facebook / image-classification updated 1 year ago

convnextv2-base-22k-384

Name: convnextv2-base-22k-384
Author: facebook

ConvNeXt V2 model pretrained using the FCMAE framework and fine-tuned on the ImageNet-22K dataset at resolution 384x384. It was introduced in the paper ConvNeXt V2: Co-designing and Scaling ConvNets with Masked Autoencoders by Woo et al. and first released in this repository.

Params

90 M

Context

—

Downloads 30d

259 K

Likes

Commercial use: allowed apache-2.0 Not gated SAFETENSORS View on Hugging Face ↗

Download history

daily snapshots · 44 days

▲ 202 K in the last 30 days (43.8%)

476 K244 K

Aug 8Aug 22Sep 6Sep 20

Can you run it?

Estimated VRAM at 8K context unless noted. Pick your hardware to see the verdict per quantization.

File	Quant	Size	Est. VRAM	Verdict on RTX 4090 · 24 GB
model.safetensors	f32	0.4 GB	0.9 GB	✅ Runs comfortably

Estimate: file size × 1.1 + KV cache at 8K + 0.5 GB overhead. Not a benchmark — how we calculate this.

Specifications

Architecture: ConvNextV2ForImageClassification
Parameters: 90 M
Tensor type: F32
Licence: apache-2.0
First seen on the Hub: 2023-02-19
Training datasets: imagenet-22k
Added to our catalog: 2026-08-08

Compare with any image-classification model

Popular comparisons

mobilenetv3_small_100.lamb_in1k 18.0 M efficientnet_b3.ra2_in1k 12.7 M vit-base-patch16-224 7.2 M nsfw_image_detection 3.4 M gender-classification 1.4 M fairface_age_image_detection 1.2 M