Create App.py

App.py

@@ -0,0 +1,116 @@
+import gradio as gr
+from transformers import pipeline
+import numpy as np
+
+from datasets import load_dataset
+dataset = load_dataset("Asseh/Ball_Classification")
+
+from sklearn.model_selection import train_test_split
+
+dataset = dataset["train"].train_test_split(test_size=0.2)
+
+from transformers import AutoImageProcessor
+
+checkpoint = "google/vit-base-patch16-224-in21k"
+image_processor = AutoImageProcessor.from_pretrained(checkpoint)
+
+from torchvision.transforms import RandomResizedCrop, Compose, Normalize, ToTensor
+
+normalize = Normalize(mean=image_processor.image_mean, std=image_processor.image_std)
+size = (
+    image_processor.size["shortest_edge"]
+    if "shortest_edge" in image_processor.size
+    else (image_processor.size["height"], image_processor.size["width"])
+)
+_transforms = Compose([RandomResizedCrop(size), ToTensor(), normalize])
+
+def transforms(examples):
+    examples["pixel_values"] = [_transforms(img.convert("RGB")) for img in examples["image"]]
+    del examples["image"]
+    return examples
+
+dataset = dataset.with_transform(transforms)
+
+from transformers import DefaultDataCollator
+
+data_collator = DefaultDataCollator()
+
+def compute_metrics(eval_pred):
+    predictions, labels = eval_pred
+    predictions = np.argmax(predictions, axis=1)
+    return accuracy.compute(predictions=predictions, references=labels)
+
+from transformers import AutoModelForImageClassification, TrainingArguments, Trainer
+
+model = AutoModelForImageClassification.from_pretrained(
+    checkpoint,
+    num_labels=len(labels),
+    id2label=id2label,
+    label2id=label2id,
+)
+
+training_args = TrainingArguments(
+    output_dir="Ball_Classification",
+    remove_unused_columns=False,
+    evaluation_strategy="epoch",
+    save_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=3,
+    warmup_ratio=0.1,
+    logging_steps=10,
+    load_best_model_at_end=True,
+    metric_for_best_model="accuracy",
+    push_to_hub=False,
+)
+
+trainer = Trainer(
+    model=model,
+    args=training_args,
+    data_collator=data_collator,
+    train_dataset=dataset["train"],
+    eval_dataset=dataset["test"],
+    tokenizer=image_processor,
+    compute_metrics=compute_metrics,
+)
+
+trainer.train()
+
+from huggingface_hub import notebook_login
+notebook_login()
+
+trainer.push_to_hub()
+
+from transformers import pipeline
+
+classifier = pipeline("image-classification", model="Ball_Classification")
+classifier(image)
+
+from transformers import pipeline
+
+classifier = pipeline("image-classification", model="Asseh/Ball_Classification")
+
+from transformers import pipeline
+
+classifier = pipeline("image-classification", model="Asseh/Ball_Classification")
+
+# Function to classify images into 7 classes
+def image_classifier(inp):
+    # Dummy classification logic
+    # Generating random confidence scores for each class
+    confidence_scores = np.random.rand(7)
+    # Normalizing confidence scores to sum up to 1
+    confidence_scores /= np.sum(confidence_scores)
+    # Creating a dictionary with class labels and corresponding confidence scores
+    classes = ['american_football', 'baseball', 'basketball', 'billiard_ball', 'bowling_ball','cricket_ball', 'football', 'golf_ball', 'hockey_ball', 'hockey_puck', 'rugby_ball', 'shuttlecock', 'table_tennis_ball', 'tennis_ball', 'volleyball']
+    result = {classes[i]: confidence_scores[i] for i in range(15)}
+    return result
+
+# Creating Gradio interface
+demo = gr.Interface(fn=image_classifier, inputs="image", outputs="label")
+
+if __name__ == "__main__":
+    demo.launch()
+