openvinotoolkit / openvinotoolkit/nncf
[TorchFX] Support weight quantization for operations with `weight_port_id != 1`
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 1.2k
- Forks
- 305
- Avg merge
- 1d 9h
- Merged PRs (30d)
- 27
Description
🚀 Feature request
The task is to support weight quantization for operations with weight_port_id != 1.
To do so, one should update the method _get_input_scale_shape of the MinMax torch_fx_backend.py https://github.com/openvinotoolkit/nncf/blob/develop/nncf/quantization/algorithms/min_max/torch_fx_backend.py#L181-L182 as it done for the torch backend https://github.com/openvinotoolkit/nncf/blob/develop/nncf/quantization/algorithms/min_max/torch_backend.py#L197-L202
Feature Use Case
No response
Are you going to submit a PR?
- Yes I'd like to help by submitting a PR!
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start with _get_input_scale_shape in nncf/quantization/algorithms/min_max/torch_fx_backend.py and compare it with the corresponding implementation in torch_backend.py. Trace how weight_port_id is handled, then verify that operations with weight_port_id != 1 receive weight quantization support without changing existing behavior.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python, pytorch
- Domain
- machine-learning
- Issue type
- Feature
- Difficulty
- 2/5
- Estimated time
- 1-3 hours
- Activity status
- Quiet
- Clarity
- Clearly specified
- Newbie friendliness
- 68/100