Back to search
Top Contributors
128
commits
4
commits
2
commits
1
commits
1
commits
1
commits
1
commits
Rust multi‑backend OCR/VLM engine (DeepSeek‑OCR-1/2, PaddleOCR‑VL, DotsOCR) with DSQ quantization and an OpenAI‑compatible server & CLI – run locally without Python.
Stars
2.2k
Forks
168
Watchers
2.2k
Open Issues
14
Overall repository health assessment
No package.json found
This might not be a Node.js project
128
commits
4
commits
2
commits
1
commits
1
commits
1
commits
1
commits
refactor(config): unify inference decode parameters handling
02b933dView on GitHubrefactor: move inline tests to crate integration suites and fix clippy warnings
5daf103View on GitHubfix(benchsuite): enforce Python/Rust prompt parity and stable strict compare
422c03aView on GitHubfix(deepseek): restrict python baseline to cpu and adjust model dtype handling
5fb5584View on GitHubfeat(benchsuite): add Python baseline support with isolated runtime environments
2f64c1eView on GitHubrefactor(benchsuite): add support for deepseek , dots.ocr , paddleocr support
f3d746bView on GitHubdocs(benchsuite): update README with detailed usage and commands
64fd4f4View on GitHubperf: refactor GLM KV cache and attention, add end-to-end timing instrumentation, and migrate to a unified benchsuite for benchmarking and strict gating
aae7a64View on GitHubfix(dsq-writer): adapt q4k/q6k quantization calls to candle 0.9.2 from_float signature
974a763View on GitHubci: remove CUDA 13 Candle git-tip workaround in extended artifacts workflow
5da4c0cView on GitHub