Qwen3.8-Flash-Next EXL3 4.05 bpw for mlx-serve (moved)
Moved to OpensourceWTF/Qwen3.8-Flash-Next-exl3, branch 4.05bpw.
Every file this repo held at commit ed79daeb, except ngram_embedding.safetensors, is on that branch with the same
bytes, each checked against this repo's sha256s. The n-gram table is turboderp's file, unchanged: get ngram_embedding.safetensors from his branch
4.05bpw_h6_ng6 (sha256 c63613666649…, 39,040,193,720 B) and put it in the pack folder as it is. This repo's history was squashed to free storage.
The quantization is turboderp's: turboderp/Qwen3.8-Flash-Next-exl3 (branch 4.05bpw_h6_ng6), made with
exllamav3. Every branch, its data and its charts:
OpensourceWTF/Qwen3.8-Flash-Next-exl3.