install the right version..this is a pain :-)
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub < https://github.com/NVIDIA/OpenSeq2Seq/issues/408?email_source=notifications&email_token=ABZRNKLTILVBPIOXLTYRFRDPZFSFNA5CNFSM4HG55S3KYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODXD6JNQ#issuecomment-499639478 , or mute the thread < https://github.com/notifications/unsubscribe-auth/ABZRNKPRXRELQXFYUTG2BUTPZFSFNANCNFSM4HG55S3A
.
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/NVIDIA/OpenSeq2Seq/issues/408?email_source=notifications&email_token=ACCS4T2ILIGSS2KTWNECJ5TPZJ5WBA5CNFSM4HG55S3KYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODXGFZWQ#issuecomment-499932378, or mute the thread https://github.com/notifications/unsubscribe-auth/ACCS4T4N5NZ762E5MK4BV2LPZJ5WBANCNFSM4HG55S3A .
Right, librosa changed API in 0.4.2
On Fri, Jun 7, 2019 at 8:52 AM David Levinthal Ph.D. < notifications@github.com> wrote:
should 408 just be reopened with the comment that version 0.6.1 works and the current release 0.6.3 does not?
On Fri, Jun 7, 2019 at 8:33 AM Boris Ginsburg notifications@github.com wrote:
David, Can you open issue on Github with more details on which version worked please? This would help us to track and add corresponding test, Thanks, Boris
On Thu, Jun 6, 2019 at 12:43 PM David Levinthal Ph.D. < notifications@github.com> wrote:
requirements.txt should install the right version..this is a pain :-)
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <
, or mute the thread <
https://github.com/notifications/unsubscribe-auth/ABZRNKPRXRELQXFYUTG2BUTPZFSFNANCNFSM4HG55S3A
.
— You are receiving this because you commented. Reply to this email directly, view it on GitHub < https://github.com/NVIDIA/OpenSeq2Seq/issues/408?email_source=notifications&email_token=ACCS4T2ILIGSS2KTWNECJ5TPZJ5WBA5CNFSM4HG55S3KYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODXGFZWQ#issuecomment-499932378 , or mute the thread < https://github.com/notifications/unsubscribe-auth/ACCS4T4N5NZ762E5MK4BV2LPZJ5WBANCNFSM4HG55S3A
.
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/NVIDIA/OpenSeq2Seq/issues/408?email_source=notifications&email_token=ABZRNKP2B547DCVPZ4C46KDPZJ747A5CNFSM4HG55S3KYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODXGHRIA#issuecomment-499939488, or mute the thread https://github.com/notifications/unsubscribe-auth/ABZRNKI4MNE2WOKTM2PXN2TPZJ747ANCNFSM4HG55S3A .
so do you need another github issue..or is 408 adequate? perfectly happy to make another but it seems a duplicate d
On Fri, Jun 7, 2019 at 1:14 PM Boris Ginsburg notifications@github.com wrote:
Right, librosa changed API in 0.4.2
On Fri, Jun 7, 2019 at 8:52 AM David Levinthal Ph.D. < notifications@github.com> wrote:
should 408 just be reopened with the comment that version 0.6.1 works and the current release 0.6.3 does not?
On Fri, Jun 7, 2019 at 8:33 AM Boris Ginsburg notifications@github.com wrote:
David, Can you open issue on Github with more details on which version worked please? This would help us to track and add corresponding test, Thanks, Boris
On Thu, Jun 6, 2019 at 12:43 PM David Levinthal Ph.D. < notifications@github.com> wrote:
requirements.txt should install the right version..this is a pain :-)
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <
, or mute the thread <
https://github.com/notifications/unsubscribe-auth/ABZRNKPRXRELQXFYUTG2BUTPZFSFNANCNFSM4HG55S3A
.
— You are receiving this because you commented. Reply to this email directly, view it on GitHub <
, or mute the thread <
https://github.com/notifications/unsubscribe-auth/ACCS4T4N5NZ762E5MK4BV2LPZJ5WBANCNFSM4HG55S3A
.
— You are receiving this because you commented. Reply to this email directly, view it on GitHub < https://github.com/NVIDIA/OpenSeq2Seq/issues/408?email_source=notifications&email_token=ABZRNKP2B547DCVPZ4C46KDPZJ747A5CNFSM4HG55S3KYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODXGHRIA#issuecomment-499939488 , or mute the thread < https://github.com/notifications/unsubscribe-auth/ABZRNKI4MNE2WOKTM2PXN2TPZJ747ANCNFSM4HG55S3A
.
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/NVIDIA/OpenSeq2Seq/issues/408?email_source=notifications&email_token=ACCS4TZSC4JS4RDJU4K64Z3PZK6RZA5CNFSM4HG55S3KYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODXG4DBY#issuecomment-500023687, or mute the thread https://github.com/notifications/unsubscribe-auth/ACCS4T2YXSFTFUJQTUWMTYTPZK6RZANCNFSM4HG55S3A .
