- Lafayette, IN, United States
-
01:24
(UTC -04:00)
Pinned Loading
-
ShokoAnime/ShokoServer
ShokoAnime/ShokoServer PublicShoko is a media metadata aggregator that solves all of the annoying edge cases that face anime's differing structure from western media. It integrates directly with media software rather than work…
-
Qwen3.8-Flash-Next-LlamaCpp-MoE-Cache-Arc-Experiments
Qwen3.8-Flash-Next-LlamaCpp-MoE-Cache-Arc-Experiments PublicPorting a CUDA MoE expert-cache design to llama.cpp's SYCL backend for Qwen3.8-Flash-Next on an Arc Pro B70; pivoted into long-context serving (sparse QSA attention, YaRN/RoPE, dynamic context-leng…
Python 1
-
Qwen3.8-vLLM-KVarN-MTP-Arc-Experiments
Qwen3.8-vLLM-KVarN-MTP-Arc-Experiments PublicQwen3.8-27B on vLLM/Intel Arc: KVarN 4-bit/2-bit KV-cache quantization for up to 550K context, with MTP speculative decoding to recover the speed cost.
Python 1
-
unbooru-tagger
unbooru-tagger PublicA multi-label anime/booru-style image tagger built on a dual-encoder, sigmoid-contrastive architecture (JoyTag / SigLIP-style), implemented in C# with TorchSharp for training and ONNX Runtime for i…
C#
-
openvinotoolkit/openvino
openvinotoolkit/openvino PublicOpenVINO™ is an open source toolkit for optimizing and deploying AI inference
If the problem persists, check the GitHub status page or contact support.




