Search results

Model does not learn with u-mup

Status: Open.

#88 In graphcore-research/unit-scaling;

· SwayStar123 opened on May 26, 2025

Add uu.Conv2d

Status: Open.

#87 In graphcore-research/unit-scaling;

· SwayStar123 opened on May 24, 2025

Scaling self-Attention when inference

Status: Open.

#86 In graphcore-research/unit-scaling;

· mmorinag127 opened on Apr 9, 2025

Support DTensor for distributed training

Status: Open.

#85 In graphcore-research/unit-scaling;

· rlrs opened on Dec 17, 2024

Attention Issue for DiT-like model

Status: Open.

#84 In graphcore-research/unit-scaling;

· kliyer-ai opened on Dec 10, 2024

Fix wandb logging bug

bugSomething isn't workingSomething isn't working

Status: Open.

#82 In graphcore-research/unit-scaling;

· thecharlieblake opened on Nov 11, 2024

Adam and SGD update learning rate scaling.

Status: Open.

#73 In graphcore-research/unit-scaling;

· norpadon opened on Oct 14, 2024

MLP scaling w.r.t. expansion_factor

Status: Open.

#70 In graphcore-research/unit-scaling;

· EIFY opened on Sep 6, 2024

[In Progress] Sharing Unit-Mamba Implementation

Status: Open.

#68 In graphcore-research/unit-scaling;

· norikazu99 opened on Aug 25, 2024

[Question] What is 'mult' parameter?

Status: Open.

#62 In graphcore-research/unit-scaling;

· mmorinag127 opened on Aug 1, 2024

[Feature Request] Any plan to support fp8 type at latest torch version?

Status: Open.

#45 In graphcore-research/unit-scaling;

· MoFHeka opened on Sep 5, 2023