Open algoritmist opened 1 month ago
"The system should be running on Linux..." I think it is actually not necessary, bare metal also seems to be fine
I suggest to use LTE with speech-to-text API instead of offline self-trained NN. The device won't need much computing power and will be much more power efficient and small. The only disadvantage that I see is that APIs are not free and not all of them support Chinese. Right now I am researching the APIs, so if you like the idea you can help me with it. Right now I am looking at assemblyAI
For locally running models we can use Whisper or RealTimeSTT, or deploy it on a server and connect to it from the device
The aim of this project is to create a portable device that provides a convenient way for communication between people speeking different languages.
Functional requirements:
User interface requirments:
Hardware requirements:
Software requirements:
Features that could be added in future:
Wheter should we should support offline translation and what type of Internet access to use (wifi or mobile network) is a subject of discussion.
Suggestions and thoughts are welcome