surya Releases
166 releases of datalab-to/surya
- v0.6.4Misc bugfixes
Fixes various minor bugs including folder loading issues, updates dependencies, and suppresses warnings.
Oct 18, 2024
- v0.6.3
Updates dependencies and raises the minimum Python version requirement.
Oct 14, 2024
- v0.6.2Refactor cell assignment
Refactors cell assignment logic into an external library and improves cell extraction capabilities.
Oct 14, 2024
- v0.6.0Table recognition model release!
Introduces a new table recognition model and enhances accuracy and memory efficiency for layout and text detection.
Oct 8, 2024
- rustymimi-0.2.2
Minimal release notes: "This includes the extended context size to 8192 so should provide 5 mins conversations on moshi-mlx."
Sep 22, 2024
- rustymimi-0.2.1
Sep 20, 2024
- v0.5.0OCR v2
Introduces OCR v2 with a custom architecture offering faster processing, better accuracy on noisy documents, and handwriting support.
Aug 16, 2024
- v0.4.15Faster text detection + layout
Switches text detection and layout model architectures to significantly boost performance across different hardware accelerators.
Jul 12, 2024
- v0.4.14
Updates compatibility to handle a new kwarg introduced in a recent transformers version update.
Jun 30, 2024
- r1.4.2PaddleSpeech r1.4.2
PaddleSpeech r1.4.2 is a comprehensive update featuring new Speech-to-Text models like WavLM and HuBERT, additions to Singing Voice Synthesis (SVS) and Text-to-Speech (TTS) architectures, and various bug fixes.
Jun 27, 2024
- v0.4.12Minor bugfixes
May 28, 2024
- v0.4.11Fix image bugs
Fixes image processing bugs and optimizes segmentation code for better stability.
May 28, 2024
- v0.4.10Change image resize
Changes image resizing backend from OpenCV to PIL to fix benchmark regressions.
May 28, 2024
- v0.4.9OCR speedups
Optimizes OCR model speed and memory usage while adding static caching capabilities.
May 27, 2024
- v0.4.8Processor improvements
Removes unnecessary conversions and fixes a color channel bug in OCR to improve results.
May 23, 2024
- v0.4.7OCR speedup
Further accelerates OCR processing by cutting processing time in half.
May 18, 2024
- v0.4.6Significant speedup for layout, line detection
Improves CPU postprocessing speed for layout and line detection significantly.
May 17, 2024
- v0.4.2Add OCR batch size option
May 8, 2024
- v1.0.0Open-Sora V1.0.0 Release
Open-Sora V1.0.0 introduces a comprehensive open-source platform for high-quality video generation. It provides source code for data processing and a Sora-like training and inference pipeline, optimized with ColossalAI and flash attention, alongside pre-trained model weights.
Mar 17, 2024
- v0.2.4Fix dependency import
Mar 6, 2024