
marlin
by IST-DASLab (organization)
FP16xINT4 LLM inference kernel that can achieve near-ideal ~4x speedups up to medium batchsizes of 16-32 tokens. (description from the project’s GitHub page)
Repository facts
- Stars
- 1,164
- Forks
- 95
- Open issues and pull requests
- 37
- Primary language
- Python
- License
- Apache License 2.0 (Apache-2.0)
- Owner
- IST-DASLab · Organization
- Created
- Jan 17, 2024
- Last push
- Sep 4, 2024
Topics
Related projects
Projects that share this one’s GitHub topics.
Rank 64: hiyouga/LlamaFactory
Unified Efficient Fine-Tuning of 100+ LLMs & VLMs (ACL 2024)
75,561 starsPythonApache-2.0Last push Oct 10, 2026
Rank 431: ymcui/Chinese-LLaMA-Alpaca
中文LLaMA&Alpaca大语言模型+本地CPU/GPU训练部署 (Chinese LLaMA & Alpaca LLMs)
18,926 starsPythonApache-2.0Last push Apr 19, 2026
Rank 959: bitsandbytes-foundation/bitsandbytes
Accessible large language models via k-bit quantization for PyTorch.
8,526 starsPythonMITLast push Sep 7, 2026
Rank 2,899: stochasticai/xTuring
Build, personalize and control your own LLMs. From data pre-processing to fine-tuning, xTuring provides an easy way to personalize open-source LLMs. Join our discord community: ht…
2,674 starsPythonApache-2.0Last push Oct 18, 2026
Rank 3,232: dvmazur/mixtral-offloading
Run Mixtral-8x7B models in Colab or consumer desktops
2,336 starsPythonMITLast push Apr 8, 2024
Rank 4,053: Tencent/AngelSlim
Model compression toolkit engineered for enhanced usability, comprehensiveness, and efficiency.
1,762 starsPythonOther licenseLast push Oct 14, 2026