hugpadhugpad

ProCreations / Nanbeige4.3-33B-DFlash2

Text Generation980M1.4K10Updated Oct 26, 2026
Hugging Face

Downloads

1.4K

Likes

10

Last updated

Oct 26, 2026

License

apache-2.0

DFlash2 block-diffusion draft model (3 layers, block 8) for Nanbeige4.3-33B. Batch-1 decode speedups on RTX PRO 6000 (Modal 600 W), SGLang 0.5.21: BF16 1.76× greedy / 1.63× T1; FP8 1.65× / 1.49×; NVFP4-A16 1.51× / 1.39×. Native exllamav3 with the EXL3 quants (Vega RTX PRO 6000…

Tasks & library

  • text generation
  • sglang

Tags

  • sglang
  • qwen3
  • dflash2
  • speculative-decoding
  • nanbeige
  • text-generation

Technical specifications

Parameters980M
Librarysglang
LanguagesNot available
Base modelNanbeige/Nanbeige4.3-33B
Licenseapache-2.0