model2vec Releases
74 releases of MinishLab/model2vec
- v0.3.1
This maintenance release improves token handling, adds optional fast encoding, and fixes a normalization bug.
Oct 29, 2024
- v0.3.0
A major feature release introducing numpy inference, local model loading, and faster tokenization capabilities.
Oct 21, 2024
- v0.2.4
A dependency update release that expands compatibility with the Hugging Face Hub library.
Oct 10, 2024
- v0.2.3
This release improves device management, adds utility functions, and enhances configuration handling.
Oct 9, 2024
- v0.2.2
Bug fixes targeting tokenizer errors, attention masks, and device selection mechanisms.
Oct 3, 2024
- v0.2.1
A maintenance release that introduces a limit on token length based on the median value.
Sep 30, 2024
- v0.2.0
A feature-rich release introducing safetensors support, hybrid tokenizers, and model cards.
Sep 28, 2024
- v0.1.2
A brief release dedicated to fixing a bug related to model dimensions.
Sep 23, 2024
- v0.1.1
Bug fixes and documentation updates including a fix for PCA and the addition of multilingual model documentation.
Sep 21, 2024
- v0.1.0
This is the initial public release of the model2vec project, establishing the foundational codebase, benchmarks, and documentation.
Sep 20, 2024
- 3.3.0
This release focuses on dependency updates for core libraries like transformers and torch, along with bug fixes for Streamlabs Polly, random voice prompts, and the new UI. It also introduces a fancy title screen.
Jun 16, 2024
- 3.2.1
Minimal release notes: "Version 3.2.1 is a bug release to address the following issues: - Translation - Story mode method 0 - Conversion fail"
Aug 24, 2023
- 3.2
The RedditVideoMakerBot update adds ElevenLabs TTS, background audio, and random voices, along with various bug fixes and UI improvements.
May 28, 2023
- 3.1
This version focuses on performance, making the program up to 40 times faster, fixing Reddit login issues across languages, and adding a transparent image mode.
Apr 13, 2023