Instructions to use nvidia/personaplex-7b-v1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Moshi
How to use nvidia/personaplex-7b-v1 with Moshi:
# pip install moshi # Run the interactive web server python -m moshi.server --hf-repo "nvidia/personaplex-7b-v1" # Then open https://localhost:8998 in your browser
# pip install moshi import torch from moshi.models import loaders # Load checkpoint info from HuggingFace checkpoint = loaders.CheckpointInfo.from_hf_repo("nvidia/personaplex-7b-v1") # Load the Mimi audio codec mimi = checkpoint.get_mimi(device="cuda") mimi.set_num_codebooks(8) # Encode audio (24kHz, mono) wav = torch.randn(1, 1, 24000 * 10) # [batch, channels, samples] with torch.no_grad(): codes = mimi.encode(wav.cuda()) decoded = mimi.decode(codes) - Notebooks
- Google Colab
- Kaggle
Personaplex and her Brazilian friends.
Oh my God! This is really fun! Since I don't speak English and the persona understands little Portuguese, the experience ended up being quite funny. I asked how I would get from Earth to the Moon… the advice was that I should go out and make friends! And, since the persona is a native of the US, she even said it would be difficult for her to make Brazilian friends. This is so much fun! When will we have an assistant who speaks Brazilian Portuguese?
Is there a way to transcribe what she says into Portuguese? Since she's a very talkative Persona, she always picks up on something I say in Portuguese and engages in a very pleasant, 'traveling' conversation. If there's a way to output the text in Portuguese, that would be very helpful.
Interesting! I don't have Portuguese conversation datasets so don't have immediate plans for Portuguese support in the model itself. But you can use a small multi-lingual LLM like tiny-aya (https://huggingface.co/CohereLabs/tiny-aya-global) to keep converting the text at the end of each sentence to Portuguese and displaying that on the UI.