Update README.md

README.md

@@ -22,6 +22,7 @@ model-index:
     - name: Accuracy
       type: accuracy
       value: 0.9742489270386266
+pipeline_tag: image-classification
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -34,20 +35,26 @@ It achieves the following results on the evaluation set:
 - Loss: 0.0868
 - Accuracy: 0.9742
 
-## Model description
+## Inference example
 
-More information needed
+```python
+from transformers import pipeline
 
-## Intended uses & limitations
+classifier = pipeline(model="lxyuan/vit-xray-pneumonia-classification")
 
-More information needed
+# image taken from https://www.news-medical.net/health/What-is-Viral-Pneumonia.aspx
+classifier("https://d2jx2rerrg6sh3.cloudfront.net/image-handler/ts/20200618040600/ri/650/picture/2020/6/shutterstock_786937069.jpg")
 
-## Training and evaluation data
+>>>
+[{'score': 0.990334689617157, 'label': 'PNEUMONIA'},
+ {'score': 0.009665317833423615, 'label': 'NORMAL'}]
 
-More information needed
+```
 
 ## Training procedure
 
+Notebook link: [here](https://github.com/LxYuan0420/nlp/blob/main/notebooks/ViT-xray-classification.ipynb)
+
 ### Training hyperparameters
 
 The following hyperparameters were used during training:
@@ -62,6 +69,44 @@ The following hyperparameters were used during training:
 - lr_scheduler_warmup_ratio: 0.1
 - num_epochs: 15
 
+```python
+from transformers import EarlyStoppingCallback
+
+training_args = TrainingArguments(
+    output_dir="vit-xray-pneumonia-classification",
+    remove_unused_columns=False,
+    evaluation_strategy="epoch",
+    save_strategy="epoch",
+    logging_strategy="epoch",
+    learning_rate=5e-5,
+    per_device_train_batch_size=16,
+    gradient_accumulation_steps=4,
+    per_device_eval_batch_size=16,
+    num_train_epochs=15,
+    save_total_limit=2,
+    warmup_ratio=0.1,
+    load_best_model_at_end=True,
+    metric_for_best_model="eval_loss",
+    greater_is_better=False,
+    fp16=True,
+    push_to_hub=True,
+    report_to="tensorboard"
+)
+
+early_stopping = EarlyStoppingCallback(early_stopping_patience=3)
+
+trainer = Trainer(
+    model=model,
+    args=training_args,
+    data_collator=data_collator,
+    train_dataset=train_ds,
+    eval_dataset=val_ds,
+    tokenizer=processor,
+    compute_metrics=compute_metrics,
+    callbacks=[early_stopping],
+)
+```
+
 ### Training results
 
 | Training Loss | Epoch | Step | Validation Loss | Accuracy |
@@ -86,4 +131,4 @@ The following hyperparameters were used during training:
 - Transformers 4.30.2
 - Pytorch 1.9.0+cu102
 - Datasets 2.12.0
-- Tokenizers 0.13.3
+- Tokenizers 0.13.3