Search results

Request access to z-lab/MiniMax-M2.7-DFlash model

Status: Open.

#148 In z-lab/dflash;

· wangxj12 opened on Jul 13, 2026

Question about drafter for diffusion models themselves

Status: Open.

#147 In z-lab/dflash;

· hsamrat007 opened on Jul 9, 2026

[Bug] DFlash speculative decoding causes cudaErrorIllegalAddress under concurrent load when CUDA graphs are enabled

Status: Open.

#146 In z-lab/dflash;

· Love-learning-Li opened on Jul 9, 2026

Docker image is not publicly available - in contrast to README.md

Status: Open.

#144 In z-lab/dflash;

· kyr0 opened on Jun 28, 2026

KIND MODEL REQUEST: gemma4:12b-it-qat

Status: Open.

#143 In z-lab/dflash;

· ttx89-dev opened on Jun 25, 2026

NVFP4 + Dflash?

Status: Open.

#142 In z-lab/dflash;

· dude-maker opened on Jun 25, 2026

是否可以自己微调或者训练自己的dflash 呢

Status: Open.

#140 In z-lab/dflash;

· aofengdaxia opened on Jun 24, 2026

DFlash speculative decoding crashes with AssertionError in unify_kv_cache_spec_page_size on hybrid model Qwen3.5-35B-A3B (H20)

Status: Open.

#139 In z-lab/dflash;

· zhanyin-kun opened on Jun 23, 2026

Access to GLM-5.1-FP8-DFlash on huggingface

Status: Open.

#136 In z-lab/dflash;

· zhujiem opened on Jun 18, 2026

Docs suggestion: cold-start TPOT undersells steady-state by ~35% (43 vs 69 tok/s) until the draft path warms

Status: Open.

#135 In z-lab/dflash;

· cipherfoxie opened on Jun 12, 2026

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