fix: separate model weights from inputs in extraction output#720
Open
xjtu-L wants to merge 1 commit into
Open
Conversation
Previously, `extractor.py` passed all placeholder params (including lifted parameters, buffers, SymInt symbols, and real inputs) to `convert_state_and_inputs(params, [])`, causing all of them to be written to `weight_meta.py`. This led to SymInt placeholders like `Program_weight_tensor_meta_s0` appearing in `weight_meta.py`, which are not real model weights. This change: - Collects `id()` of original model parameters and buffers in `wrapper()` - Passes them to `GraphExtractor` via `param_buffer_ids` - Splits `params` into `weights` (params/buffers by identity) and `example_inputs` (real inputs including SymInt) - Writes weights to `weight_meta.py` and inputs to `input_meta.py` - Updates `utils.py` to handle dict-style `example_inputs`
|
Thanks for your contribution! |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This PR separates model weights (parameters + buffers) from real input tensors during extraction, so that weight_meta.py only contains weights and input_meta.py only contains inputs.