List view

Clarify DFT behavior when <code>inverse=True</code>

topic: spec clarificationClarification of the ONNX spec neededClarification of the ONNX spec needed

Status: Open.

#5920 In onnx/onnx;

· justinchuby opened on Feb 8, 2024

[Spec issue] Rounding or truncation behavior for Cast operator is not specified in the spec

bug

contributions welcome

topic: operatorIssues related to ONNX operatorsIssues related to ONNX operators

topic: spec clarificationClarification of the ONNX spec neededClarification of the ONNX spec needed

Status: Open.

#3876 In onnx/onnx;

· max-ku opened on Dec 2, 2021

Ambiguity in Pool spec (AveragePool and MaxPool)

contributions welcome

topic: operatorIssues related to ONNX operatorsIssues related to ONNX operators

topic: spec clarificationClarification of the ONNX spec neededClarification of the ONNX spec needed

Status: Open.

#1922 In onnx/onnx;

· hariharans29 opened on Apr 9, 2019

Spec issue: GridSample

contributions welcome

module: spec

topic: operatorIssues related to ONNX operatorsIssues related to ONNX operators

Status: Open.

#4033 In onnx/onnx;

· broune opened on Feb 21, 2022

[Spec] DepthToSpace <code>mode</code> attribute is counter-intuitive

contributions welcome

module: spec

questionQuestions about ONNXQuestions about ONNX

topic: documentationIssues related to ONNX documentationIssues related to ONNX documentation

topic: operatorIssues related to ONNX operatorsIssues related to ONNX operators

topic: spec clarificationClarification of the ONNX spec neededClarification of the ONNX spec needed

Status: Open.

#6069 In onnx/onnx;

· zhenhuaw-me opened on Apr 8, 2024

Parser treats a float attribute as int and produces invalid model

bug

module: parseronnx parseronnx parser

Status: Open.

#5748 In onnx/onnx;

· justinchuby opened on Nov 10, 2023

Make InstanceNormalization weights and bias optional

contributions welcome

topic: operatorIssues related to ONNX operatorsIssues related to ONNX operators

Status: Open.

#6552 In onnx/onnx;

· justinchuby opened on Nov 23, 2024

Expanding to 0-size shape <code>[0, ...]</code> behaves differently in PyTorch and ONNX

contributions welcome

topic: spec clarificationClarification of the ONNX spec neededClarification of the ONNX spec needed

Status: Open.

Feature

#6664 In onnx/onnx;

· himalayjor opened on Jan 28, 2025

[RFC] FP6 quantization data type

contributions welcome

module: spec

rfcRequest for CommentsRequest for Comments

Status: Open.

Feature

#7048 In onnx/onnx;

· TedThemistokleous opened on Jun 13, 2025

Add support for FP6 quantization data types (FLOAT6E2M3 and FLOAT6E3M2) addressing RFC #7048

module: spec

Status: Open (in progress).

onnx/onnxnumber 7145

#7145 In onnx/onnx;

· MagellaX opened on Jul 18, 2025

ONNX printer does not print initializers when they are types like float16

bug

Status: Open.

#7053 In onnx/onnx;

· justinchuby opened on Jun 15, 2025

Introduce BitCast (reinterpret cast)

topic: operatorIssues related to ONNX operatorsIssues related to ONNX operators

Status: Open.

#6887 In onnx/onnx;

· justinchuby opened on Apr 11, 2025

[Feature request] Allow Type Inference Without Shape Inference in ONNX

contributions welcome

good first issueGood for getting startedGood for getting started

module: shape inferenceIssues related to shape inferenceIssues related to shape inference

topic: enhancementRequest for new feature or operatorRequest for new feature or operator

Status: Open.

#7100 In onnx/onnx;

· galagam opened on Jul 2, 2025

Missing type support in parser for various types (float16, bfloat16, ...)

bug

contributions welcome

module: parseronnx parseronnx parser

Status: Open.

#6152 In onnx/onnx;

· TinaAMD opened on May 31, 2024

[RFC] Equivalent subgraphs and compiled subgraphs

module: spec

rfcRequest for CommentsRequest for Comments

Status: Open.

Feature

#7301 In onnx/onnx;

· justinchuby opened on Sep 13, 2025

ONNX 1.18.0 wheel building fails with Python 3.12: cmake uses system Python instead of runtime Python

topic: buildIssues related to ONNX builds and packagesIssues related to ONNX builds and packages

Status: Open.

#7226 In onnx/onnx;

· coderabbitai opened on Aug 12, 2025

Clarify specification of mixed precision Pow op.

contributions welcome

topic: operatorIssues related to ONNX operatorsIssues related to ONNX operators

topic: spec clarificationClarification of the ONNX spec neededClarification of the ONNX spec needed

Status: Open.

#7071 In onnx/onnx;

· gramalingam opened on Jun 23, 2025

Clarify mixed precision semantics in Pow operator specification

Status: Open (in progress).

onnx/onnxnumber 7139

#7139 In onnx/onnx;

· Copilot opened on Jul 17, 2025

Reference implementation of ReduceLogSumExp crashes with integer input

bug

contributions welcome

good first issueGood for getting startedGood for getting started

topic: operatorIssues related to ONNX operatorsIssues related to ONNX operators

Status: Open.

#7141 In onnx/onnx;

· cbourjau opened on Jul 17, 2025

Hardmax, LogSoftMax, Softmax behavior with scalar inputs

contributions welcome

questionQuestions about ONNXQuestions about ONNX

topic: operatorIssues related to ONNX operatorsIssues related to ONNX operators

Status: Open.

#7155 In onnx/onnx;

· konstantin-pueckler-qc opened on Jul 22, 2025

Add BitCast operator

review needed: operators approversRequire reviews from members of operators-approversRequire reviews from members of operators-approvers

Status: Open (in progress).

onnx/onnxnumber 7084

#7084 In onnx/onnx;

· alx256 opened on Jun 26, 2025

Specify that present_key and past_key relationship for Attention

topic: spec clarificationClarification of the ONNX spec neededClarification of the ONNX spec needed

Status: Open.

#7389 In onnx/onnx;

· justinchuby opened on Oct 14, 2025

[RFC] What is missing from ONNX's representation of quantized models?

rfcRequest for CommentsRequest for Comments

topic: operatorIssues related to ONNX operatorsIssues related to ONNX operators

Status: Open.

#7435 In onnx/onnx;

· justinchuby opened on Oct 29, 2025

ONNX 1.19 tests fails on osx-arm64 platfrom (eg ShapeInferenceError, TypeInferenceError)

bug

Status: Open.

#7376 In onnx/onnx;

· Andrii-Osipov opened on Oct 8, 2025

Allow float input to Mod when fmod=0

contributions welcome

topic: operatorIssues related to ONNX operatorsIssues related to ONNX operators

Status: Open.

#7520 In onnx/onnx;

· justinchuby opened on Dec 12, 2025