From the model card
Fields below are copied from the tags and counters on the HuggingFace repository timm/convnext_femto.d1_in1k at our last fetch. They are set by the uploader, not verified by us; rows with no tag are omitted. How this page is made.
- Publisher (HF namespace)
- timm
- Pipeline tag
- image-classification
- Library
- timm, Transformers
- Framework tags
- PyTorch
- Weight formats
- safetensors
- License tag
apache-2.0— read the license file in the repo before relying on it- Papers cited
- arXiv:2201.03545
- Datasets declared
- imagenet-1k
- Downloads (HF counter at last fetch)
- 362,600
- Likes (HF counter at last fetch)
- 1
- Model card
- https://huggingface.co/timm/convnext_femto.d1_in1k
Use cases
- Image classification on severely compute-constrained hardware
- Backbone for lightweight object detection or segmentation
- Ablation studies on ConvNeXt scaling behavior
- Mobile or microcontroller image classification pipelines
Pros
- Apache-2.0 license
- Extremely small and fast — suitable for edge inference
- timm compatibility provides easy fine-tuning and feature extraction
- ConvNeXt architectural improvements over older ConvNet designs
Cons
- Femto scale trades accuracy heavily for speed — not competitive on challenging datasets
- ImageNet-1K pretraining — may not transfer well without fine-tuning on domain data
- Less community documentation than larger ConvNeXt variants
- timm dependency required; not native Transformers pipeline
Tags
timmpytorchsafetensorsimage-classificationtransformersdataset:imagenet-1karxiv:2201.03545license:apache-2.0region:us