Update README.md

README.md

@@ -1,9 +1,34 @@
-
-# ViT(google/vit-base-patch16-224-in21k) finetuned on document classifaction task over DocLayNet-base dataset
-
+---
+datasets:
+- pierreguillou/DocLayNet-base
+metrics:
+- accuracy
+base_model:
+- facebook/deit-base-distilled-patch16-224
+library_name: transformers
+tags:
+- vision
+- document-layout-analysis
+- document-classification
+- deit
+- doclaynet
+---
+
+# Data-efficient Image Transformer(DeiT) for Document Classification(DocLayNet)
+
+This model is a fine-tuned Data-efficient Image Transformer(DeiT) for document layout classification based on the DocLayNet dataset.
+
+Trained on images of the document categories from DocLayNet dataset where the categories namely(with their indexes) are :
+
+{'financial_reports': 0,
+ 'government_tenders': 1,
+ 'laws_and_regulations': 2,
+ 'manuals': 3,
+ 'patents': 4,
+ 'scientific_articles': 5} 
 ## Model description
 
-ViT(google/vit-base-patch16-224-in21k) finetuned on document classification
+DeiT(facebook/deit-base-distilled-patch16-224) finetuned on document classification
 
 
 
@@ -17,18 +42,21 @@ https://huggingface.co/datasets/pierreguillou/DocLayNet-base
 hyperparameters:
 
 {
-    'batch_size': 64,
+    'batch_size': 128,
     'num_epochs': 20,
     'learning_rate': 1e-4,
-    'weight_decay': 0.05,
-    'warmup_ratio': 0.2,
+    'weight_decay': 0.1,
+    'warmup_ratio': 0.1,
     'gradient_clip': 0.1,
     'dropout_rate': 0.1,
     'label_smoothing': 0.1
     'optmizer': 'AdamW'
 }
+
 ## Evaluation results
-Test Loss: 0.8622, Test Acc: 81.36%
+
+Test Loss: 0.8134, Test Acc: 81.56%
+
 
 ## Usage