Readme updated

README.md

@@ -4,6 +4,12 @@ tags:
 model-index:
 - name: diffusion-detection
   results: []
+license: apache-2.0
+datasets:
+- imagenet-1k
+metrics:
+- accuracy
+pipeline_tag: image-classification
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -11,21 +17,32 @@ should probably proofread and complete it, then remove this comment. -->
 
 # diffusion-detection
 
-This model was trained from scratch on the None dataset.
+This model was trained to distinguish real world images (negative) from machine generated ones (postive).
 
-## Model description
+## Model usage
 
-More information needed
+```python
+from transformers import BeitImageProcessor, BeitForImageClassification
+from PIL import Image
 
-## Intended uses & limitations
+processor = BeitImageProcessor.from_pretrained('TimKond/diffusion-detection')
+model = BeitForImageClassification.from_pretrained('TimKond/diffusion-detection')
 
-More information needed
+image = Image.open("2980_saltshaker.jpg")
+
+inputs = processor(images=image, return_tensors="pt")
+outputs = model(**inputs)
+logits = outputs.logits
+
+predicted_class_idx = logits.argmax(-1).item()
+print("Predicted class:", model.config.id2label[predicted_class_idx])
+```
 
 ## Training and evaluation data
 
-More information needed
+[BEiT-base-patch16-224-pt22k](https://huggingface.co/microsoft/beit-base-patch16-224-pt22k) was loaded as a base model for further fine tuning:
 
-## Training procedure
+As negatives a subsample of 10.000 images from [imagenet-1k](https://huggingface.co/datasets/imagenet-1k) was used. Complementary 10.000 positive images were generated using [Realistic_Vision_V1.4](https://huggingface.co/SG161222/Realistic_Vision_V1.4). 
 
 ### Training hyperparameters
 
@@ -44,4 +61,4 @@ The following hyperparameters were used during training:
 - Transformers 4.29.2
 - Pytorch 1.11.0+cu113
 - Datasets 2.12.0
-- Tokenizers 0.13.3
+- Tokenizers 0.13.3