DeepSeek-V4.1-Flash EXL3 packs whose routed experts stream from SSD, for mlx-serve.
David Tai
OpensourceWTF
AI & ML interests
None yet
Recent Activity
updated a model 2 days ago
OpensourceWTF/Qwen3.8-Flash-Next-exl3 updated a model 3 days ago
OpensourceWTF/DeepSeek-V4.1-Flash-MTPLX-streaming-q2 updated a model 3 days ago
OpensourceWTF/DeepSeek-V4.1-Flash-streaming-repack-exl3Organizations
Packs for MTPLX on Apple silicon, most of them streaming their experts from SSD.
-
OpensourceWTF/DeepSeek-V4.1-Flash-MTPLX-streaming-q2
Updated • 1.16k • 4 -
OpensourceWTF/Hy3-oQ2e-MTPLX-streaming
Text Generation • 9B • Updated • 70 • 3 -
OpensourceWTF/Kimi-K3-Q2_K-t158-MTPLX-streaming
Text Generation • 57B • Updated • 373 -
OpensourceWTF/Qwen3.8-Flash-Next-MTPLX-oQ4-MTP
180B • Updated • 287 • 1
turboderp's Qwen3.8-Flash-Next EXL3 quants in mlx-serve's layout.
Other streaming and MLX packs.
-
OpensourceWTF/Beam-1.0-501B-A23B-NVFP4-q8-mlx-stream
Text Generation • 3B • Updated • 398 • 1 -
OpensourceWTF/MiMo-V2.6-Flash-MOPD-fp8res-mlx-stream
Text Generation • 6B • Updated • 393 -
OpensourceWTF/Muse-Glimmer-30B-MLX-q4
Text Generation • 28B • Updated • 218 • 1 -
OpensourceWTF/DeepSeek-V4.1-Flash-1017-exl3-2.81bpw
Updated • 477 • 4
GLM-5.3 and GLM-5.3-Flash EXL3 packs whose experts stream from SSD, for mlx-serve + mlx-stream.
DeepSeek-V4.1-Flash EXL3 packs whose routed experts stream from SSD, for mlx-serve.
Other streaming and MLX packs.
-
OpensourceWTF/Beam-1.0-501B-A23B-NVFP4-q8-mlx-stream
Text Generation • 3B • Updated • 398 • 1 -
OpensourceWTF/MiMo-V2.6-Flash-MOPD-fp8res-mlx-stream
Text Generation • 6B • Updated • 393 -
OpensourceWTF/Muse-Glimmer-30B-MLX-q4
Text Generation • 28B • Updated • 218 • 1 -
OpensourceWTF/DeepSeek-V4.1-Flash-1017-exl3-2.81bpw
Updated • 477 • 4
Packs for MTPLX on Apple silicon, most of them streaming their experts from SSD.
-
OpensourceWTF/DeepSeek-V4.1-Flash-MTPLX-streaming-q2
Updated • 1.16k • 4 -
OpensourceWTF/Hy3-oQ2e-MTPLX-streaming
Text Generation • 9B • Updated • 70 • 3 -
OpensourceWTF/Kimi-K3-Q2_K-t158-MTPLX-streaming
Text Generation • 57B • Updated • 373 -
OpensourceWTF/Qwen3.8-Flash-Next-MTPLX-oQ4-MTP
180B • Updated • 287 • 1
GLM-5.3 and GLM-5.3-Flash EXL3 packs whose experts stream from SSD, for mlx-serve + mlx-stream.
turboderp's Qwen3.8-Flash-Next EXL3 quants in mlx-serve's layout.