README.md

---
library_name: onnx
pipeline_tag: feature-extraction
license: bsd-3-clause
tags:
- eidora
- eidora-model-zoo
- onnx
- onnxruntime
- embeddings
- image
- alexnet
- imagenet
- compute:light
- modality:image
base_model: torchvision/alexnet-imagenet1k-v1
datasets:
- imagenet-1k
metrics:
- cosine-similarity
model-index:
- name: alexnet_imagenet1k_4096
  results: []
---

# alexnet_imagenet1k_4096

alexnet_imagenet1k_4096 is a light image embedding model for fast first-pass visual grouping in EIDORA. Choose it when you want a small, classic baseline that runs comfortably on ordinary laptops.

## Best For

- Fast baseline visual grouping.
- Small or exploratory image projects on ordinary laptops.
- Regression testing the EIDORA ONNX package pipeline.

## Not Ideal For

- Fine-grained visual similarity where modern self-supervised models perform better.
- Text, video, or audio inputs.
- Production-quality semantic image retrieval when a stronger model is acceptable.

## Compute Tier

Light: small download, low memory, faster CPU runtime. Intended for laptop CPU use and large first-pass projects.

## Inputs

- `image`: required image input from `media_source`.

## Output

The primary output is `embedding`, a float32 tensor shaped `[batch, 4096]`. Embeddings are already normalized and are intended for cosine similarity.

## Usage In EIDORA

EIDORA shows this package as a light image embedding model in the Model Zoo. Use it for discovery maps, grouping, retrieval, and related embedding workflows.

## Usage Examples

Download the complete package from Hugging Face and keep the generated files together:

```python
from huggingface_hub import snapshot_download

package_dir = snapshot_download("eidora/alexnet_imagenet1k_4096")
# EIDORA reads config.yaml, README.md, and model.onnx from this folder.
```

## Preprocessing

- `image`: rescale uses `1/255`; normalize inside the ONNX graph with mean `[0.485, 0.456, 0.406]` and std `[0.229, 0.224, 0.225]`.

## Authorship And Citation

This ONNX package was produced by EIDORA from the original AlexNet ImageNet model. EIDORA converted the model to ONNX and is not the original model creator. Please cite ImageNet Classification with Deep Convolutional Neural Networks and the original model repository when using this converted model.

Original model: https://proceedings.neurips.cc/paper/2012/hash/c399862d3b9d6b76c8436e924a68c45b-Abstract.html

Original paper: https://proceedings.neurips.cc/paper/2012/hash/c399862d3b9d6b76c8436e924a68c45b-Abstract.html

Authors: Alex Krizhevsky, Ilya Sutskever, Geoffrey E. Hinton

```bibtex
@inproceedings{krizhevsky2012imagenet,
  title={ImageNet Classification with Deep Convolutional Neural Networks},
  author={Krizhevsky, Alex and Sutskever, Ilya and Hinton, Geoffrey E.},
  booktitle={Advances in Neural Information Processing Systems},
  year={2012}
}
```

## Training Data And Provenance

Base model: `torchvision/alexnet-imagenet1k-v1`. Source repository: https://pytorch.org/vision/stable/models/generated/torchvision.models.alexnet.html. Known training data: ImageNet-1K supervised classification data. Package payload size: 228020495 bytes.

## Evaluation And Validation

The package validation checks that the ONNX graph loads with ONNX Runtime CPU execution, runs the declared fixtures, returns finite float32 embeddings with the declared shape, and matches the artifact hash recorded in `eidora-model.yaml`.

## Limitations And Safety

AlexNet is an older supervised ImageNet model. It is useful as a lightweight baseline, but modern CLIP, DINOv2, SigLIP, or domain-specific models will usually produce stronger semantic similarity.

## License And Attribution

This package uses license `bsd-3-clause`. Upstream license: BSD-3-Clause for TorchVision code; ImageNet-trained weights distributed by PyTorch under their documented model terms. Converted to ONNX for EIDORA from the TorchVision AlexNet ImageNet weights.

## Version

Package version: 1.0.0. ONNX opset: 17. Exporter: eidora-onnx-exporter 0.1.0.