# Human-1 by Josh Talks: A Full-Duplex Conversational Modeling Framework in Hindi using Real-World Conversations

Source: [Human-1 by Josh Talks: A Full-Duplex Conversational Modeling Framework in Hindi using Real-World Conversations](<https://www.sota2.com/research/paper/69f03eea55a3c32af89f9eed>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-05-27T05:00:13.467Z

Authors: Bhaskar Singh, Shobhit Banga, Mahima Manik, Pranav Sharma

Published: 2026-04-25T13:18:40.000Z

## Abstract

Full-duplex spoken dialogue systems can model natural conversational behaviours such as interruptions, overlaps, and backchannels, yet such systems remain largely unexplored for Indian languages. We present the first open, reproducible full-duplex spoken dialogue system for Hindi by adapting Moshi, a state-of-the-art duplex speech architecture, using a custom Hindi tokeniser and training on 26,000 hours of real spontaneous conversations collected from 14,695 speakers with separate speaker channels, enabling direct learning of turn-taking and overlap patterns from natural interactions. To support Hindi text generation, we replace the original English tokeniser and reinitialise text-vocabulary-dependent parameters while retaining the pre-trained audio components. We propose a two-stage training recipe -- large-scale pre-training followed by fine-tuning on 1,000 hours of conversational data. Evaluation through the prompted dialogue continuation paradigm with both automatic metrics and human judgments demonstrates that the resulting model generates natural and meaningful full-duplex conversational behaviour in Hindi. This work serves as a first step toward real-time duplex spoken dialogue systems for Hindi and other Indian languages.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Turn-taking Analysis on Hindi Telephone Conversations (test)](<https://www.sota2.com/research/sota/turn-taking-analysis-on-hindi-telephone-conversations-test.md>): Turn-taking Analysis; Hindi Telephone Conversations (test); IPU Duration: 38.9.

- [Language Modeling on Hindi conversational Whisper-v3 transcriptions (test)](<https://www.sota2.com/research/sota/language-modeling-on-hindi-conversational-whisper-v3-transcriptions-test.md>): Language Modeling; Hindi conversational Whisper-v3 transcriptions (test); Perplexity (PPL): 356.9.

- [Human Evaluation on Hindi spoken dialogue (test)](<https://www.sota2.com/research/sota/human-evaluation-on-hindi-spoken-dialogue-test.md>): Human Evaluation; Hindi spoken dialogue (test); Naturalness: 4.1.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2604.23295>)

[PDF](<https://arxiv.org/pdf/2604.23295.pdf>)