Instructions to use SenseTime/deformable-detr with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use SenseTime/deformable-detr with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("object-detection", model="SenseTime/deformable-detr")# Load model directly from transformers import AutoImageProcessor, AutoModelForObjectDetection processor = AutoImageProcessor.from_pretrained("SenseTime/deformable-detr") model = AutoModelForObjectDetection.from_pretrained("SenseTime/deformable-detr") - Notebooks
- Google Colab
- Kaggle
Conflict between the config file and model code in Deformable DETR
#3
by ZeQ - opened
Hello,
I noticed a conflict in the Deformable DETR implementation. The id2label mapping in the config file does not have the last value representing the "no-object" class, which conflicts with the Deformable DETR implementation in the Hugging Face codebase. Specifically, in the modeling_deformable_detr.py file, within the DeformableDetrLoss(nn.Module) class, at line 2184, it is indicated that the "no-object" class must be the last ID (logits.shape[-1] - 1).
Thank you!
cc @qubvel-hf