Bulk Upload of Lora Models

My proposal pertains to the bulk uploading of LoRAs.

Current Situation:

I have a collection of Pokémon and Digimon LoRAs that I would like to upload, but this process would take an entire day.

Therefore, I propose a bulk upload feature.

We could implement a system similar to the LoRA Trainer, where the model is uploaded as a zip file.

Requirements for the zip:

Folder structure:

/model/model.safetensors

/images/*.png (Include all images in the folder)

/meta/modeldata.txt

/meta/modeldesc.txt

modeldata.txt contains (preferably via template) all necessary details such as triggers, training details, etc.

modeldesc.txt contains the title, tags, and description.

Each zip would represent a model, which can then be uploaded either via a temporary SFTP or (perhaps more securely) a Google Drive link.

The most important aspect would be providing an example zip file that includes blank templates for the txt data.

An interesting challenge arises if there are different versions of a model. This could potentially be managed with a "Switch" file located in the root of the zip. This would be a txt file with the model ID indicating where the zip should be directed, for example.