Entonytang

Can you tell me the methods you compile kaldi.......(aslo means : how can you get the file like kaldi_matrix.a.........) [email protected] From: Dzmitry Bahdanau Date: 2015-11-20 23:34 To: rizar/attention-lvcsr CC: Entonytang...

After change the configure command, problem solved...... this steps: $LVSR/bin/run.py train wsj_paper6 $LVSR/exp/wsj/configs/wsj_paper6.yaml this default configuration trains model using CPU.......how to use GPU instead.......

After adding "device =gpu3" while I find GPU Process in GPU 2(device K40).....using default wsj_paper6.yaml..... it costs 65 seconds per steps(1 epoch = 3700 steps), I think this speed is...

thanks, solved..... while at the 830 steps. the program stoped without any warnings......while GPU Process is still there.......the bokeh-server is also there. and wsj_paper6.yaml doesn't seems to be the setting...

I don't think so, I use another core and try again. the result is similar...... (The best_valid_sequence_log_likelihood: 503.460199693 which is same as the result after 830 steps.) while only pretraining_model.zip|...

I know. I use the command: $LVSR/exp/wsj/decode.sh wsj_paper6 valid 200...... and I hope you can give some descriptions about the config file like wsj_paper.yaml......

I got the same error. ValueError: Shape must be rank 4 but is rank 3 for 'Conv2D' (op: 'Conv2D') with input shapes: [757,100,39], [41,11,1,32].