NVIDIA/personaplex là dự án mã nguồn mở trên GitHub, viết chủ yếu bằng Python, với 12.1k sao. PersonaPlex code.
Tóm tắt dựng từ metadata GitHub của chính dự án — chưa có bài review TopGit. Trang sẽ tự động cập nhật khi bài review đầy đủ được xuất bản.
VÌ SAO CHƯA CÓ REVIEW
TopGit viết bài đầy đủ cho repo có nhiều sao nhất và được yêu cầu nhiều nhất. Trang này là snapshot trong thời gian chờ — xem README gốc ở tab READ ME.
PersonaPlex: Voice and Role Control for Full Duplex Conversational Speech Models
PersonaPlex is a real-time, full-duplex speech-to-speech conversational model that enables persona control through text-based role prompts and audio-based voice conditioning. Trained on a combination of synthetic and real conversations, it produces natural, low-latency spoken interactions with a consistent persona. PersonaPlex is based on the Moshi architecture and weights.
NEW: the system prompt (voice + text) now stays in the model's context for the whole conversation instead of dropping out of context. The prompt also acts as the attention sink. This enabled endless stable conversations while preserving voice and role. However, conversation memory is still limited to a sliding window of the last ~2.5 minutes ((2048-system_prompt_frames)/12.5). Also the system prompt is loaded in one prefill instead of stepping though token-by-token. Both are on by default; see Endless Conversations.
CPU Offload: If your GPU has insufficient memory, use the --cpu-offload flag to offload model layers to CPU. This requires the accelerate package (pip install accelerate):
Access the Web UI from a browser at localhost:8998 if running locally, otherwise look for the access link printed by the script:
Access the Web UI directly at https://11.54.401.33:8998
Endless Conversations
PersonaPlex attends to a context of 2048 frames. The server and offline script now keep the system prompt pinned in that context for the whole conversation, and compute it in a single pass at the start of each session. Both moshi.server and moshi.offline accept:
Flag
Default
Description
--pin-system-prompt / --no-pin-system-prompt
on
Keep the system prompt in context for the whole conversation; the rest of the context keeps sliding.
Load the system prompt in one batched pass instead of frame by frame.
--max-pinned-frames
512
Longest system prompt that can be pinned, in frames (12.5 per second). Set it to the longest system prompt that you would expect the server to encounter. There is no issue if the system prompt is shorter than this value.
Each session logs the prompt length, for example Pinned 176 system prompt frames (max_pinned_frames=512, sliding window 2824 frames). A prompt takes 12.5 frames per second of voice prompt plus one per text token, plus 11; set --max-pinned-frames a little above your longest prompt. Pinned frames come out of the 2048-frame window. python -m moshi.bench_pinning measures the overhead on your GPU.
Offline Evaluation
For offline evaluation use the offline script that streams in an input wav file and produces an output wav file from the captured output stream. The output file will be the same duration as the input file.
Add --cpu-offload to any command below if your GPU has insufficient memory (requires accelerate package). Or install cpu-only PyTorch for offline evaluation on pure CPU.
PersonaPlex supports a wide range of voices; we pre-package embeddings for voices that sound more natural and conversational (NAT) and others that are more varied (VAR). The fixed set of voices are labeled:
The model is trained on synthetic conversations for a fixed assistant role and varying customer service roles.
Assistant Role
The assistant role has the prompt:
You are a wise and friendly teacher. Answer questions or provide advice in a clear and engaging way.
Use this prompt for the QA assistant focused "User Interruption" evaluation category in FullDuplexBench.
Customer Service Roles
The customer service roles support a variety of prompts. Here are some examples for prompting style reference:
You work for CitySan Services which is a waste management and your name is Ayelen Lucero. Information: Verify customer name Omar Torres. Current schedule: every other week. Upcoming pickup: April 12th. Compost bin service available for $8/month add-on.
You work for Jerusalem Shakshuka which is a restaurant and your name is Owen Foster. Information: There are two shakshuka options: Classic (poached eggs, $9.50) and Spicy (scrambled eggs with jalapenos, $10.25). Sides include warm pita ($2.50) and Israeli salad ($3). No combo offers. Available for drive-through until 9 PM.
You work for AeroRentals Pro which is a drone rental company and your name is Tomaz Novak. Information: AeroRentals Pro has the following availability: PhoenixDrone X ($65/4 hours, $110/8 hours), and the premium SpectraDrone 9 ($95/4 hours, $160/8 hours). Deposit required: $150 for standard models, $300 for premium.
