search

Kohulan / DECIMER-Image_Transformer

DECIMER Image Transformer is a deep-learning-based tool designed for automated recognition of chemical structure images. Leveraging transformer architectures, the model converts chemical images into SMILES strings, enabling the digitization of chemical data from scanned documents, literature, and patents.
MIT License
214 stars 52 forks source link

Can I get the trained model as a .h5 or tf model format? #49

Closed weathon closed 1 year ago

weathon commented 1 year ago

Issue Type

Support

Source

GitHub (source)

DECIMER Image Transformer Version

2.0.0

OS Platform and Distribution

No response

Python version

No response

Current Behaviour?

Can I get the trained model as a .h5 or tf model format that I can load? I did not find the model in the repo

Which images caused the issue? (This is mandatory for images related issues)

No response

Standalone code to reproduce the issue

# Not code related question
# Can I get the trained model as a .h5 or tf model format that I can load? I did not find the model in the repo

Relevant log output

No response

Code of Conduct

Kohulan commented 1 year ago

Hi @weathon ,

You can find the models here: https://zenodo.org/record/7624994, which is already listed on ReadME.md. It is in a .pb format which works with Tensorflow without any issues. Hope that helps 😃 .

weathon commented 1 year ago

Ah I found it in read me. Thanks. It took me a while though because it is an icon not a link so I skipped it when skimming. I suggest change it into a link. But anyway, thank you!

On Fri, Jun 16, 2023 at 12:17 AM Kohulan Rajan @.***> wrote:

Hi @weathon https://github.com/weathon ,

You can find the models here: https://zenodo.org/record/7624994, which is already listed on ReadME.md. It is in a .pb format which works with Tensorflow without any issues. Hope that helps 😃 .

  • Kohulan

— Reply to this email directly, view it on GitHub https://github.com/Kohulan/DECIMER-Image_Transformer/issues/49#issuecomment-1594224809, or unsubscribe https://github.com/notifications/unsubscribe-auth/AJ3CXHGIYWLIZGSB3KAJY63XLQB63ANCNFSM6AAAAAAZI2R72M . You are receiving this because you were mentioned.Message ID: @.***>

weathon commented 1 year ago

This is a tf model not a Keras model so I am not sure if I can take only a few layers out. I want only the decoder, is it possible t do that?

weathon commented 1 year ago

@Kohulan

Kohulan commented 1 year ago

@weathon, We don't use Keras .h5 format, maybe you could use some converters?