🎉 We are excited to have surpassed 50,000 downloads (stable+nightly) in less than 3 weeks! We've added many new features and models. Expect something cool in the next release! github.com/linkedin/Liger…

🚀🚀 v0.3.0 Release 1. Large Vision Language Model Support (Qwen-VL) 2. Patch Kernels on Model Instances 3. SWIFT Trainer Integration 4. KL Divergence Kernel and extended cross entropy For the full release note, see github.com/linkedin/Liger… We are also hosting our 1st IRL

🚀 Liger Kernel has surpassed **100,000+ downloads** after a month! We're humbled by the many success stories shared by both the research community and enterprises. Our commitment remains strong: we'll continue to support more kernels and models while improving performance.

Exciting talk on LLM deployment and serving!

Our tech report on Liger Kernel is now available on arXiv: arxiv.org/abs/2410.10989 We discuss the advantages of Triton, kernel fusion, kernel optimization (e.g. chunking, in-place, etc), the design philosophy of the interface, testing best practice, and benchmarks on a wide

EMNLP 2025 Noah Lee Thankfully, we now have more than 1,300 open checkpoints in the Hugging Face community🙌 And thanks to groundbreaking work by Liger Kernel team, you can use ORPO with up to 10 times more efficient memory usage🤯 github.com/linkedin/Liger…

Liger Kernel v0.4.0 has arrived! github.com/linkedin/Liger… 1. Full AMD Support: We have partnered with embeddedllm.com to adjust the Triton configuration to fully support AMD! With version 0.4.0, you can run multi-GPU training with 26% higher speed and 60% lower memory

v0.4.1: Gemma 2 Support, CrossEntropy Patching FIx, and GroupNorm github.com/linkedin/Liger… 1. **Gemma 2 Support**: The long pending gemma 2 is finally supported thanks to Chun-Chih Tseng! He has implemented the nasty softcapping in fused linear cross entropy

With Liger-Kernel, you can now fine-tune Qwen2-VL Qwen on Modal by one command with 10%+ throughput and 50% less memory, thanks to the contribution from Tyler Romero! github.com/linkedin/Liger…

What just happened 👀👀

LinkedIn's Engineering blog on Liger Kernel is now live!! linkedin.com/blog/engineeri… Since its initial release in August 2024, Liger-Kernel has grown rapidly across the community, accumulating 3,500+ stars and 350k+ downloads. We have integrated with mainstream training

Lots to announce today! Axolotl 0.6.0 is out with: - new axolotl cli command - updated dependencies to latest on Transformers, TRL, Accelerate, PEFT, bitsandbytes, deepspeed, Liger Kernel - Support for Cut Cross Entropy - ReLoRA optimizer reset fixes - ADOPT optimizer 🧵

Liger Kernel v0.5 github.com/linkedin/Liger… Post Training Loss: Introducing the first open-source optimized post-training losses in Liger Kernel with ~80% memory reduction, featuring DPO, CPO, ORPO, SimPO, JSD, and more. No more OOM nightmares for post-training ML researchers!

Happy Chinese Lunar New Year! The Liger team is excited to announce our milestone of 1,000,000 downloads! Since our initial release in August 2024, we have observed tremendous growth in both usage and community engagement. We appreciate all the recognition and will continue to

🚀 Liger Kernel is Now Live on Glows.ai! We're stoked to announce that Liger has landed on Glows.ai! No more environment setup headaches - just pure, seamless Liger magic at your fingertips. 🔥 Unbeatable Pricing: Get mind-blowing RTX 4090 compute

VeOmni x Liger to enable scalable and efficient training!

Cool work!