Steps to try out DeepSpeech with pre-release 0.5.0a11 model

Steps.md

Downloaded the DeepSpeech GPU wheel for my OS + intended version of Python (3.7)

https://github.com/mozilla/DeepSpeech/releases/download/v0.5.0-alpha.11/deepspeech_gpu-0.5.0a11-cp37-cp37m-manylinux1_x86_64.whl

Downloaded the model (as per link from https://discourse.mozilla.org/t/deep-speech-0-5-0-pre-release-model-checkpoint/41236) and expanded files

https://drive.google.com/file/d/1vNokDsmUdwP9rk1c68JUOS0SCO5lHv3m/view?usp=sharing

Conda environment for DeepSpeech GPU

conda create -n deepspeech_051a11 python=3.7

conda activate deepspeech_051a11

conda install cudatoolkit=10.0 numpy

pip3 install 'tensorflow-gpu==1.13.1'

pip install /home/neil/main/Projects/DeepSpeech-Jun-2019/pre-built_binaries/deepspeech_gpu-0.5.0a11-cp37-cp37m-manylinux1_x86_64.whl

cd /home/neil/main/Projects/DeepSpeech-Jun-2019/models/

deepspeech --model deepspeech-0.5.0-models/output_graph.pbmm --alphabet deepspeech-0.5.0-models/alphabet.txt --lm deepspeech-0.5.0-models/lm.binary --trie deepspeech-0.5.0-models/trie --audio ../test_wavs/p225_27282.wav

deepspeech --model deepspeech-0.5.0-models/output_graph.pbmm --alphabet deepspeech-0.5.0-models/alphabet.txt --lm deepspeech-0.5.0-models/lm.binary --trie deepspeech-0.5.0-models/trie --audio ../test_wavs/p225_27283.wav

deepspeech --model deepspeech-0.5.0-models/output_graph.pbmm --alphabet deepspeech-0.5.0-models/alphabet.txt --lm deepspeech-0.5.0-models/lm.binary --trie deepspeech-0.5.0-models/trie --audio ../test_wavs/p225_27280.wav