Instructions to use snap-research/efficientformer-l1-300 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- timm
How to use snap-research/efficientformer-l1-300 with timm:
import timm model = timm.create_model("hf_hub:snap-research/efficientformer-l1-300", pretrained=True) - Notebooks
- Google Colab
- Kaggle
EfficientFormer (backbone) + SegFormer (head)
Hi Bearnardd,
Thank you for sharing pretrained efficientformer (classification) models.
I'd like to test efficientformer by change head of efficientformer into another head (i.e. SegFormer).
Is there any way to easily try to test for it?
Hi @skyeom !
At the moment EfficientFormer is still not on main branch since I had a little bug in the implementation which I just fixed but I also got several comments to apply in my PR (which you can follow https://github.com/huggingface/transformers/pull/20459) so it will probably be finished in the next week. In case of changing the head of the model you can use EfficientFormer as a base model and add custom head on top of it. You can take a look at this blog post on how to achieve that https://towardsdatascience.com/adding-custom-layers-on-top-of-a-hugging-face-model-f1ccdfc257bd).