subframe7536 / maple-font
Maple Mono: Open source monospace font with round corner, ligatures and Nerd-Font icons for IDE and terminal, fine-grained customization options. 带连字和控制台图标的圆角等宽字体，中英文宽度完美2:1，细粒度的自定义选项
subframe7536 / maple-font
om-ai-lab / VLX-Seek
VLX-Seek is an efficient VLM that grounds fine-grained visual targets with high precision in wild environments.
1038lab / ComfyUI-RMBG
A ComfyUI custom node designed for advanced image background removal and object, face, clothes, and fashion segmentation, utilizing multiple models including RMBG-2.0, VRMBG-3.0, FeyNoBg, Finegrain Box Segmenter, INSPYRENET, BEN, BEN2, BiRefNet, Lucida, SDMatte, SAM, SAM2, SAM3, Florence-2, YOLOv8/YOLO11, Big-Lama, and GroundingDINO.
Peterande / D-FINE
D-FINE: Redefine Regression Task of DETRs as Fine-grained Distribution Refinement [ICLR 2025 Spotlight]
llm-as-a-verifier / llm-as-a-verifier
LLM-as-a-Verifier is a general-purpose framework that provides fine-grained feedback for any agent without requiring additional training. It achieves SOTA performance across coding, robotics, and medical agentic benchmarks.
sbi-dev / sbi
sbi is a Python package for simulation-based inference, designed to meet the needs of both researchers and practitioners. Whether you need fine-grained control or an easy-to-use interface, sbi has you covered.
google / grain
Library for reading and processing ML training data.
Tencent / ContextPilot
[EMNLP26] contextpilot: teaching agents for proactive context management via fine-grained rl
davefano / omarchy-trackpad-plus
Fine-grained, per-device trackpad controls and pointer-feel tuning for Omarchy.
Dimweaker / jev-libero
Fine-grained robot control with Jev, physics previews, and configurable LIBERO tasks.
lab-emi / GrainSpeech
Official repository for “GrainSpeech: Less Context, More Detail for Compact Speech Synthesis.”
LiorMoment / darkslide-gr
Load film looks and your own LUTs into your GR III / GR IIIx — a local browser editor for film, presets, grain and shutdown screen. No firmware included.
yyyyu120 / W2-VLA
World-to-Wrist: Task-Conditioned Future Wrist Modeling for Fine-Grained Robot Manipulation
Swapnil-jain / integer-mult-kappa
Integer multiplication below n log n: a conditional witness kappa = 6685733/10^10 (> 2^-11) for T(n) = O(n (log n)^(1-kappa)) in the multitape model, building on OpenAI's problem and Colkitt's integer-mult-bounds: certificates, independent replays, Lean 4 checks