Update README.md

README.md

@@ -1,8 +1,39 @@
----
-tags:
-- clip
-library_name: open_clip
-pipeline_tag: zero-shot-image-classification
-license: mit
----
-# Model card for CLIC-ViT-B-16-224-CogVLM
+# Model Card for CLIC-ViT-B-16-224-CogVLM
+
+## Model Details
+
+<!-- Provide the basic links for the model. -->
+
+- **Model-details:** : Fine-tuned with CLIC using CogVLM relabelled 1M dataset
+
+## Model Usage
+### With OpenCLIP
+
+```
+import torch
+from PIL import Image
+import open_clip
+
+model, _, image_processor = open_clip.create_model_and_transforms('hf-hub:nmndeep/CLIC-ViT-B-16-224-CogVLM')
+
+
+image = image_processor(Image.open(urlopen(
+    'https://images.pexels.com/photos/869258/pexels-photo-869258.jpeg?auto=compress&cs=tinysrgb&w=1260&h=750&dpr=1'))).unsqueeze(0)
+
+model.eval()
+
+tokenizer = open_clip.get_tokenizer(f'hf-hub:nmndeep/CLIC-ViT-B-16-224-CogVLM')
+
+texts= ["a diagram", "a dog", "a cat", "snow"]
+text = tokenizer(texts)
+
+with torch.no_grad(), torch.autocast("cuda"):
+    image_features = model.encode_image(image)
+    text_features = model.encode_text(text)
+    image_features /= image_features.norm(dim=-1, keepdim=True)
+    text_features /= text_features.norm(dim=-1, keepdim=True)
+
+    text_probs = (100.0 * image_features @ text_features.T).softmax(dim=-1)
+idx = torch.argmax(text_probs)
+print("Output label:", texts[idx])  
+```