Uploading Models#

Publish model files, configuration, tokenizer assets, and the model card in one revision. A complete release lets a reader download and load the same artifact set later.

1mega models upload OWNER/MODEL ./release . \2  --commit-message "Publish model release"

For a large tree, use the resumable uploader:

mega upload-large-folder OWNER/MODEL ./release --revision main

Use a Hugging Face-compatible client#

The standard huggingface_hub client can create the repository and upload a folder when its endpoint is pointed at MEGA:

1from huggingface_hub import HfApi2 3api = HfApi(endpoint="https://mega.tensorplay.cn", token="YOUR_MEGA_TOKEN")4api.create_repo("OWNER/MODEL", repo_type="model", exist_ok=True)5api.upload_folder(6    repo_id="OWNER/MODEL",7    repo_type="model",8    folder_path="./release",9    commit_message="Publish model release",10)

The same workflow is available through hf upload after setting HF_ENDPOINT. See Hugging Face Compatibility for supported operations and limits.

Before tagging a release, verify the model from a clean download and ensure the card, license, configuration, tokenizer, index, and every referenced shard are in the same commit. See Large Files and Model Repositories.