Search results

"expected scalar type Float but found Float8_e8m0fnu" when running MXFP8 MoE

moe

Status: Open.

#3287 In pytorch/ao;

· goldhuang opened on Nov 4, 2025

TorchAO multi-gpu support tracker

tracker

Status: Open.

#3284 In pytorch/ao;

· jcaip opened on Nov 3, 2025

[moe training] Update torchao docsite with MoE training docs

moe

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

Status: Open.

#3281 In pytorch/ao;

· danielvegamyhre opened on Nov 3, 2025

macos failed to import torchao.quantization.quantize_

autoquant

Status: Open.

#3280 In pytorch/ao;

· ownia opened on Nov 3, 2025

Using <code>WandaSparsifier</code> on Whisper causes odd <code>input_features</code> for <code>PerChannelNormObserver</code>

sparsity

Status: Open.

#3278 In pytorch/ao;

· rasgaard opened on Nov 3, 2025

Proposal to add a beginner-friendly introduction tutorial for TorchAO

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

Status: Open.

#3274 In pytorch/ao;

· smishra8 opened on Nov 1, 2025

[DOCS] Quick Start Guide PT2E Example does not work as is. Undefined objects

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

Status: Open.

#3270 In pytorch/ao;

· cjm715 opened on Oct 31, 2025

[RFC] NVFP4 Rounding Modes

Status: Open.

#3264 In pytorch/ao;

· syed-ahmed opened on Oct 30, 2025

Add support for parameter quantization to all configs

tracker

Status: Open.

#3252 In pytorch/ao;

· jcaip opened on Oct 28, 2025

TorchAO Missing 3.13T (free-threading) Wheels

Status: Open.

#3243 In pytorch/ao;

· Qubitium opened on Oct 24, 2025

Quantized OpenVLA 7B (4-bit) runs slower than default model

Status: Open.

#3236 In pytorch/ao;

· arunmadhusud opened on Oct 24, 2025

nvfp4: doesn't work with vLLM + lm-eval

Status: Open.

#3234 In pytorch/ao;

· vkuzo opened on Oct 23, 2025