https://github.com/ai4bharat/elaichi
ELAICHI: Enhancing Low-resource TTS by Addressing Infrequent and Low-frequency Character Bigrams
Science Score: 13.0%
This score indicates how likely this project is to be science-related based on various indicators:
-
○CITATION.cff file
-
✓codemeta.json file
Found codemeta.json file -
○.zenodo.json file
-
○DOI references
-
○Academic publication links
-
○Academic email domains
-
○Institutional organization owner
-
○JOSS paper metadata
-
○Scientific vocabulary similarity
Low similarity (7.1%) to scientific vocabulary
Repository
ELAICHI: Enhancing Low-resource TTS by Addressing Infrequent and Low-frequency Character Bigrams
Basic Info
- Host: GitHub
- Owner: AI4Bharat
- License: cc0-1.0
- Default Branch: master
- Size: 9.77 KB
Statistics
- Stars: 1
- Watchers: 5
- Forks: 0
- Open Issues: 0
- Releases: 0
Metadata Files
README.md
☕ ELAICHI [WIP]
ELAICHI: Enhancing Low-resource TTS by Addressing Infrequent and Low-frequency Character Bigrams
Abstract
Recent advancements in Text-to-Speech (TTS) technology have led to natural-sounding speech for English, primarily due to the availability of large-scale, high-quality web data. However, many other languages lack access to such resources, relying instead on limited studioquality data. This scarcity results in synthesized speech that often suffers from intelligibility issues, particularly with low-frequency character bigrams. In this paper, we propose three solutions to address this challenge. First, we leverage high-quality data from linguistically or geographically related languages to improve TTS for the target language. Second, we utilize low-quality Automatic Speech Recognition (ASR) data recorded in non-studio environments, which is refined using denoising and speech enhancement models. Third, we apply knowledge distillation from large-scale models using synthetic data to generate more robust outputs. Our experiments with Hindi demonstrate significant reductions in intelligibility issues, as validated by human evaluators. We propose this methodology as a viable alternative for languages with limited access to high-quality data, enabling them to collectively benefit from shared resources.
Getting Started
We use the VITS model as implemented by jaywalnut/vits, with an extended vocabulary to accommodate Indian scripts and additional emotion embeddings. Note that all models released in this work have been trained with a neutral emotion label.
Model Checkpoints
All model checkpoints are public and available here
Owner
- Name: AI4Bhārat
- Login: AI4Bharat
- Kind: organization
- Email: opensource@ai4bharat.org
- Location: India
- Website: https://ai4bharat.org
- Twitter: AI4Bharat
- Repositories: 37
- Profile: https://github.com/AI4Bharat
Artificial-Intelligence-For-Bhārat : Building open-source AI solutions for India!
GitHub Events
Total
- Issues event: 1
- Watch event: 7
- Member event: 1
- Push event: 3
- Create event: 2
Last Year
- Issues event: 1
- Watch event: 7
- Member event: 1
- Push event: 3
- Create event: 2
Issues and Pull Requests
Last synced: 10 months ago
All Time
- Total issues: 1
- Total pull requests: 0
- Average time to close issues: N/A
- Average time to close pull requests: N/A
- Total issue authors: 1
- Total pull request authors: 0
- Average comments per issue: 0.0
- Average comments per pull request: 0
- Merged pull requests: 0
- Bot issues: 0
- Bot pull requests: 0
Past Year
- Issues: 1
- Pull requests: 0
- Average time to close issues: N/A
- Average time to close pull requests: N/A
- Issue authors: 1
- Pull request authors: 0
- Average comments per issue: 0.0
- Average comments per pull request: 0
- Merged pull requests: 0
- Bot issues: 0
- Bot pull requests: 0
Top Authors
Issue Authors
- ninurFarm (1)