Veriafy - Universal File Classification Platform

Veriafy now supports GPU acceleration for dramatically faster classification. Here's everything you need to know.

Performance Improvements

With GPU acceleration enabled:

Operation	CPU	GPU	Speedup
Image classification	45ms	4ms	11x
Document classification	80ms	12ms	7x
Batch (100 images)	4.5s	180ms	25x

Supported GPUs

Veriafy supports:

•**NVIDIA**: RTX 30/40 series, A100, H100
•**AMD**: RX 6000/7000 series (ROCm)
•**Apple**: M1/M2/M3 (Metal)

Configuration

Docker

docker run -d --gpus all -p 8080:8080 \
  -e VERIAFY_GPU=1 \
  veriafy/veriafy:latest-gpu

Python SDK

client = Veriafy(gpu=True) result = client.classify("image.jpg", model="veriafy/nsfw-classifier") ```

CLI

veriafy classify image.jpg --gpu

Batch Processing

GPU acceleration shines with batch processing:

# Process 1000 images in seconds
results = client.classify_batch(
    files=image_paths,
    model="veriafy/nsfw-classifier",
    batch_size=64  # Optimal for most GPUs
)

Memory Requirements

GPU VRAM	Recommended Batch Size
4 GB	16
8 GB	32
16 GB	64
24 GB+	128

Cloud Deployment

For cloud deployments, we recommend:

•**AWS**: p3.2xlarge or g4dn.xlarge
•**GCP**: n1-standard-8 + T4
•**Azure**: NC6s_v3

See our [GPU deployment guide](/docs/advanced/gpu) for detailed instructions.

GPU Acceleration: 10x Faster Classification