Hi @LabRAI 🤗
Niels here from the open-source team at Hugging Face. I discovered your work on Arxiv regarding ReAD (Reinforcement-Guided Capability Distillation) and was wondering whether you would like to submit it to hf.co/papers to improve its discoverability. You can submit it at https://huggingface.co/papers/submit.
The paper page lets people discuss about your work and find related artifacts. I noticed the GitHub link in the paper (https://github.com/LabRAI/ReAD) currently returns a 404; I assume it might be made public soon!
It'd be great to make the resulting distilled checkpoints and the distillation datasets available on the 🤗 hub to improve their discoverability and visibility. We can add tags so that people find them when filtering https://huggingface.co/models and https://huggingface.co/datasets.
Uploading models
Hosting on Hugging Face gives your models more visibility through metadata tags and download statistics. You can use the PyTorchModelHubMixin class to add from_pretrained and push_to_hub methods to your model code, or simply upload checkpoints via the UI.
Uploading datasets
Making your generated distillation data available on 🤗 would also be very beneficial for the community. It allows users to do:
from datasets import load_dataset
dataset = load_dataset("your-hf-org-or-username/your-dataset")Besides easy access, the dataset viewer allows people to explore the data directly in their browser.
Let me know if you're interested or need any guidance!
Cheers,
Niels
ML Engineer @ HF 🤗
Hi @LabRAI 🤗
Niels here from the open-source team at Hugging Face. I discovered your work on Arxiv regarding ReAD (Reinforcement-Guided Capability Distillation) and was wondering whether you would like to submit it to hf.co/papers to improve its discoverability. You can submit it at https://huggingface.co/papers/submit.
The paper page lets people discuss about your work and find related artifacts. I noticed the GitHub link in the paper (
https://github.com/LabRAI/ReAD) currently returns a 404; I assume it might be made public soon!It'd be great to make the resulting distilled checkpoints and the distillation datasets available on the 🤗 hub to improve their discoverability and visibility. We can add tags so that people find them when filtering https://huggingface.co/models and https://huggingface.co/datasets.
Uploading models
Hosting on Hugging Face gives your models more visibility through metadata tags and download statistics. You can use the PyTorchModelHubMixin class to add
from_pretrainedandpush_to_hubmethods to your model code, or simply upload checkpoints via the UI.Uploading datasets
Making your generated distillation data available on 🤗 would also be very beneficial for the community. It allows users to do:
Besides easy access, the dataset viewer allows people to explore the data directly in their browser.
Let me know if you're interested or need any guidance!
Cheers,
Niels
ML Engineer @ HF 🤗