comp Releases
321 releases of trycompai/comp
- v2.16.0
Major update adding DeepSeek API integration with streaming reasoning support, MCP protocol support, and expanded AI model options including SiliconFlow provider.
Feb 16, 2025
- v3.2.3
Adds configuration flexibility by allowing environment variables to be set via file reading, following the `<variable_name>_FILE` pattern.
Feb 6, 2025
- v3.2.2
Minimal release notes: "- Use `better-sqlite3` on the `.example.env` and also add `DB_FILENAME`. (ec5a99f)"
Jan 24, 2025
- v3.2.1
Jan 22, 2025
- v1.3.2
Improves network filtering and CLI usability with dropdown menus, adapter selection, and enhanced chart time values.
Jan 6, 2025
- v2.15.8v2.15.8 Support OpenAI Realtime API
Adds support for the OpenAI Realtime API and introduces new features for custom model names and Claude 3.5 Haiku.
Nov 11, 2024
- v2.15.7v2.15.7 Add Claude haiku 3.5 and Support ChatGLM
Introduces Claude 3.5 Haiku and ChatGLM support, along with fixes for Gemini and status text issues.
Nov 5, 2024
- v2.15.6v2.15.6 XAI support
Adds XAI support and fixes Tauri fetch implementation issues.
Oct 24, 2024
- v3.1.2GUI for precise control
Introduces a new Graphical User Interface (GUI) for precise control over utterance generation, allowing users to generate multiple realizations and modify pitch/duration via drag-and-drop, with support for voice swapping across 7000+ languages.
Oct 7, 2024
- v3.1.1Improved Vocoder through end-to-end Training
Focuses on audio quality improvement by training the vocoder end-to-end on synthetic spectrograms and increasing its capacity, resulting in significant quality gains without slowing down inference speed.
Sep 22, 2024
- v3.1Improved TTS in 7000 Languages
Improves the TTS model for 7000 languages by making prosody prediction stochastic and expanding support for IPA modifiers and language pretraining data.
Jul 25, 2024
- v3.0TTS in 7000 Languages
A major release introducing support for over 7000 languages through extrapolation from a 462-language model, alongside improved TTS quality and watermarking features.
Jun 10, 2024
- 2.pPrompting Controlled Emotional TTS
Adds functionality for controlled emotional TTS, enabling users to condition models on emotional prompts during training and transfer emotion to synthesized speech during inference.
Jun 10, 2024
- v1.11.1v1.11.1 - BaseUrl Hotfix
This hotfix resolves BaseUrl issues, allowing the application to function correctly on servers using subpaths or base URLs (e.g., example.com/jellyfin).
Jun 8, 2024
- v1.11.0v1.11.0 - External Web Client
The application architecture is updated to use the external web client, ensuring that updates to jellyfin-web take effect immediately.
Jun 7, 2024
- 0.0.3
May 21, 2024
- 0.0.20.0.2 Release
Minimal release notes: "- Separate logic between 3 main components: Crawler, Parser, Scraper - Add cache layer to CLI app - Add docs - Update"
Apr 17, 2024
- 0.0.1
Mar 16, 2024
- 0.2
A major update introducing new neural network modules, support for the Mistral model, and various performance improvements.
Mar 1, 2024
- 0.1
This is the first functional release of Luminal, a deep learning framework focused on static computation and aggressive optimization. It successfully runs the full Llama 1 7B model, though performance is currently slow.
Aug 11, 2023