Skip to content

Session config to create weightless EPContext models during compilation#27197

Merged
adrianlizarraga merged 3 commits intomainfrom
adrianl/WeightlessEpContext_AppEnable
Mar 9, 2026
Merged

Session config to create weightless EPContext models during compilation#27197
adrianlizarraga merged 3 commits intomainfrom
adrianl/WeightlessEpContext_AppEnable

Conversation

@adrianlizarraga
Copy link
Copy Markdown
Contributor

Description

Adds a session configuration entry named ""ep.enable_weightless_ep_context_nodes" that an application enables to explicitly request that EPs create compiled models (i.e., EPContext models) with EPContext nodes that do not store model weights internally. Instead, the weights should be provided by ONNX Runtime as explicit inputs to the EPContext nodes.

Motivation and Context

If the weights are originally stored in an external file, this allows multiple models to share the same external weights file.

@javier-intel
Copy link
Copy Markdown
Contributor

Looks good to me

@javier-intel
Copy link
Copy Markdown
Contributor

Hi @adrianlizarraga, is this still planned for the next ORT release?

@adrianlizarraga adrianlizarraga marked this pull request as ready for review March 6, 2026 22:14
@adrianlizarraga adrianlizarraga merged commit 6f55965 into main Mar 9, 2026
102 of 106 checks passed
@adrianlizarraga adrianlizarraga deleted the adrianl/WeightlessEpContext_AppEnable branch March 9, 2026 17:01
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants