ReasonCLIP Models
Collection
Visually Grounded Commonsense Reasoning Supervision for CLIP (ECCV 2026) • 25 items • Updated • 1
# Load model directly
from transformers import AutoProcessor, AutoModelForZeroShotImageClassification
processor = AutoProcessor.from_pretrained("RISys-Lab/ReasonCLIP-B32-S0-Des")
model = AutoModelForZeroShotImageClassification.from_pretrained("RISys-Lab/ReasonCLIP-B32-S0-Des")Quick Links
Model Details
- Model: ReasonCLIP-B32-S0-Des
- Base model: openai/clip-vit-base-patch32
- Architecture: CLIP ViT-B/32
- Image resolution: 224
- Training stage: Stage 0 - Descriptive
- Training data: CC12M-Refined
Method
Resources
- GitHub: RISys-Lab/ReasonCLIP
- Paper: arXiv:2606.26794
Usage
from transformers import CLIPModel, CLIPProcessor
model_id = "RISys-Lab/ReasonCLIP-B32-S0-Des"
model = CLIPModel.from_pretrained(model_id)
processor = CLIPProcessor.from_pretrained(model_id)
For the full checkpoint list, see the ReasonCLIP model card.
- Downloads last month
- 31
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("zero-shot-image-classification", model="RISys-Lab/ReasonCLIP-B32-S0-Des") pipe( "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/hub/parrots.png", candidate_labels=["animals", "humans", "landscape"], )