Traditional archiving formats (ZIP, Tar.gz) were designed in the 90s for floppy disks. Today, they are the main bottleneck of IT infrastructure. They force you to download entire 500GB datasets, cannot stream individual files from the cloud, and cause extremely expensive GPUs to sit idle waiting for data.

NRA (Neural Ready Archive) is a next-generation binary format. It combines enterprise-grade deduplication, ultra-fast Zstd compression, and B+ Tree indexing so you can train neural networks directly from the public cloud.

⚡ Why Legacy Formats Are Dead (Our Benchmarks)

We ran a stress test on 60,000 small files (CIFAR-10):

🏆 Competitive Radar: NRA vs Everyone

NRA v4.5 is the only format that scores maximum across all technical parameters:

🚀 Try It Now: Train Online Without Downloading

Use our ready-made dataset on Hugging Face

pip install nra torch

import nra
import torch
from torch.utils.data import Dataset, DataLoader

class NraStreamDataset(Dataset):
    def __init__(self, url):
        self.url = url
        # The manifest downloads in 150ms. The archive itself stays in the cloud!
        self.file_ids = nra.CloudArchive(url).file_ids()
        self._archive = None
        
    def __len__(self):
        return len(self.file_ids)
        
    def __getitem__(self, idx):
        if self._archive is None:
            self._archive = nra.CloudArchive(self.url)
        raw_bytes = self._archive.read_file(self.file_ids[idx])
        return torch.tensor([len(raw_bytes)], dtype=torch.float32)

# 🤗 Our ready-made dataset on Hugging Face (NRA format)
dataset = NraStreamDataset(
    "https://huggingface.co/datasets/zevatov/nra-cifar10/resolve/main/cifar10.nra"
)
loader = DataLoader(dataset, batch_size=256, num_workers=4)

for batch in loader:
    # Training starts at second 0. Zero bytes on your SSD!
    pass

🤗 Open the dataset on Hugging Face →

🏗️ How Cloud Architecture Works (Zero-Disk I/O)

PyTorch DataLoader
       │
       ▼
NRA Core (Rust)  ──── B+ Tree manifest lookup O(1)
       │
       ▼
HTTP GET Range: bytes=X-Y  →  HuggingFace / S3
       │
       ▼
Zstd decompress in RAM  →  raw bytes  →  GPU tensor

Zero Disk I/O. Your SSD is never touched. Data flows: Cloud → RAM → GPU.

🛠️ The NRA Ecosystem

1. Python SDK (pip install nra): Integration into PyTorch and TensorFlow.
2. NRA CLI (cargo install nra-cli): Console utility for servers.
3. NRA GUI: Desktop application for visual archive management. (In development: zevatov/nra-manager-pro)
4. FUSE Mount: Mount .nra archives as virtual drives (nra-cli mount).
5. 🤗 HuggingFace: zevatov/nra-cifar10 — ready-to-use NRA dataset.

🗺️ Roadmap

📚 Deep Documentation

For full architectural documentation, whitepapers, and archiving benchmarks, visit our Official GitHub Repository.

License

The nra-core, nra-cli, and nra-python components are distributed under the MIT license.