README.md

---
datasets:
- mikehemberger/plantnet300K
language:
- en
pipeline_tag: text-to-image
base_model:
- google/embeddinggemma-300m
---
# Plantgen
| ![purple flower](https://raw.githubusercontent.com/groloch/plantgen/refs/heads/main/docs/purple_flower.png) | ![blue flower](https://raw.githubusercontent.com/groloch/plantgen/refs/heads/main/docs/blue_flower.png) | ![orange flower](https://raw.githubusercontent.com/groloch/plantgen/refs/heads/main/docs/orange_flower.png) | ![red flower](https://raw.githubusercontent.com/groloch/plantgen/refs/heads/main/docs/red_flower.png) | ![white flower](https://raw.githubusercontent.com/groloch/plantgen/refs/heads/main/docs/white_flower.png) | ![yellow flower](https://raw.githubusercontent.com/groloch/plantgen/refs/heads/main/docs/yellow_flower.png) |
|:---:|:---:|:---:|:---:|:---:|:---:|
| purple flower | blue flower | orange flower | red flower | white flower | yellow flower |
## Overview
This is a re-implementation of the full [Stable Diffusion 3 pipeline](https://arxiv.org/pdf/2403.03206) trained on the [PlantNet300k](https://github.com/plantnet/PlantNet-300K) dataset.

## Usage

To use the pipeline, you need to clone the [official repo](https://github.com/groloch/plantgen).

[Here](https://github.com/groloch/plantgen/blob/main/src/plantgen/scripts/hf_hub_test.py) are some examples on how to load the pipeline from the hub. Note that you need gated access to embeddinggemma-300m.