tensorflow / tensorflow/model-optimization
Quantization Aware Training support for KerasLayer
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 1.6k
- Forks
- 349
- Avg merge
- 3d 2h
- Merged PRs (30d)
- 1
Description
System information
- TensorFlow version: latest
Motivation
Trying to combine Quantization Aware Training with transfer learning.
I am using feature vectors from tfhub, so when I download the feature vector it comes as a hub.KerasLayer.
following this tutorial https://www.tensorflow.org/model_optimization/guide/combine/cqat_example, Clustering is working well on the model but when it comes to QAT there is always an error that the layer is not supported. Is there any workaround or plan for support?
As examples:
Model architecture that I am trying to use
Just for reference, the KerasLayer is the feature vector of MobileNet V2
Error when trying QAT
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 by reproducing the Quantization Aware Training tutorial linked in the issue with the MobileNet V2 tfhub.KerasLayer and inspect the reported unsupported-layer error. The issue names no repository files or tests; done would require a defined support path or documented workaround for applying QAT to KerasLayer-based transfer-learning models.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- keras, python, tensorflow
- Domain
- machine-learning
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 25/100