microsoft / microsoft/onnxruntime

Type 'tensor(bfloat16)' of input parameter of operator (InstanceNormalization) is invalid.

Open
#20,051 1 comment 0 reactions 0 assignees View on GitHub
feature request
Dominant language
C++
Stars
21.9k
Forks
4.2k
Avg merge
4d 11h
Merged PRs (30d)
184

Description

### Describe the feature request

bf16 seems not supported for InstanceNormalizationin 1.17. When can we have the InstanceNormalizationsupport bf16?``

### Describe scenario use case

InstanceNormalization is used in many CNN models and also in Stable Diffusion. We need this to support bf16 as welll

Contributor guide

Open the contributing guide

Research direction

No file or test is named in the issue. Start by locating the InstanceNormalization operator implementation and the existing coverage for its input types; confirm the current bfloat16 failure, with done defined as successful bfloat16 support for the operator in the CNN and Stable Diffusion scenarios described.

Written by the indexing model from the issue text.

Assessment

Tech stack
cpp
Domain
machine-learning
Issue type
Feature
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
38/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.