Search results

qwen3.5系列的草稿模型，没标nonthink和think标志的，是同时支持两种模式么？

Status: Open.

#134 In z-lab/dflash;

· yangyan00800-blip opened on Jun 12, 2026

Acceptance rate lower than benchmark with Qwen3-8B + Dflash

Status: Open.

#133 In z-lab/dflash;

· wsb853529465 opened on Jun 11, 2026

request for z-lab/GLM-5.1-FP8-DFlash

Status: Open.

#132 In z-lab/dflash;

· yumingfan-0219 opened on Jun 11, 2026

Model Request: DFlash for Llama 3.3 70B

Status: Open.

#131 In z-lab/dflash;

· wzyyy-lab opened on Jun 10, 2026

MODEL REQUEST: Dflash for MiMo V2.5 (non-Pro)

Status: Open.

#130 In z-lab/dflash;

· alexanderjacuna opened on Jun 8, 2026

The dflash model seems does not has good performance in dgx spark box

Status: Open.

#129 In z-lab/dflash;

· carvinrui opened on Jun 8, 2026

larger gemma models with fp8 are generating junk

Status: Open.

#128 In z-lab/dflash;

· programmeddeath1 opened on May 29, 2026

请问能否释放Qwen3.5系列模型的3层DFlash草稿模型？

Status: Open.

#127 In z-lab/dflash;

· pbhfcycssjlmm opened on May 29, 2026

question about sliding window attention

Status: Open.

#126 In z-lab/dflash;

· ZelinMa557 opened on May 27, 2026

CUDA 12.8 install of dflash[sglang] fails

Status: Open.

#123 In z-lab/dflash;

· Sylvan820 opened on May 19, 2026

Ablation on Attention Mask Type

Status: Open.

#122 In z-lab/dflash;

· zhi-dian opened on May 15, 2026

DFLASH for GLM51

Status: Open.

#121 In z-lab/dflash;

· dongyibo opened on May 15, 2026