teru00801/hawks-qwen3_5-35b-a3b-merged-0612-fsdp
This repository contains the merged/fused HF-format model.
It is intended as the precision-preserving source for downstream conversions,
including MLX quantization and future runtime formats.
Source
- Base model:
unsloth/Qwen3.5-35B-A3B
- GGUF runtime package:
teru00801/hawks-qwen3_5-35b-a3b-gguf-0612-fsdp
MLX conversion
Run this on Apple Silicon:
mlx_lm.convert --model teru00801/hawks-qwen3_5-35b-a3b-merged-0612-fsdp -q --upload-repo <your-mlx-repo>