Ok, now I got it. librosa 0.6.3 works fine in NVIDIA TensorFlow container if it was installed with pip install librosa -U
. This command should update joblib
as well. But requirements.txt
contains joblib==0.11
which was another fix for older librosa versions. And librosa 0.6.3 isn't compatible with joblib 0.11. I've updated joblib and librosa versions in requirements.txt. Thanks for the issue report!
will this work with natively installed environments? ie..not using the docker images? d
On Fri, Jun 7, 2019 at 2:45 PM Vitaly Lavrukhin notifications@github.com wrote:
Ok, now I got it. librosa 0.6.3 works fine in NVIDIA TensorFlow container if it was installed with pip install librosa -U. This command should update joblib as well. But requirements.txt contains joblib==0.11 which was another fix for older librosa versions. And librosa 0.6.3 isn't compatible with joblib 0.11. I've updated joblib and librosa versions in requirements.txt. Thanks for the issue report!
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/NVIDIA/OpenSeq2Seq/issues/408?email_source=notifications&email_token=ACCS4T34UP7WYG4HWR4L4Q3PZLJG5A5CNFSM4HG55S3KYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODXHCKFY#issuecomment-500049175, or mute the thread https://github.com/notifications/unsubscribe-auth/ACCS4T6JWSDL367TZ4OLUVDPZLJG5ANCNFSM4HG55S3A .
It should work now. The main problem with previous requirements.txt was that it: 1) installed the latest librosa (0.6.3 and its dependency joblib 0.13.2) 2) uninstalled joblib 0.13.2 and installed required joblib 0.11 (making librosa broken)
thanks.. :-)
On Fri, Jun 7, 2019 at 3:01 PM Vitaly Lavrukhin notifications@github.com wrote:
It should work now. The main problem with previous requirements.txt was that it:
- installed the latest librosa (0.6.3 and its dependency joblib 0.13.2)
- uninstalled joblib 0.13.2 and installed required joblib 0.11 (making librosa broken)
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/NVIDIA/OpenSeq2Seq/issues/408?email_source=notifications&email_token=ACCS4T3F7SUX6ZYVT5EVB6LPZLLDLA5CNFSM4HG55S3KYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODXHDIGQ#issuecomment-500053018, or mute the thread https://github.com/notifications/unsubscribe-auth/ACCS4TYOTT4VWRBHZ2IQ2IDPZLLDLANCNFSM4HG55S3A .
after installation following general installation instruction, I tried to test it with bash scripts/run_all_tests.sh, which gives below errors,
ImportError: Failed to import test module: utils Traceback (most recent call last): File "/home/yinyan/anaconda3/lib/python3.6/unittest/loader.py", line 462, in _find_test_path package = self._get_module_from_name(name) File "/home/yinyan/anaconda3/lib/python3.6/unittest/loader.py", line 369, in _get_module_from_name import(name) File "/workspace/yinyan/cnn_seq2seq_work/OpenSeq2Seq/open_seq2seq/utils/init.py", line 2, in
from .funcs import train, infer, evaluate
File "/workspace/yinyan/cnn_seq2seq_work/OpenSeq2Seq/open_seq2seq/utils/funcs.py", line 18, in
from open_seq2seq.data import WKTDataLayer
File "/workspace/yinyan/cnn_seq2seq_work/OpenSeq2Seq/open_seq2seq/data/init.py", line 3, in
from .speech2text.speech2text import Speech2TextDataLayer
File "/workspace/yinyan/cnn_seq2seq_work/OpenSeq2Seq/open_seq2seq/data/speech2text/speech2text.py", line 12, in
import librosa
File "/home/yinyan/anaconda3/lib/python3.6/site-packages/librosa/init.py", line 12, in
from . import core
File "/home/yinyan/anaconda3/lib/python3.6/site-packages/librosa/core/init.py", line 110, in
from .audio import # pylint: disable=wildcard-import
File "/home/yinyan/anaconda3/lib/python3.6/site-packages/librosa/core/audio.py", line 16, in
from .. import util
File "/home/yinyan/anaconda3/lib/python3.6/site-packages/librosa/util/init.py", line 67, in
from .utils import # pylint: disable=wildcard-import
File "/home/yinyan/anaconda3/lib/python3.6/site-packages/librosa/util/utils.py", line 111, in
def valid_audio(y, mono=True):
File "/home/yinyan/anaconda3/lib/python3.6/site-packages/librosa/cache.py", line 49, in wrapper
if self.location is not None and self.level >= level:
AttributeError: 'CacheManager' object has no attribute 'location'
anyone can help? thanks!