How do I save a model, using Huggingface, after adding tensorflow layers to it. I am able to save it using native tensorflow, but not Huggingface. Specifially, I am using “model = TFBertModel.from_pretrained(‘bert-base-cased’)”, and then adding a few layers to it. After I add the layers it is no longer saveable using transformers. I get the error “AttributeError: ‘Functional’ object has no attribute ‘save_model’”. Please help.
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| Saving custom and/or finetuned models without the HUB | 3 | 1034 | March 2, 2022 | |
| How to save my model to use it later | 15 | 169587 | November 10, 2024 | |
| How to save and load fine-tune model | 4 | 24398 | October 25, 2021 | |
| Save custom transformer as PreTrainedModel | 1 | 884 | September 7, 2021 | |
| Save a tensorflow model with a transformer layer | 0 | 1642 | January 21, 2022 |