Search results

[JAX] Extend permutation primitive to multi-GPU

MoE

Status: Open.

#2536 In NVIDIA/TransformerEngine;

· tdophung opened on Dec 19, 2025

Linear layer reset_parameters() changes bias zero init to random init near 0

bugSomething isn't workingSomething isn't working

Status: Open.

#2529 In NVIDIA/TransformerEngine;

· jstjohn opened on Dec 17, 2025

LayerNormLinear reset_parameters() leads to the wrong initialization.

bugSomething isn't workingSomething isn't working

Status: Open.

#2528 In NVIDIA/TransformerEngine;

· jstjohn opened on Dec 17, 2025

FA3 fails in the newest version of FA3

bugSomething isn't workingSomething isn't working

Status: Open.

#2527 In NVIDIA/TransformerEngine;

· MaciejBalaNV opened on Dec 17, 2025

What is the updated NVFP4 recipe for Nemotron 3 SUPER/ULTRA?

Status: Open.

#2526 In NVIDIA/TransformerEngine;

· TheTinyTeddy opened on Dec 17, 2025

Blockwise (1x128 and 128x128) FP8 grouped quantization

MoE

Status: Open.

#2525 In NVIDIA/TransformerEngine;

· ptrendx opened on Dec 16, 2025

cannot import name 'load_framework_extension' from 'transformer_engine.common'

bugSomething isn't workingSomething isn't working

Status: Open.

Bug

#2515 In NVIDIA/TransformerEngine;

· LiYufengzz opened on Dec 13, 2025

[Docs] Add datatype × hardware support matrix

Status: Open.

#2512 In NVIDIA/TransformerEngine;

· sbhavani opened on Dec 11, 2025

[Feature Request] Add Native Sparse Attention (NSA) support

Status: Open.

#2511 In NVIDIA/TransformerEngine;

· sbhavani opened on Dec 11, 2025

Refactor the grouped functionality to use NVTEGroupedTensor

2.12.0

MoE

Status: Open.

#2510 In NVIDIA/TransformerEngine;

· ptrendx opened on Dec 11, 2025

Examples: Using TE with LLaMA-Factory

Status: Open.

#2509 In NVIDIA/TransformerEngine;

· sbhavani opened on Dec 11, 2025

Problem encountered while compiling version v2.6

bugSomething isn't workingSomething isn't working

Status: Open.

#2504 In NVIDIA/TransformerEngine;

· 191220042 opened on Dec 11, 2025