Don't embed external initializers into the proto to avoid 2GB limit #817
Conversation
javier-intel
requested review from
MayureshV1,
ericcraw,
javier-intel and
Copilot
There was a problem hiding this comment.
Pull Request Overview
This PR modifies the model proto generation to avoid embedding external initializers directly into the proto, preventing potential issues when models exceed the 2GB protobuf limit. The changes introduce a mechanism to handle initializers with external data by updating their metadata to use a special memory address marker instead of including the actual data.
Key changes:
- Add logic to exclude initializer data from proto generation when
load_user_initializer_is enabled - Implement metadata updating for external initializers using the
*/_ORT_MEM_ADDR_/*marker - Add comprehensive logging for debugging initializer processing
Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.
… restores the metadata so OV can read them back Signed-off-by: bfilipek <[email protected]>
intbf
force-pushed
the
dont_write_ext_initializers
branch
from
fd29454 to
2b34773
Compare
…re are external initializers in memory (more than one) Signed-off-by: bfilipek <[email protected]>
There was a problem hiding this comment.
Pull Request Overview
Copilot reviewed 1 out of 1 changed files in this pull request and generated 3 comments.
Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.
| } | ||
|
|
||
| static void ReadExternalDataFields(const ONNX_NAMESPACE::TensorProto* src_init, std::string& location, size_t& offset, size_t& length) { | ||
| // Remove constness as we need to use mutable_external_data() to get the entries to read. |
There was a problem hiding this comment.
There's a trailing space after the period in the comment on line 497.
| // Remove constness as we need to use mutable_external_data() to get the entries to read. | |
| // Remove constness as we need to use mutable_external_data() to get the entries to read. |
Copilot uses AI. Check for mistakes.
| } | ||
| else { | ||
| // Debug info for file-based initializers | ||
| LOGS(logger, VERBOSE)<< "File-based initializer: " |
There was a problem hiding this comment.
Missing space between LOGS(logger, VERBOSE) and << operator.
| LOGS(logger, VERBOSE)<< "File-based initializer: " | |
| LOGS(logger, VERBOSE) << "File-based initializer: " |
Copilot uses AI. Check for mistakes.
Signed-off-by: bfilipek <[email protected]>
Description
When calling subgraph.ToProto don't save initializers, so that we don't bloat the model proto. This is especially important for models with ext initializers in memory that in total contain more than 2GB of data.
Once we have the list of those external initializers, we can update ther metadata inside the proto, so that they leverage the special marker:
*/_ORT_MEM_ADDR_/*. (The support for this location was recently fixed inside OV with openvinotoolkit/openvino#31632)Motivation and Context
CVS-173057, CVS-172710
Todo/questions