seamless communication on raspberry pi 4 or Nvidia Jetson

[kiashann]

Is it possible to use pretrained model for inference on raspberry pi 4 or Nvidia Jetson?
What is the latest hardware requirements to use on boards?
Anyone can help me, or do you have such experience to use pretrained model like this on boards?

[Juan1108888]

I also want to know!
For small models, maybe try

[astrowar]

I tested it on a jetson orin nano. The medium and large models exceed the amount of memory available.
The small model takes 47s to translate a 5s audio and uses 2GB (inference only, the model loads is fast < 7s). The audio output has good quality. The translation gets a little sluggish on the first word in some tests. but the overall result is good in the two languages I tested.
To work you need to add the resample to 16Khz in the input

[Juan1108888]

How ASR and T2ST?

[cndn]

Hey @kiashann - we don't have active plan to support specific hardwares. If you're able to get our project working on a new hardware platform or have insights on improvements, would be appreciated if you could contribute knowledge!

[sytpb]

how about the raspberry PI5 , focus on it .