Click any tag below to further narrow down your results
Links
tufazip is a Transformer-based compressor for the enwik9 benchmark, reported at 105,388,353 bytes total once the decoder archive is counted. The repo gives the exact commands, file hashes, and hardware needed to reproduce the decode.
- Compressed enwik9 is 100,958,324 bytes, with a 4,430,029-byte decoder ZIP, for a ranked total of 105,388,353 bytes.
- It is an online Transformer initialized from a quantized offline Cycle2 prior, and needs two NVIDIA B200 GPUs.
- Compression takes about 631,548 seconds and decompression about 645,748 seconds, so it is far from a practical everyday tool.
JPEG XL is now supported in Chrome and Firefox, and the author argues it's the best archival format thanks to lossless quality and JPEG transcoding that shrinks legacy files about 20% without any loss. AVIF still wins for lossy web delivery, so JXL is a storage and lossless play more than a web-image replacement.
- Lossless JPEG transcoding cuts legacy .jpg files by about 20% and can reconstruct the original bit-for-bit, so archives can shrink without generation loss.
- For lossless images, AVIF comes out 30–122% larger than JXL and decodes more slowly, while AVIF still beats JXL on lossy file size.
- JXL supports up to 32 bits per channel, HDR, and custom channels, though browser gain-map support is still catching up and hardware acceleration is only planned.