Created by: thumarrushik

Hi,

I am facing a below-mentioned problem when evaluating pretrained models. Any help would be appreciated. Also, Thank you for providing DeepSpeech Pretrained Models.

Operating System : AWS Ubuntu 16.04

DeepSpeech/examples/tiny$ sudo sh run_test_golden.sh
Download language model ...
download_lm_en.sh: 8: [: 099a601759d467cd0a8523ff939819c5: unexpected operator
--2018-05-13 22:32:15--  http://paddlepaddle.bj.bcebos.com/model_zoo/speech/common_crawl_00.prune01111.trie.klm
Resolving paddlepaddle.bj.bcebos.com (paddlepaddle.bj.bcebos.com)... 103.235.46.61
Connecting to paddlepaddle.bj.bcebos.com (paddlepaddle.bj.bcebos.com)|103.235.46.61|:80... connected.
HTTP request sent, awaiting response... 416 Requested Range Not Satisfiable

    The file is already fully retrieved; nothing to do.

download_lm_en.sh: 20: [: 099a601759d467cd0a8523ff939819c5: unexpected operator
Download LibriSpeech model ...
download_model.sh: 8: [: 1f72d0c5591f453362f0caa09dd57618: unexpected operator
--2018-05-13 22:32:50--  http://cloud.dlnel.org/filepub/?uuid=117cde63-cd59-4948-8b80-df782555f7d6
Resolving cloud.dlnel.org (cloud.dlnel.org)... 202.108.23.203
Connecting to cloud.dlnel.org (cloud.dlnel.org)|202.108.23.203|:80... connected.
HTTP request sent, awaiting response... 200 OK
Cookie coming from cloud.dlnel.org attempted to set domain to baidu.com
Length: unspecified [application/x-tar]
Saving to: ‘./librispeech_model.tar.gz’

./librispeech_model.ta     [                    <=>   ] 159.29M  4.54MB/s    in 62s

2018-05-13 22:33:54 (2.58 MB/s) - ‘./librispeech_model.tar.gz’ saved [167029229]

download_model.sh: 20: [: 1f72d0c5591f453362f0caa09dd57618: unexpected operator
mean_std.npz
params.tar.gz
README.md
vocab.txt
-----------  Configuration Arguments -----------
alpha: 2.5
batch_size: 8
beam_size: 500
beta: 0.3
cutoff_prob: 1.0
cutoff_top_n: 40
decoding_method: ctc_beam_search
error_rate_type: wer
lang_model_path: models/lm/common_crawl_00.prune01111.trie.klm
mean_std_path: models/librispeech/mean_std.npz
model_path: models/librispeech/params.tar.gz
num_conv_layers: 2
num_proc_bsearch: 8
num_proc_data: 8
num_rnn_layers: 3
rnn_layer_size: 2048
share_rnn_weights: 1
specgram_type: linear
test_manifest: data/tiny/manifest.test-clean
trainer_count: 8
use_gpu: 0
use_gru: 0
vocab_path: models/librispeech/vocab.txt
------------------------------------------------
Illegal instruction (core dumped)
Failed in evaluation!

run_test_golden.sh file

DeepSpeech/examples/tiny$ cat run_test_golden.sh 
#! /usr/bin/env bash

cd ../.. > /dev/null

# download language model
cd models/lm > /dev/null
sh download_lm_en.sh
if [ $? -ne 0 ]; then
    exit 1
fi
cd - > /dev/null


# download well-trained model
cd models/librispeech > /dev/null
sh download_model.sh
if [ $? -ne 0 ]; then
    exit 1
fi
cd - > /dev/null


# evaluate model
CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 \
python -u test.py \
--batch_size=8 \
--trainer_count=8 \
--beam_size=500 \
--num_proc_bsearch=8 \
--num_proc_data=8 \
--num_conv_layers=2 \
--num_rnn_layers=3 \
--rnn_layer_size=2048 \
--alpha=2.5 \
--beta=0.3 \
--cutoff_prob=1.0 \
--cutoff_top_n=40 \
--use_gru=False \
--use_gpu=False \
--share_rnn_weights=True \
--test_manifest='data/tiny/manifest.test-clean' \
--mean_std_path='models/librispeech/mean_std.npz' \
--vocab_path='models/librispeech/vocab.txt' \
--model_path='models/librispeech/params.tar.gz' \
--lang_model_path='models/lm/common_crawl_00.prune01111.trie.klm' \
--decoding_method='ctc_beam_search' \
--error_rate_type='wer' \
--specgram_type='linear'

if [ $? -ne 0 ]; then
    echo "Failed in evaluation!"
    exit 1
fi


exit 0

Swig Decoders

DeepSpeech/examples/tiny$ pip show swig-decoders
---
Metadata-Version: 1.0
Name: swig-decoders
Version: 1.1
Summary: CTC decoders
Home-page: UNKNOWN
Author: UNKNOWN
Author-email: UNKNOWN
License: UNKNOWN
Location: /usr/local/lib/python2.7/dist-packages/swig_decoders-1.1-py2.7-linux-x86_64.egg
Requires: 
Classifiers:
You are using pip version 8.1.1, however version 10.0.1 is available.
You should consider upgrading via the 'pip install --upgrade pip' command.