Alternatives

Products that do what Lossless compressor that can shrink llama3 to 68% does

1. 18F
   80% faster, 50% less memory, 0% loss of accuracy Llama finetuning▲385

   Hi HN! I'm just sharing a project I've been working on during the LLM Efficiency Challenge - you can now finetune Llama with QLoRA 5x faster than Huggingface's original implementation on your own local GPU. Some highlights: 1. Manual autograd engine - hand derived backprop steps. 2. QLoRA / LoRA 80% faster, 50% less memory. 3. All kernels written in OpenAI's Triton language. 4. 0% loss in accuracy - no approximation methods - all exact. 5. No change of hardware necessary. Supports NVIDIA GPUs since 2018+. CUDA 7.5+. 6. Flash Attention support via Xformers. 7. Supports 4bit and 16bit…

   2023 · github.com

2. 2AI
   An interactive guide to compression basics▲256

   2017 · unwttng.com

3. 3
   TurboQuant▲295

   New LLM compression algorithm by Google

   Mar 2026 · research.google

4. 4
   Unsloth▲241

   Finetune LLMs 2x faster, 80% less memory

   2025

5. 5TA
   Trigrad, a novel image compression with interesting results▲175

   2015 · ruarai.github.io

6. 6JI
   JavaScript image compressor▲41

   2017 · github.com

7. 7L3
   Llama 3.3 70B Sparse Autoencoders with API access▲201

   2024 · goodfire.ai

8. 8IM
   I made a Image compressor that is free▲5

   2025 · image-compressor-five-azure.vercel.app

9. 9OS
   Open-source load balancer for llama.cpp▲151

   Stateful load balancer customized for llama.cpp (with a reverse proxy).

   2024 · github.com

10. 10XC
    XPress Compress v1.0 – A config-less compression algorithm▲19

    2019 · github.com

11. 11UF
    UltraCompress – first mathematically lossless 5-bit LLM compression▲6

    May 2026 · github.com

12. 12IM
    I made a new compression algorithm▲9

    Jan 2026 · github.com

13. 13OA
    Overfitted a 900KB Transformer to Compress a 100MB CSV into 7MB▲112

    I built an experiment that uses an overfitted transformer and arithmetic coding to compress individual files. Instead of training the model to generalize, I train a 900KB transformer to memorize a single file and predict the next byte. Those predictions are fed into an arithmetic coder to produce the compressed output. On a 100MB NYC taxi CSV, it compresses to about 7MB (~0.5 bits/byte). On a 100MB slice of enwik9, it compresses to about 21MB (~1.68 bits/byte). It's pretty slow right now (roughly 20–30 minutes of training and 45 minutes each for compression and decompression on my…

    Jun 2026

14. 14ED
    Excssive, drag and drop CSS compressor▲19

    2011 · excssive.com

15. 15TB
    TurboPFor – Bringing Fastest Integer Compression Incl. SIMD to Java▲14

    2015 · github.com

16. 16DR
    DivANS – Rust simd compression algorithm demo in WASM in the browser▲36

    2018 · dropbox.github.io

17. 17CV
    Compress videos by up to 60%▲6

    2016 · compressify.herokuapp.com

18. 18LL
    Lightweight lossless file minifier/optimizer▲23

    2014 · github.com

19. 19AB
    A better zip decoder▲35

    2019 · github.com

20. 20TE
    Tzar, Easy compression and extraction for any compression format▲7

    2021 · github.com

21. 21CS
    Claude-Stream-Compactor▲5

    Feb 2026 · github.com

22. 22CL
    Cssfact – Lossy CSS compressor based on binary matrix factorization▲7

    2024 · github.com

23. 23LS
    Limitd – simple daemon for rate limiting HA applications▲15

    2015 · github.com

24. 24AL
    A LESS compiler for all platforms▲9

    2011 · wearekiss.com

Ranked by how close each launch is in meaning, then by votes. Refine with a description →