tensorflow / tensorflow/models
attention_ocr: Support TensorFlow Serving's Classification API
@xavigibert is already working on this.
Since Sep 17, 2020.
- Dominant language
- Python
- Stars
- 77.7k
- Forks
- 44.8k
- PR merge metrics
- No merged PRs in 30d
Description
Prerequisites
Please answer the following question for yourself before submitting an issue.
- I checked to make sure that this feature has not been requested already.
1. The entire URL of the file you are using
https://github.com/tensorflow/models/blob/master/research/attention_ocr/python/model_export.py
2. Describe the feature you request
Add a Classification SignatureDef to the SavedModel, so that the model can be served through the Tensorflow Serving Classification API as described in https://stackoverflow.com/questions/63931981/error-when-serving-attention-ocr-model-error-expected-one-or-two-output-ten?sem=2
3. Additional context
The Classification API expects the output to contain 2 tensors, one for predictions and the other for scores. The existing code exports 17 tensors. The tensors that need to be exported for the classification API are endpoints.predicted_text and endpoints.normalized_seq_conf.
4. Are you willing to contribute it? (Yes or No)
yes
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.
Assessment
This issue has not been assessed yet.