Search results

在训练的stage2，请问代码是如何将attention冻住的

Status: Open.

#1727 In microsoft/unilm;

· liu-zongxi opened on Jul 21, 2025

The download link for trcor is invalid.

Status: Open.

#1726 In microsoft/unilm;

· ruc-G opened on Jul 10, 2025

ReSA Eval crashed

Status: Open.

#1725 In microsoft/unilm;

· MengAiDev opened on Jul 8, 2025

Questions on BEiT v2 VQ-KD Pretraining

Status: Open.

#1724 In microsoft/unilm;

· IKnowWhoo opened on Jul 4, 2025

When will you open source the data synthesis code of paper "Scaling Laws of Synthetic Data for Language Model"

Status: Open.

#1722 In microsoft/unilm;

· butterluo opened on Jul 3, 2025

LHRMs (Large Hybrid-Reasoning Models) code & checkpoints referenced in paper 2505.14631 not found

Status: Open.

#1719 In microsoft/unilm;

· almogtavor opened on May 31, 2025

Differential Transformer loss spikes while training.

Status: Open.

#1718 In microsoft/unilm;

· fasil-saidalavi opened on May 20, 2025

Ablation tests with the same headdim as v with differential transformers

Status: Open.

#1716 In microsoft/unilm;

· RuiWang1998 opened on May 13, 2025

use wavlm to separate mix-audio

Status: Open.

#1715 In microsoft/unilm;

· yangwyy opened on May 10, 2025

Can BEiT2 weights be used for the image component of a BEiT3 pretrained model instead of performing BEiT3 pretraining?

Status: Open.

#1713 In microsoft/unilm;

· IKnowWhoo opened on May 4, 2025

Latent model release

Status: Open.

#1712 In microsoft/unilm;

· pli76 opened on Apr 23, 2025

configuration_bitnet.py

Status: Open.

#1711 In microsoft/unilm;

· werruww opened on Apr 19, 2025