Gameplay-Classcode-10

Gameplay-Classcode-10 is a vision-language model fine-tuned from google/siglip2-base-patch16-224 using the SiglipForImageClassification architecture. It classifies gameplay screenshots or thumbnails into one of ten popular video game titles.

Classification Report:
                precision    recall  f1-score   support

      Among Us     0.9990    0.9920    0.9955      1000
  Apex Legends     0.9737    0.9990    0.9862      1000
      Fortnite     0.9960    0.9910    0.9935      1000
 Forza Horizon     0.9990    0.9820    0.9904      1000
     Free Fire     0.9930    0.9860    0.9895      1000
Genshin Impact     0.9831    0.9890    0.9860      1000
    God of War     0.9930    0.9930    0.9930      1000
     Minecraft     0.9990    0.9990    0.9990      1000
        Roblox     0.9832    0.9960    0.9896      1000
      Terraria     1.0000    0.9910    0.9955      1000

      accuracy                         0.9918     10000
     macro avg     0.9919    0.9918    0.9918     10000
  weighted avg     0.9919    0.9918    0.9918     10000

The model predicts one of the following game categories:

• 0: Among Us
• 1: Apex Legends
• 2: Fortnite
• 3: Forza Horizon
• 4: Free Fire
• 5: Genshin Impact
• 6: God of War
• 7: Minecraft
• 8: Roblox
• 9: Terraria

---

Run with Transformers 🤗

!pip install -q transformers torch pillow gradio

import gradio as gr
from transformers import AutoImageProcessor, SiglipForImageClassification
from PIL import Image
import torch

# Load model and processor
model_name = "prithivMLmods/Gameplay-Classcode-10"  # Replace with your actual model path
model = SiglipForImageClassification.from_pretrained(model_name)
processor = AutoImageProcessor.from_pretrained(model_name)

# Label mapping
id2label = {
    0: "Among Us",
    1: "Apex Legends",
    2: "Fortnite",
    3: "Forza Horizon",
    4: "Free Fire",
    5: "Genshin Impact",
    6: "God of War",
    7: "Minecraft",
    8: "Roblox",
    9: "Terraria"
}

def classify_game(image):
    """Predicts the game title based on the gameplay image."""
    image = Image.fromarray(image).convert("RGB")
    inputs = processor(images=image, return_tensors="pt")

    with torch.no_grad():
        outputs = model(**inputs)
        logits = outputs.logits
        probs = torch.nn.functional.softmax(logits, dim=1).squeeze().tolist()

    predictions = {id2label[i]: round(probs[i], 3) for i in range(len(probs))}
    predictions = dict(sorted(predictions.items(), key=lambda item: item[1], reverse=True))
    return predictions

# Gradio interface
iface = gr.Interface(
    fn=classify_game,
    inputs=gr.Image(type="numpy"),
    outputs=gr.Label(label="Game Prediction Scores"),
    title="Gameplay-Classcode-10",
    description="Upload a gameplay screenshot or thumbnail to identify the game title (Among Us, Fortnite, Minecraft, etc.)."
)

# Launch the app
if __name__ == "__main__":
    iface.launch()

---

Intended Use

This model can be used for:

• Automatic tagging of gameplay content for streamers and creators
• Organizing gaming datasets
• Enhancing searchability in gameplay video repositories
• Training AI systems for game-related content moderation or recommendations