Image Classification
Transformers
Safetensors
resnet
nsfw
sfw
safety
image-safety
moderation
Not-For-All-Audiences
Instructions to use najicreator90856/is-it-nsfw_ai-moderator with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use najicreator90856/is-it-nsfw_ai-moderator with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("image-classification", model="najicreator90856/is-it-nsfw_ai-moderator") pipe("https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/hub/parrots.png")# Load model directly from transformers import AutoImageProcessor, AutoModelForImageClassification processor = AutoImageProcessor.from_pretrained("najicreator90856/is-it-nsfw_ai-moderator") model = AutoModelForImageClassification.from_pretrained("najicreator90856/is-it-nsfw_ai-moderator", device_map="auto") - Notebooks
- Google Colab
- Kaggle
π Is It NSFW? AI Moderator (ResNet-18)
is-it-nsfw_ai-moderator is a high-precision, lightweight image classification model designed for real-time content moderation. Based on the ResNet-18 architecture, it provides an exceptional balance between inference speed and detection accuracy.
π Key Highlights
- Accuracy: 99.44% on validation sets.
- Optimized for Speed: Runs flawlessly on consumer-grade hardware (tested on GTX 1060).
- Fine-tuned for Reality: Unlike many models, this one has been specifically fine-tuned to reduce false positives on "tricky" SFW content (fashion, illustrations, complex lighting) while maintaining zero tolerance for explicit content.
π Evaluation Results
| Metric | Value |
|---|---|
| Accuracy | 99.44% |
| Eval Loss | 0.0278 |
π» How to Use (Python)
from transformers import AutoImageProcessor, AutoModelForImageClassification
import torch
from PIL import Image
model_id = "najicreator90856/is-it-nsfw_ai-moderator"
processor = AutoImageProcessor.from_pretrained(model_id)
model = AutoModelForImageClassification.from_pretrained(model_id)
def predict(image_path):
image = Image.open(image_path).convert("RGB")
inputs = processor(images=image, return_tensors="pt")
with torch.no_grad():
logits = model(**inputs).logits
predicted_label = logits.argmax(-1).item()
return model.config.id2label[predicted_label]
# print(predict("your_image.jpg"))
- Point of Contact: catcomp83@gmail.com
β Support & Donations
If you find this model useful for your messenger, community, or commercial project, please consider supporting the development. Your donations help pay for the compute power needed to keep this project alive and updated.
π³ Standard Donations
- DonationAlerts: https://www.donationalerts.com/r/najicreator
πͺ Crypto Support
| Asset | Network | Address |
|---|---|---|
| USDT | TON | UQDcC3o7u4LS07pr-cQI-Gxx3hPjXKrRrNQdGBj0WV7SADe3 |
| USDT | SOL | 77EtoHbBaoTowoQ7sR48k8QhuSHjWUnRjouFh1juCkiA |
| USDT | TRON (TRC20) | TLB4ZqS6HG1jkgyHNuQWtbhj4o4ypbn3p8 |
| USDT/ETH | ERC20 | 0x1e838c1f4c7580bb2f34703f9e9b0af1b1ae92fd |
| BTC | Bitcoin | 12AJT86N1V1D4g2Ee9PWBJG2z8DoHpjHbQ |
| ETH | Ethereum | 0x1e838c1f4c7580bb2f34703f9e9b0af1b1ae92fd |
Disclaimer: This model is for moderation and safety purposes. Users must comply with local laws and regulations regarding the handling of sensitive content.
- Downloads last month
- 29