Deepspeech c windows error incompatible model

[JRMeyer]

>>> EternalAmbiguity
[August 15, 2019, 4:10am]

I downloaded the 0.5.1 nuget package for DeepSpeech, then set up a WPF
project similar to the .Net example in the source. I also downloaded the
0.5.1 models from the appropriate release. I try to initialize with
something like:

ds = new DeepSpeechClient.DeepSpeech(); slash
ds.CreateModel('C: slash slash Users??? slash slash Desktop slash slash Deep
Speech slash slash deepspeech-0.5.1-models slash slash output_graph.pbmm', N_CEP, N_CONTEXT,
'C: slash slash Users??? slash slash Desktop slash slash Deep
Speech slash slash deepspeech-0.5.1-models slash slash alphabet.txt', BEAM_WIDTH);

When I hit that piece of the code I get a System.ArgumentException
'Error incompatible model'. I'm not sure what this error means.

The model files are in the appropriate folder -

20.9

Any tips?

[This is an archived TTS discussion thread from discourse.mozilla.org/t/deepspeech-c-windows-error-incompatible-model]

[JRMeyer]

>>> kdavis
[August 15, 2019, 9:49am]

As you're the windows
expert, do you have any ideas?

[Archived Post]

[JRMeyer]

>>> lissyx
[August 15, 2019, 10:14am]



> When I hit that piece of the code I get a System.ArgumentException
> 'Error incompatible model'. I'm not sure what this error means.

Can you share explicit stderr?

That error means you are loading a different model version than what is
supported by the binary ...

[Archived Post]

[JRMeyer]

>>> lissyx
[August 15, 2019, 10:15am]

github.com

#### mozilla/DeepSpeech/blob/v0.5.1/native_client/deepspeech.cc slash #L714

<br/> return DS_ERR_FAIL_CREATE_SESS;<br/> }<br/> <br/> int graph_version = model->graph_def.version();<br/> if (graph_version < DS_GRAPH_VERSION) {<br/> std::cerr << 'Specified model file version (' << graph_version << ') is '<br/> << 'incompatible with minimum version supported by this client ('<br/> << DS_GRAPH_VERSION << '). See '<br/> << 'https://github.com/mozilla/DeepSpeech/#model-compatibility '<br/> << 'for more information' << std::endl;<br/> return DS_ERR_MODEL_INCOMPATIBLE;<br/> }<br/> <br/> for (int i = 0; i < model->graph_def.node_size(); ++i) {<br/> NodeDef node = model->graph_def.node(i);<br/> if (node.name() == 'input_node') {<br/> const auto& shape = node.attr().at('shape').shape();<br/> model->n_steps = shape.dim(1).size();<br/> model->n_context = (shape.dim(2).size()-1)/2;<br/> model->n_features = shape.dim(3).size();<br/> model->mfcc_feats_per_timestep = shape.dim(2).size() DEEPSPEECH.cdx deepspeech.commands DEEPSPEECH.pages DEEPSPEECH.warc.gz discourse.mozilla.org html-to-markdown.sh shell-conver-html-to-split-posts.sh sorted-deepspeech-posts shape.dim(3).size();<br/>

[Archived Post]

[JRMeyer]

>>> EternalAmbiguity
[August 15, 2019, 1:05pm]

Okay, I think I figured it out. I previously tried to use the latest
0.6.0 alpha 4 package but ran into trouble. I went back to the 0.5.1
nuget package but ran into the incompatible model issue.

I just cleaned and rebuilt my visual studio solution and it works.
Thanks for all the help.

[Archived Post]