Casual Conversations
The model is also trained on real conversations from the Fisher English Corpus with LLM-labeled prompts for open-ended conversations. Here are some example prompts for casual conversations:
You enjoy having a good conversation.
You enjoy having a good conversation. Have a casual discussion about eating at home versus dining out.
You enjoy having a good conversation. Have an empathetic discussion about the meaning of family amid uncertainty.
You enjoy having a good conversation. Have a reflective conversation about career changes and feeling of home. You have lived in California for 21 years and consider San Francisco your home. You work as a teacher and have traveled a lot. You dislike meetings.
You enjoy having a good conversation. Have a casual conversation about favorite foods and cooking experiences. You are David Green, a former baker now living in Boston. You enjoy cooking diverse international dishes and appreciate many ethnic restaurants.
Use the prompt You enjoy having a good conversation. for the "Pause Handling", "Backchannel" and "Smooth Turn Taking" evaluation categories of FullDuplexBench.
Generalization
Personaplex finetunes Moshi and benefits from the generalization capabilities of the underlying Helium LLM. Thanks to the broad training corpus of the backbone, we find that the model will respond plausibly to out-of-distribution prompts and lead to unexpected or fun conversations. We encourage experimentation with different prompts to test the model's emergent ability to handle scenarios outside its training distribution. As an inspiration we feature the following astronaut prompt in the WebUI:
You enjoy having a good conversation. Have a technical discussion about fixing a reactor core on a spaceship to Mars. You are an astronaut on a Mars mission. Your name is Alex. You are already dealing with a reactor core meltdown on a Mars mission. Several ship systems are failing, and continued instability will lead to catastrophic failure. You explain what is happening and you urgently ask for help thinking through how to stabilize the reactor.
License
The present code is provided under the MIT license. The weights for the models are released under the NVIDIA Open Model license.
Citation
If you use PersonaPlex in your research, please cite our paper:
@misc{roy2026personaplexvoicerolecontrol,
title={PersonaPlex: Voice and Role Control for Full Duplex Conversational Speech Models},
author={Rajarshi Roy and Jonathan Raiman and Sang-gil Lee and Teodor-Dumitru Ene and Robert Kirby and Sungwon Kim and Jaehyeon Kim and Bryan Catanzaro},
year={2026},
eprint={2602.06053},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2602.06053},
}
Trang TopGit này là một snapshot — tab "Readme" hiển thị nguyên văn README của repo (đã bỏ link, giữ ảnh). Repo GitHub ở github.com/NVIDIA/personaplex là nguồn chính thức.
NVIDIA/personaplex có bao nhiêu sao?
NVIDIA/personaplex có 12.1k sao GitHub — tải lại trang để xem số mới nhất, hoặc xem trực tiếp github.com/NVIDIA/personaplex. TopGit phản chiếu số sao của GitHub nhưng không cam kết đến từng phút.
NVIDIA/personaplex có phải mã nguồn mở không?
Có — NVIDIA/personaplex phát hành theo license MIT, nghĩa là mã nguồn mở để đọc, fork và (tùy license) tái sử dụng. Mã: github.com/NVIDIA/personaplex.
NVIDIA/personaplex còn đang phát triển không?
Commit gần nhất trên NVIDIA/personaplex là 8 tháng trước (theo timestamp GitHub). Repo có 1.6k fork — một chỉ báo về mức độ quan tâm của cộng đồng.
NVIDIA/personaplex dùng license gì?
NVIDIA/personaplex phát hành theo license MIT. Nên mở file LICENSE trên GitHub để xác nhận — license metadata đôi khi lệch với thực tế dự án.
NVIDIA/personaplex là gì?
NVIDIA/personaplex (NVIDIA/personaplex) là dự án Python trên GitHub. Theo mô tả gốc: PersonaPlex code.
NVIDIA/personaplex viết bằng ngôn ngữ gì?
NVIDIA/personaplex chủ yếu viết bằng Python. Trường "language" của GitHub dựa trên phần lớn byte ở nhánh mặc định.
Đọc đầy đủ README ở tab phía trên.
personaplex có đáng để bạn bỏ thời gian?
ChatGPT, Claude và Perplexity đều đọc được trang này. Hỏi thử xem họ nghĩ gì về personaplex.