Can I upload private models to Hugging Face?

Yes, the skill allows you to configure repository visibility to either Public or Private during the setup configuration.

Do I need a Hugging Face token to use this?

Yes, you must have a Hugging Face write token configured in your environment to create repositories and upload files.

What formats does funsloth-upload support?

It supports LoRA adapters, merged 16-bit models, and various GGUF quantization levels such as Q4_K_M, Q5_K_M, and Q8_0.

Does it generate the model documentation automatically?

Yes, it creates a comprehensive README.md including YAML metadata, training hyperparameters, loss metrics, and copy-pasteable usage examples.

Funsloth Model Upload

Name: Funsloth Model Upload
Author: chrisvoncsefalvay

bychrisvoncsefalvay

•

データサイエンスとML

Automates the generation of professional model cards and the deployment of fine-tuned LLMs to the Hugging Face Hub.

概要

Funsloth-upload is a specialized utility designed to streamline the post-training workflow for machine learning engineers. It handles the complexities of exporting models into multiple formats—including LoRA adapters, merged 16-bit weights, and GGUF quantizations—while automatically generating comprehensive, standard-compliant README documentation. By integrating directly with the Hugging Face Hub, it ensures that your fine-tuned models are shared with all necessary metadata, usage examples, and performance metrics, making it an essential tool for completing the Unsloth fine-tuning lifecycle.

主な機能

Automated Hugging Face Hub repository creation and file upload
Configurable GGUF quantization levels (Q4_K_M, Q5_K_M, Q8_0)
4 GitHub stars
Multi-format export including LoRA, merged 16-bit, and GGUF
Professional model card (README.md) generation with YAML metadata
Built-in validation and verification of uploaded assets

ユースケース

Deploying fine-tuned models to Hugging Face after training with Unsloth
Standardizing documentation for internal or public model repositories
Creating quantized GGUF versions of models for use in Ollama or llama.cpp

概要

主な機能

Automated Hugging Face Hub repository creation and file upload
Configurable GGUF quantization levels (Q4_K_M, Q5_K_M, Q8_0)
4 GitHub stars
Multi-format export including LoRA, merged 16-bit, and GGUF
Professional model card (README.md) generation with YAML metadata
Built-in validation and verification of uploaded assets

ユースケース

Deploying fine-tuned models to Hugging Face after training with Unsloth
Standardizing documentation for internal or public model repositories
Creating quantized GGUF versions of models for use in Ollama or llama.cpp