No matching distribution found for tensorflow==1.15.2 (from deepspeech-training==0.7.3)

Please be more specific than “throws a lot of http errors”, I can’t read in your mind nor on your computer.

Sorry about that, it was getting late and I’d assumed I made some lazy mistake. I cleared the notebook now, restarted the runtime, and retraced the steps now it’s working perfectly. The ldc93s1 script too. I can just now download this folder and work with it on my computer?

So I did just that: cloned on colab, zipped the folder and copied it to drive, then downloaded from drive (took 20 minutes as compared to cloning locally which didn’t give me anything even after waiting for 2 hours), and now everything’s set up locally.

I guess that concludes the thread.

For posterity having the issue in the title: check you python3 --version. The reason why it didn’t work with the anaconda environment was that there python3 pointed to python 3.8 and python pointed to python3.6. I’m quite certain the reason for the tensorflow issue was that. You can set your python3 alias to python3.6 to solve it.

I’d like to thank Olaf and lissyx for helping me with this.

As I understood in this thread next lines are just warning.

swig/python detected a memory leak of type 'Output *', no destructor found.
swig/python detected a memory leak of type 'std::vector< Output,std::allocator< Output > > *', no destructor found.
swig/python detected a memory leak of type 'std::vector< std::vector< Output,std::allocator< Output > > > *', no destructor found.
swig/python detected a memory leak of type 'Alphabet *', no destructor found.
swig/python detected a memory leak of type 'Output *', no destructor found.
swig/python detected a memory leak of type 'std::vector< Output,std::allocator< Output > > *', no destructor found.
swig/python detected a memory leak of type 'std::vector< std::vector< Output,std::allocator< Output > > > *', no destructor found.
swig/python detected a memory leak of type 'Alphabet *', no destructor found.

I have a proper scorer branch(0.7.3). it is working well. but these warnings are annoying. Does this warning affect the result?

how to solve them? or at least how to hide them?

Info :

  • DeepSpeech version: 0.7.3
  • ubuntu 18.04
  • see this warnings in test level.

no

Just use latest 0.7.4 or 0.8, as documented it was fixed.

1 Like

Maybe some of you have already resolved the issue
# [No matching distribution found for tensorflow==1.15.2 (from deepspeech-training==0.7.3)](https://discourse.mozilla.org/t/no-matching-distribution-found-for-tensorflow-1-15-2-from-deepspeech-training-0-7-3/61485)
. but, the problem could be python 3,7 and above. If you downgrade python to 3.6, this issue gets resolved. or downgrade python, than install TF 1.15.2 than upgrade python again. give it a try.

Could you fix it? I got this error : Could not find a version that satisfies the requirement tensorflow==2.3

I have used 0.8.2 with 1.15.2 TF and it worked fine. maybe I have not yet upgraded TF to 2.3

You have to use TF 1.15 for now, it is not compatible with 2.x

We don’t support 2.3

Please upgrade and use 1.15.4 as documented, it contains bugfixes important for deepspeech

1 Like