Search results

为什么dual cache模式下，humaneval的准确率为0

Status: Open.

#66 In NVlabs/Fast-dLLM;

· zqc214 opened on Jan 17, 2026

Batch Inference using Fast-dLLM

Status: Open.

#65 In NVlabs/Fast-dLLM;

· Gaurav7888 opened on Jan 16, 2026

batch size > 1 and temperature != 0.0 not supported by the generate function file

Status: Open.

#64 In NVlabs/Fast-dLLM;

· Laurence-Wu opened on Jan 16, 2026

Discrepancy in GSM8K evaluation results compared to Table 1

Status: Open.

#63 In NVlabs/Fast-dLLM;

· JueVioletGrace opened on Jan 11, 2026

[v2] Reproducibility: release paper training script, preprocessing, and exact hyperparameters

Status: Open.

#62 In NVlabs/Fast-dLLM;

· yeonjoon-jung01 opened on Jan 5, 2026

Why does Dream’s cache “global update” always update from current_block_start instead of using a sampling strategy?

Status: Open.

#61 In NVlabs/Fast-dLLM;

· lizhuo008 opened on Dec 30, 2025

(v2) Open source code for transferring AR model to dLLM?

Status: Open.

#60 In NVlabs/Fast-dLLM;

· Xalp opened on Dec 30, 2025

请问为什么在eval_gsm8k.sh中，dual cache+paralle的steps参数是传入length，而不是steps

Status: Open.

#59 In NVlabs/Fast-dLLM;

· zqc214 opened on Dec 29, 2025

do we need chat template when evaluating llada-1.5

Status: Open.

#58 In NVlabs/Fast-dLLM;

· mynameisissac opened on Dec 28, 2025

Question Regarding the Llada Parallel decoding in GSM8K

Status: Open.

#54 In NVlabs/Fast-dLLM;

· littlestone111 opened on Dec 18, 2025

About warm-up stage and cache reuse stage in dual cache+parallel

Status: Open.

#52 In NVlabs/Fast-dLLM;

· quaternior opened on Dec 11, 2025

vllm support

Status: Open.

#50 In NVlabs/Fast-dLLM;

· hsliuustc0106 opened on Dec 8, 2025