NielsRogge
Hi @FFY0 🤗 Niels here from the open-source team at Hugging Face. I discovered your work on Arxiv and noticed that you are preparing the code for release: https://github.com/FFY0/DefensiveKV. That's fantastic! Your paper, "Taming the Fragility of KV Cache Eviction in LLM Inference," got featured on Hugging Face's daily papers: https://huggingface.co/papers/2510.13334. The paper page lets people discuss your paper and find associated artifacts. You can also claim the paper as yours, which will show up on your public profile at HF, and add GitHub and project page URLs. We would be thrilled if you would consider hosting your DefensiveKV implementation and any associated artifacts (such as code for the method itself, or examples demonstrating its application) on the 🤗 Hub once they are ready for release. This would significantly improve their discoverability and visibility within the AI community. We can add relevant tags so that people can easily find them. ## Uploading the implementation Your method for KV cache eviction could be very valuable to the community. You could potentially host the implementation as a custom model or library, for example, by leveraging the [PyTorchModelHubMixin](https://huggingface.co/docs/huggingface_hub/package_reference/mixins#huggingface_hub.PyTorchModelHubMixin) class, which adds `from_pretrained` and `push_to_hub`. This allows users to easily download and integrate your method into their LLM inference pipelines. See here for a guide: https://huggingface.co/docs/hub/models-uploading. Alternatively, a Hugging Face Space could be used to demonstrate your method. We can provide a ZeroGPU [grant](https://huggingface.co/docs/hub/en/spaces-gpus#community-gpu-grants), which gives you A100 GPUs for free. Let me know if you're interested/need any help regarding this once your code and artifacts are ready for release! Cheers, Niels ML Engineer @ HF 🤗