Memo Feedback
Feedback
Guidelines
If you are encountering issues related to downloading, conversion, or translation failure, you may first refer to the Frequently Asked Questions below to solve your problem."
- Github: https://github.com/Makememo/MemoAI/issues
- Email: feedback@memo.ac
Frequently Asked Questions
Why can't I download the model, always stays at 0?
- Go to Model Download to download and import it yourself.
- Please proxy huggingface.co
Why is the output content blank?
If an errorTypeError: Cannot read properties of undefined (reading 'st') is reported, it may be the following problems:
- The installation path of Memo has Chinese characters, generally for Windows devices.
- The device has 8GB of memory and uses a large model for conversion.
Why is the output content repeated?
Repetitive outputs usually occur with long blank space or background noise.
- Output prompt: This is a song, please ignore the blank segments and background noise, and focus on the problem in the output.
- Please change the model or manually trim off the noise or blank segments.
Currently, there is no good technical solution, and we need to wait for improvements in the AI model.
Why can't I use AI to summarize?
Please switch to the English interface, or use Wenxin Yiyuan. Note that non-Mainland China services need to go to Settings - Proxy Settings to turn on the system proxy.
Why can't I translate or why is there a translation error?
- Google Translate, AI translation, and other non-Mainland China services need to go to Settings - Proxy Settings to turn on the system proxy.
- Please check if there is any problem with your custom Key, such as expiration, arrears, no word quota, etc.
- Is Microsoft Translator translating too long content? If so, please switch to Volcano Translate to translate long content.
Why did the conversion process get stuck?
Choosing Medium or Large models with too small memory will result in no available memory for inference. We recommend switching to a smaller model.
A JavaScript error ocurred in the main progress error occurs after Windows is installed
Please install this component: Microsoft Visual C++ Redistributable, scroll down and install the x86 or x64 version.
Installation Issues
Some Common Problems
- There is a repetition in the output, please switch to a different model or manually trim off blank sections. Currently, there is no good technical solution, and we need to wait for improvements in the AI model.
- After opening, you can't output any text, and there is a blank screen. Confirm whether your machine has 16GB of memory and above, and if so, confirm that you have not selected the wrong transcription language. If not, please use a smaller model.
Why can't I open it after installation?
For Mac, the minimum compatible system version is 11, and the recommended version is 13 and above. Win requires Win10 and above, and also check whether you used a Chinese path.
What should I do if there is a Windows runtime error?
If the error "A JavaScript error ocurred in the main progress" occurs during Windows installation, please install this component: Microsoft Visual C++ Redistributable latest supported downloads, scroll down, and install the x86 or x64 version.
If the installation is normal:
- Please run in administrator mode. Or check if it is an ARM architecture chip. Currently, ARM chips for Windows are not supported.
- Please check if the model's file path contains Chinese characters.
Why can't I install Windows by double-clicking?
Generally, the security software has turned on firewall mode, please turn off the firewall.
Speech Recognition Issues
Why can't YouTube be downloaded and local videos cannot be used?
If you are an M series chip user, please install Rosetta. If not, please check the system version, macOS 12, Win 10 and above.
Why is the conversion quality poor?
For better conversion quality, please select the Mediun or Large model. Correspondingly, the speed will be slow.
Why can't it accurately segment sentences?
Please use a medium or larger model + prompt: Each identification waits for the discussion to end instead of breaking in the middle to achieve segmentation. It was provided in the group by @Xinbao.
Why is the output content repeated?
This is a model illusion problem. We will use VAD to solve it in the future, please be patient.
Download & Translation Issues
Failed to Get Media Information
Please go to settings, you need to go to settings - Proxy Settings - Turn on System Proxy, if it still doesn't work, please set the proxy link and port.
Why can't I download content or why is there a translation error?
Please go to Settings, clear the proxy in the Proxy Settings, and then restart. Note: The cache capacity of the current test version has not iterated, and the upgrade to the new installation package will be automatically cleaned, please export the relevant conversion content in time.
Why is the translation error?
This is a Google Translate capability issue, proxy issue. It is recommended to check the proxy configuration in the settings as well as the proxy tool. Google Translate server testing tool: https://github.com/GoodCoder666/GoogleTranslate_IPFinder
The reason for the Open AI translation error is that it is limited to 3/min and a maximum of 7000 tokens can be entered at a time.
Submit Feedback to Memo
Please go to Memo AI Feedback to submit.