This is a fine-tuned ViT (Vision Transformer) model for Pokémon image classification. The model is trained to classify upto Gen9 (1025) Pokémon images.

Intended Use

This model is designed for image classification tasks, specifically for identifying Pokémon characters. It can be used for:

Pokémon-themed apps
Educational projects
Pokémon identification in images

Note: The model is not designed for general-purpose image classification.

How to Use

Here's how you can load and use the model with the Hugging Face transformers library:

python

from transformers import ViTForImageClassification, ViTImageProcessor
from PIL import Image
import torch

# Define the device
device = "cuda" if torch.cuda.is_available() else "cpu"

# Load the model and image processor
model_id = "skshmjn/Pokemon-classifier-gen9-1025"
model = ViTForImageClassification.from_pretrained(model_id).to(device)
image_processor = ViTImageProcessor.from_pretrained(model_id)

# Load and process an image
img = Image.open('test.jpg').convert("RGB")
inputs = image_processor(images=img, return_tensors='pt').to(device)

# Make predictions
outputs = model(**inputs)
predicted_id = outputs.logits.argmax(-1).item()
predicted_pokemon = model.config.id2label[predicted_id]

# Print predicted class
print(f"Predicted Pokémon Pokédex number: {predicted_id+1}")
print(f"Predicted Pokémon: {predicted_pokemon}")