Hi @donaldssh 🤗
I'm Niels and work as part of the open-source team at Hugging Face. I discovered your work through Papers with Code as your paper "K-Merge: Online Continual Merging of Adapters for On-device Large Language Models" got accepted to ACL 2026!
The Hugging Face paper page (https://huggingface.co/papers/2510.13537) allows researchers to find artifacts (like models and datasets) linked to papers. You can also claim the paper as yours to show it on your public Hugging Face profile.
I noticed on your GitHub repository that the code is "coming soon!". Once you are ready to publish, would you like to host the pre-trained LoRA adapters and model checkpoints on https://huggingface.co/models?
Hosting on Hugging Face will provide better discoverability and visibility for your work. We can add metadata tags to your model cards so users can easily search and find them, and link them directly to your Hugging Face paper page.
If you are using PEFT/LoRA, they can be easily uploaded and integrated with the peft and transformers libraries, so that users can load them with from_pretrained directly from the Hub. Alternatively, one can use hf_hub_download to fetch individual weights.
Let me know if you're interested or if you need any assistance when the code is ready to be released!
Kind regards,
Niels
Hi @donaldssh 🤗
I'm Niels and work as part of the open-source team at Hugging Face. I discovered your work through Papers with Code as your paper "K-Merge: Online Continual Merging of Adapters for On-device Large Language Models" got accepted to ACL 2026!
The Hugging Face paper page (https://huggingface.co/papers/2510.13537) allows researchers to find artifacts (like models and datasets) linked to papers. You can also claim the paper as yours to show it on your public Hugging Face profile.
I noticed on your GitHub repository that the code is "coming soon!". Once you are ready to publish, would you like to host the pre-trained LoRA adapters and model checkpoints on https://huggingface.co/models?
Hosting on Hugging Face will provide better discoverability and visibility for your work. We can add metadata tags to your model cards so users can easily search and find them, and link them directly to your Hugging Face paper page.
If you are using PEFT/LoRA, they can be easily uploaded and integrated with the
peftandtransformerslibraries, so that users can load them withfrom_pretraineddirectly from the Hub. Alternatively, one can usehf_hub_downloadto fetch individual weights.Let me know if you're interested or if you need any assistance when the code is ready to be released!
Kind regards,
Niels