model-00010-of-00012.safetensors 1.3 GBmodel-00009-of-00012.safetensors 1.3 GBmodel-00003-of-00012.safetensors 1.3 GBmodel-00005-of-00012.safetensors 1.3 GBmodel-00007-of-00012.safetensors 1.3 GBmodel-00008-of-00012.safetensors 1.3 GBmodel-00004-of-00012.safetensors 1.3 GBmodel-00006-of-00012.safetensors 1.3 GBmodel-00011-of-00012.safetensors 1.3 GBmodel-00002-of-00012.safetensors 1.3 GBmodel-vision-00001-of-00001.safetensors 921 MBmodel-00001-of-00012.safetensors 675 MBmodel-00012-of-00012.safetensors 496 MBMLX/VLM conversion of empero-ai/Qwythos-27B-v1, pinned to revision 7c72a9c714cf66281cb222c4aa0aef368d84c94f.
This standard artifact intentionally disables the embedded MTP layer and does not include mtp.safetensors.
The source configuration's 1M-token YaRN settings are preserved, but this release was locally exercised only through 19,069 input tokens—not at 1M.
python -m mlx_vlm.generate \
--model <local-path-or-repo-id> \
--image image.jpg \
--prompt "Describe this image." \
--max-tokens 256
Review the source model card for intended use, training, license, and deployment considerations. This conversion does not change the source model's safety characteristics.
有文件?提交种子。
提交种子empero-ai/Qwythos-27B-v1