surya-koritala / Glyd
Lossless AI compression: 33% less GPU memory, bit for bit.
Why this job: matched by what the repo talks about: memory, lossless, gpu, compression
What it does
Glyd keeps a bf16 model's weights and KV cache in fewer bits in GPU memory and rebuilds the exact values as the model runs. It is the same model in about a third less memory, so it fits a smaller GPU or serves more requests on the same one.
Get started
No single install command. Read the steps in the README (opens in new tab)
- Maintenance
- Active
- Last commit 26 days ago
- Stars
- 200
- On GitHub
- Licence
- Custom
- Read before business use
- Downloads for
- 2 systems
- Mac, Linux
- Latest version
- v0.32.1
- Released 26 days ago
- Built with
- Rust
- 54% of the code
Is it looked after?
- Contributors
- Not stated
- Last 90 days
- Issues
- 0% closed
- 1 still open
- Pull requests
- 112 merged
- 1 waiting
- Releases
- 67
- 30+ in the last year
More: download files, languages, topics and project record
Download files, v0.32.1
- Mac3 files, 22 downloads
- Linux6 files, 47 downloads
Languages
- Rust 54%
- Shell 21%
- Python 17%
- HTML 6%
Topics
- benchmark
- c-abi
- cli
- compression
- cuda
- data-compression
- deduplication
- delta-compression
- entropy-coding
- gpu
- kv-cache
- llm
Project record
- JobAgent memory
- StatusListed
- Commits1,172
- Forks2
- Watching0
- Stars in 28 days+42
- CreatedSep 19, 2026
- First seen by usNov 7, 2026
- CheckedNov 8, 2026