---
tags:
- depth-to-robot
- image-to-image
- cyclegan
---

# Depth2Robot GAN Model

This model transforms depth maps into robot-style images using CycleGAN.
<div style="display: flex; justify-content: center;">
  <img src="testOutput/depth2image/custom_real.png" alt="depth map" title="depth map" width="45%">
  <img src="testOutput/depth2image/custom_fake.png" alt="stylized depth map" title="stylized depth map" width="45%">
</div>
<div style="display: flex; justify-content: center;">
  <img src="testOutput/image2depth/custom_real.png" alt="depth map" title="depth map" width="45%">
  <img src="testOutput/image2depth/custom_fake.png" alt="stylized depth map" title="stylized depth map" width="45%">
</div>
## Model Description

- This model was trained on depth maps and robot images.
- It converts grayscale depth maps to colorful robot-style imagery.
- Trained using CycleGAN architecture.

## Usage

```python
import torch
from huggingface_hub import hf_hub_download

# Download the model
model_path = hf_hub_download(repo_id="Borcherding/depth2AnythingCycleGAN_RobotsV2", filename="latest_net_G.pth")

# Load the model (you need to define the Generator class)
model = Generator()
model.load_state_dict(torch.load(model_path), strict=False)
model.eval()

# Use the model for inference
# ...
```