Search results

Is there no tutorial for dynamic quantization of BERT model in torch.ao?

Status: Open.

#3157 In pytorch/ao;

· Esttelle opened on Oct 12, 2025

NotImplementedError: Int4PreshuffledTensor dispatch: attempting to run unimplemented operator/function:

Status: Open.

#3144 In pytorch/ao;

· seungduk-yanolja opened on Oct 10, 2025

README should highlight our huggingface models

topic: documentationUse this tag if this PR adds or improves documentationUse this tag if this PR adds or improves documentation

Status: Open.

#3137 In pytorch/ao;

· andrewor14 opened on Oct 9, 2025

Implement NVFP4Tensor.fill_

Status: Open.

#3134 In pytorch/ao;

· andrewor14 opened on Oct 9, 2025

Fix test_quantize_pt2e_qat.py::test_mixing_qat_ptq in CI

Status: Open.

#3125 In pytorch/ao;

· jainapurva opened on Oct 6, 2025

Access to compact internal representation for <code>target_dtype=torch.uint4</code>

question

triaged

Status: Open.

#3122 In pytorch/ao;

· mseeger opened on Oct 6, 2025

TorchAOBaseTensor should provide an interface to convert to high precision tensor

Status: Open.

#3118 In pytorch/ao;

· vkuzo opened on Oct 2, 2025

[mxfp8 moe traing] to_mxfp8_a2a_dequant flat perf vs high precision a2a

moe

mx

Status: Open.

#3112 In pytorch/ao;

· danielvegamyhre opened on Oct 2, 2025

FLUX FP8 Training Error

bugSomething isn't workingSomething isn't working

float8

Status: Open.

Bug

#3111 In pytorch/ao;

· faruknane opened on Oct 1, 2025

Extend NVFP4 QAT to support emulation

Status: Open.

#3102 In pytorch/ao;

· andrewor14 opened on Sep 30, 2025

Model with subclass fails to export

Status: Open.

#3099 In pytorch/ao;

· metascroy opened on Sep 30, 2025

MoE training conversion function should log full FQN of swapped parameters

moe

Status: Open.

#3097 In pytorch/ao;

· danielvegamyhre opened on Sep 29, 2025