---
license: apache-2.0
language:
- en
tags:
- microscopy
- denoising
- fluorescence-microscopy
- image-restoration
- 2d
- 3d
---

# FMDNet Weights (`smthooper/fmdnet-weights`)

## Model Summary

FMDNet is a pretrained denoising model for fluorescence microscopy data.
This repository provides checkpoint weights consumed by the `fmdnet` Python package.

Provided files:
- `fmdnet_3d.pt`
- `fmdnet_2d.pt`

## Intended Use

- Denoising fluorescence microscopy images/volumes.
- Load with:
  - `FMDNet.from_pretrained("fmdnet-3d")`
  - `FMDNet.from_pretrained("fmdnet-2d")`

## Out-of-Scope / Non-Intended Use

- Clinical diagnosis or treatment decisions.
- Safety-critical deployment without domain-specific validation.
- Data modalities outside fluorescence microscopy.

## Limitations

- Performance depends on imaging conditions.
- Memory limits may apply for large volumes.

## Training / Fine-Tuning Context

Base checkpoints are intended as general pretrained denoisers.
Fine-tuning is supported by the `fmdnet` package in:
- `supervised`
- `clean`
- `noisy`

## Version Compatibility

| Model repo/tag | Expected package version | Notes |
|---|---|---|
| `main` (current) | `fmdnet==0.0.1` | Initial public release |

## License

Model weights license: Apache-2.0

Code repository license: Apache-2.0

## Citation

If you use this model, please cite:

```text
[ADD FULL PAPER CITATION]