How to install specific native_client version from taskcluster

Deep Speech
#1

Now that there is tooling to add specific versions from the taskcluster, I’m attempting to install v0.1.1 of the native client via the following command:

/content/DeepSpeech/util/taskcluster.py --branch "v0.1.1" --target /content/DeepSpeech/native_client

When doing so I’m getting a URL error. Here is the full output:

Downloading https://index.taskcluster.net/v1/task/project.deepspeech.deepspeech.native_client.v0.1.1.cpu/artifacts/public/native_client.tar.xz ...
Traceback (most recent call last):
  File "/content/DeepSpeech/util/taskcluster.py", line 87, in <module>
    maybe_download_tc(target_dir=args.target, tc_url=get_tc_url(args.arch, args.artifact, args.branch))
  File "/content/DeepSpeech/util/taskcluster.py", line 51, in maybe_download_tc
    urllib.request.urlretrieve(tc_url, target_file, reporthook=(report_progress if progress else None))
  File "/usr/lib/python3.6/urllib/request.py", line 248, in urlretrieve
    with contextlib.closing(urlopen(url, data)) as fp:
  File "/usr/lib/python3.6/urllib/request.py", line 223, in urlopen
    return opener.open(url, data, timeout)
  File "/usr/lib/python3.6/urllib/request.py", line 532, in open
    response = meth(req, response)
  File "/usr/lib/python3.6/urllib/request.py", line 642, in http_response
    'http', request, response, code, msg, hdrs)
  File "/usr/lib/python3.6/urllib/request.py", line 570, in error
    return self._call_chain(*args)
  File "/usr/lib/python3.6/urllib/request.py", line 504, in _call_chain
    result = func(*args)
  File "/usr/lib/python3.6/urllib/request.py", line 650, in http_error_default
    raise HTTPError(req.full_url, code, msg, hdrs, fp)
urllib.error.HTTPError: HTTP Error 404: Not Found

I seem to be passing through the wrong version number. I tried various combinations but am pretty much just guessing (0.1.1, 0.1.1-gpu, etc).

I don’t see a list of version numbers in the documentation however.

#2

No, it happens that we had set expirations for artifacts to 6 months. So we need to rebuild that 0.1.1 version.

#3

Ah gotcha. When I use pip install deepspeech it defaults to 0.1.1. I’d be happy to use the master version of deepspeech to eliminate the native client version problem but not sure how to do that as there is no flag to specify version.

Right now I’m not beholden to any model version as I’m building a small prototype for Mongolian. Thanks

Robert Ritz

#4

You can rely on alphas, we publish them on Pypi and NPM now. And the taskcluster.py has been indeed augmented with a version flag: https://github.com/mozilla/DeepSpeech/blob/master/util/taskcluster.py#L75-L76

So passing --branch v0.2.0-alpha.9 should get you those artifacts: https://tools.taskcluster.net/index/project.deepspeech.deepspeech.native_client.v0.2.0-alpha.9

#5

Ok great. I’ll try it out tomorrow and report back.

#6

The taskcluster branch flag is working perfectly. However pip install deepspeech still always installs 0.1.1.

To better understand how it works, this seems to be connected to the deepspeech files downloaded from Git LFS? The larger issue I’m having is with generating a trie from my language model. Per a previous thread the issue is probably my deepspeech and native client being different versions.

However now I’m trying to determine how to install the alpha version of deepspeech to correspond with the native client.

#7

pip install deepspeech==0.2.0a9: https://pypi.org/project/deepspeech/0.2.0a9/

1 Like
#8

Great. Looks like I should really brush up on my pip. This is so obvious it’s killing me.

Unfortunately I’m still having the problem with generating the trie. But I will leave that for another post. Going to retrace my creation steps and do some trial and error. Thanks!

#9

But if you don’t give more details, we cannot help you.

#10

It’s the same error I had in a different thread and you postulated it was a version mismatch causing it. Now that I’ve eliminated that I’m going to ensure my pipeline for language model creation is working and then document it.

If I can’t solve it I will share the details. Even if I do solve it I’ll post the solution.

Robert Ritz

1 Like
#11

Sure, don’t hesitate to share any progress :slight_smile: