Closed joewale closed 2 years ago
@joewale I haven't tested tlinux. So you've installed trition inference server on tlinux (腾讯云服务器) and installed ctc decoder successfully? But you see an error after triton inference server ?
Could you give more info about the environment ?
BTW, why not use docker to deploy? Is there any concern or requirement ? Thanks!
Thank you for quick reply. The info about the environment as follows:
Linux 0484488c97c7 5.4.32-1-tlinux4-0001 #1 SMP Tue Jul 21 14:24:37 CST 2020 x86_64 x86_64 x86_64 GNU/Linux
Linux version 5.4.32-1-tlinux4-0001 (root@3c7067f9a859) (gcc version 8.3.1 20191121 (Red Hat 8.3.1-5) (GCC)) #1 SMP Tue Jul 21 14:24:37 CST 2020
cmake version 3.18.2
Python 3.8.3
I use the docker to deploy, but use the docker images with tlinux system as base docker image.
Hi, @Slyne , the log of installation is given above, please help to see if there is a problem in my installation. Looking forward to your reply, thank you!
The error in tritonserver as follows
[I0324 02:36:09.087216 1526 pb_stub.cc:402] Failed to process the request(s) for model 'scoring_0_0', message: TypeError: ctc_beam_search_decoder_batch() takes no keyword arguments
At:
/home/speech_asr_async_svr/model_repo/scoring/1/model.py(181): execute
Hi, @Slyne , the log of installation is given above, please help to see if there is a problem in my installation. Looking forward to your reply, thank you!
The error in tritonserver as follows
[I0324 02:36:09.087216 1526 pb_stub.cc:402] Failed to process the request(s) for model 'scoring_0_0', message: TypeError: ctc_beam_search_decoder_batch() takes no keyword arguments At: /home/speech_asr_async_svr/model_repo/scoring/1/model.py(181): execute
Is it possible to share the Dockerfile ?
H
ctc_decoder_install.log Hi, @Slyne , the log of installation is given above, please help to see if there is a problem in my installation. Looking forward to your reply, thank you! The error in tritonserver as follows
[I0324 02:36:09.087216 1526 pb_stub.cc:402] Failed to process the request(s) for model 'scoring_0_0', message: TypeError: ctc_beam_search_decoder_batch() takes no keyword arguments At: /home/speech_asr_async_svr/model_repo/scoring/1/model.py(181): execute
Is it possible to share the Dockerfile ?
Hi, @Slyne . Is it possible to contact with wechat? my wechat: joewale
Remove
blank_id=
space_id=
...
in model.py will solve this issue.
Confirmed with @joewale
Thanks very much!
Hi, Slyne,
I follow your example for wenet in the folder : runtime/server/x86_gpu. And I run the command: tritonserver --model-repository=model_repo,The error appears when run the above command in the tlinux system, while successfully in ubuntu system. What is the problem? Looking forward to your reply, thank you!
0315 02:18:17.627657 8077 pb_stub.cc:402] Failed to process the request(s) for model 'scoring_0_1', message: TypeError: ctc_beam_search_decoder_batch() takes no keyword arguments