Issue with Plugins of speechtotext
Hi , I am trying to use speech to text plugin in my project and I downloaded deepspeech weight from macro.I have created folder containing audio file but I am getting the issue attached here.Please find let me know possible solutions Thank you !
Best Answer
-
Hi,
Thanks for your interest in this plugin.
The latest version of the DeepSpeech model requires a recent version of libstdc++6, as indicated by the error message. To upgrade libstdc++6, you will need to have admin access to the underlying server hosting DSS. If you have this access, and can confirm which Linux version you are running, I can guide you with these operations.
Alternatively, you can use an older version of the DeepSpeech model using the previous version of this plugin: [EDIT] https://update.dataiku.com/dss/8.0/plugins/speech-to-text-cpu/0.1.0/dss-plugin-speech-to-text-cpu-0.1.0.zip
Note that in that case, you will need to install the plugin manually from the ZIP archive: https://doc.dataiku.com/dss/latest/plugins/installing.html#installing-from-a-zip-file. You will no longer be able to benefit from the latest updates from the Store.
Please let me know which direction you'd like to take, and I'll be able to advise.
Cheers,
Alex
Answers
-
Thank you Alex, I am first wanna try with older version by manual download of that plugin if again the issue continues let you know!
-
Hi Alex,
I am need your support to extract the zip file. to proceed with your help.I am not able to go with your solution now.Let me know if you are possible with any chat in mail gmeet or any?
-
Not able to downlaod manually
Note that in that case, you will need to install the plugin manually from the ZIP archive: https://doc.dataiku.com/dss/latest/plugins/installing.html#installing-from-a-zip-file. You will no longer be able to benefit from the latest updates from the Store.
-
Hi,
Can you clarify what is not working? Are you able to download the ZIP file?
If so, you need to uninstall the existing plugin, and install it again from the ZIP file by following the documentation I linked to in my previous post: https://doc.dataiku.com/dss/latest/plugins/installing.html#installing-from-a-zip-file
Best regards,
Alex
-
Hi ,
Now downloaded the file manually but in running the model again not came the output successfully.PFA
in the Speechmacro they able see the models and sample.But I am not able to import that!
-
Hi, Give your time to discuss to explore in these plugin for 10mins Need your help.I created one project>created two folder>managed>speech/macro(where I add my audio file ml.wav in both the folder ) > and use the plugin to interlink and perform the run it giving the error instead of translation as text
-
Hi ,
Please let me know step by step procedure can explore sentimental analysis -> speech to text plugin with free trial .
Please share if you have any video/document or any demo
-
Hi,
After investigation, I confirm that the libstdc++6 requirements also applies to the previous version of the plugin. Could you please try again with the 0.1 version?
The procedure to uninstall the previous version and re-install from the ZIP archive is the same as my previous article.
We will evaluate how to reduce the system requirements of this plugin in a future plugin release.
Best regards,
Alex
-
Hi,
Even Tried with other version of 5.1 and all available versions.Now come with this error attached.PFA.
-
Hi,
I am afraid screenshots are not sufficient to diagnose the issue. Could you please send us a job diagnosis by following this procedure?
If the resulting file is too large for mail (> 15 MB), you can use https://dl.dataiku.com to send it to us. Please don't forget to send the link that is generated when you upload the file.
Best regards,
Alex
-
Hi Alex,
I uploaded my job diagnosis as in video format for better understanding! Please have a look in the Plik (dataiku.com) dashboard. the file in the name of after manual installation of plugin_Bala.zip and with existing plugin_Bala.mp4 .Please let me know if you got that
-
I am afraid I cannot use a video to diagnose the problem. Please send the job diagnosis ZIP archive.
-
https://dl.dataiku.com/file/BeCypG76E26Kwbkn/ZxYB4X8O1vS9JFN5/with%20existing%20plugin_Bala.mp4
please check
and i don't know how to perform job diagnosis than this method
-
To generate a job diagnosis, please follow the steps highlighted on this page (linked in my previous message): https://doc.dataiku.com/dss/latest/troubleshooting/problems/job-fails.html#getting-a-job-diagnosis
- Go to the page of the affected job
- Click on Actions > Download job diagnosis (top right of your screen)
- This will download the job diagnosis to your local machine
Please upload this job diagnosis to dl.dataiku.com and please don't forget to send the link that is generated when you upload the file.
-
Please find the attached Job Diagnosis and let me know how to fix the same.Thank you for supporting! attached JD below
-
Hi,
Thanks. Reading the job diagnosis, it looks like the code-env of the plugin has not been updated.
Could you navigate to the Speech-to-Text plugin page and create a new code-env from this screen?
Click on "Create a new one".
Then, please try to run the same recipe again. If it fails, please send us the new job diagnosis.
Best regards,
Alex
-
Hi Alex,
I am updating the code env and send you the JD again
-
Hi Please find the updated or created new code env work and corresponding JD is here attached
-
Hi,
Reading the job diagnosis, it looks like that the 0.1 version of the plugin has not been installed.
Could you please make sure again that you have followed all the steps from my previous messages, by replaying the following procedure:
0. Verify that your DSS version is 8.0.X (? > About)
1. Uninstall your current Speech-to-Text plugin
2. Reinstall the plugin using this ZIP archive: https://update.dataiku.com/dss/8.0/plugins/speech-to-text-cpu/0.1.0/dss-plugin-speech-to-text-cpu-0.1.0.zip
3. Create a new code-env for the plugin after installation
4. Re-run your test project to validate the correct behavior of the plugin
Please re-run the "Download DeepSpeech model" macro before re-running the "Speech to text" recipe.
---
If you still get a failure after following all of these steps, I would ask you to provide all the following elements:
- The failed job diagnosis: https://doc.dataiku.com/dss/latest/troubleshooting/problems/job-fails.html#getting-a-job-diagnosis
- The code-env creation diagnosis:
- An export of the test project you are using:
Best regards,
Alex
-
Hi Alex,
I've tried the plugin and it's not working. Read through all comments and different methods mentione above to solve the issue without any success. Any chance you can check job diagnosis file?
-
Hi,
The job diagnosis indicates that the underlying library Deepspeech requires libstdc++6, as indicated here: https://deepspeech.readthedocs.io/en/latest/SUPPORTED_PLATFORMS.html#linux-amd64-without-gpu.
I understand that you are running the latest version of the plugin (0.3) which relies on Deepspeech 0.9.1.
One option is to revert to an older version of the plugin, with an older version of Deepspeech, which works with older versions of libstdc++6. You can find this version here: https://update.dataiku.com/dss/8.0/plugins/speech-to-text-cpu/0.1.0/dss-plugin-speech-to-text-cpu-0.1.0.zip
Then, you can follow the steps of this post: https://update.dataiku.com/dss/8.0/plugins/speech-to-text-cpu/0.1.0/dss-plugin-speech-to-text-cpu-0.1.0.zip to reinstall.
Otherwise, you will need admin (sudo) access to the server hosting Dataiku DSS in order to update libstdc++. If you have this access and would like to pursue this quest, I would kindly ask which operating system the server runs on to guide you through these operations.
Cheers,
Alex
-
Hi again,
I've tried the older version now with some other errors.
I've added error log files and a short step explanation what I tried to get this to work. But without success. To answer you question, I am running DSS through Virtual Box.
Thanks in advance for any help, looking fwd to get this working.
