convnextv2 / README.md
Flor's picture
card: update from fastforks 1cbf1671
773180f verified
|
Raw History Blame Contribute Delete
5.86 kB
metadata
library_name: fastforks
pipeline_tag: image-classification
license: cc-by-nc-4.0
tags:
  - classification
  - image-classification
  - cuda
  - pytorch
  - fastforks
  - fastforks-card

ConvNeXt V2 (fastforks)

ConvNeXt V2 (Woo et al. 2023): ConvNeXt blocks with global response normalisation, FCMAE pre-training.

This repo is the model card only: it holds no weights. fastforks downloads the authors' weights from the links below and checks them against the SHA-256 in the model card. Code: https://github.com/kleinicke/fastforks. Kernels for the kernels backend: https://huggingface.co/fastforks/kernels. Original code: https://github.com/huggingface/pytorch-image-models at 0df212b369a5.

Author weight repos: https://huggingface.co/timm/convnextv2_base.fcmae_ft_in22k_in1k.

Task: classification. Tier: stable. Outputs: logits. Backends: portable, kernels.

Input: one RGB image. Output: ImageNet-1k class logits.

Variants

Weights are the authors' files (not hosted here). SHA-256 shortened to 12 hex digits; the full hashes are in card.toml.

variant weights (author link) SHA-256 weights licence commercial use
convnextv2_base.fcmae_ft_in22k_in1k@1 convnextv2_base.fcmae_ft_in22k_in1k.safetensors 3b04d2fd7f52 CC-BY-NC-4.0 no

Licence

Original-code strings, recorded as they are (research-use package; check them before any other use). Code: Apache-2.0.

Training data: ImageNet-21k / ImageNet-22k (full ImageNet, fall 2011 release), ImageNet-1k (ILSVRC 2012 classification).

Licence evidence: https://huggingface.co/timm/convnextv2_base.fcmae_ft_in22k_in1k/blob/9e250ed8f88b436472d2b24af26f82a8aa8c719d/README.md

Input size

fixed 224x224: the input is resized (short side 256) and centre-cropped to it Tested up to (portable, fp16, RTX 5060 Ti 16 GB): convnextv2_base.fcmae_ft_in22k_in1k 224x224 (0.3 GiB peak allocated, 0.4 reserved)

Speed and memory

Classification, synthetic

model input time (ms): original fp32 original portable kernels kernels speed-up: vs original vs portable peak memory (MiB, allocated / reserved): original portable kernels
convnextv2_base.fcmae_ft_in22k_in1k 224x224 10.7 14.6 10.2 2.94 3.7x 3.5x 542 / 566 299 / 376 318 / 380

Startup and first call

model load (s): original portable kernels first call (s): original portable kernels
convnextv2_base.fcmae_ft_in22k_in1k 2.70 1.23 2.12 0.63 0.03 0.16

torch.compile and TensorRT

model input time: portable compile TensorRT fp16 kernels preparation: compile (cold first call) ONNX + engine build peak MiB (alloc / reserved): portable compile kernels TensorRT (engine + torch)
convnextv2_base.fcmae_ft_in22k_in1k 224x224 10.2 ms 3.85 ms 2.13 ms 2.94 ms 14.4 s 35.1 s 299 / 376 324 / 356 318 / 380 486 / 518

Fixed-size cells: time (ms) / peak allocated / peak reserved memory (MiB).

Hardware, protocol and caveats: benchmarks (RTX 5060 Ti 16 GB, sm_120). Original = the original code named above at the precision of portable (fp16, or (fp32) where fp16 is unsupported or not finite); original fp32 = its default fp32 run. Peak memory is the torch allocator's.

Accuracy

On public datasets

fastforks: the model's own resize / crop on the GPU; original: the library's pretrained loader and its PIL eval transform. Original = the original code named under Details, same checkpoint and protocol.

variant dataset original fp32 original fp16 portable kernels note
convnextv2_base.fcmae_ft_in22k_in1k ImageNet-1k val, 5000 random images, top-1 / top-5 % 87.32 / 98.18 87.32 / 98.18 87.10 / 98.14 87.10 / 98.14

Measured on NVIDIA GeForce RTX 5060 Ti, torch 2.14.1+cu130, fastforks 939adfe8, 2026-11-06.

Parity with the original code

Every variant is checked against reference outputs of the original code at its pinned commit. CPU, float32: max |diff| <= 0.002 vs the original. GPU, float16 (the default), portable and kernels: mean |diff| <= 0.02 and p99 <= 0.08 (absolute error of the output logits).

Usage

from fastforks import classification

# non-commercial weights: run with FASTFORKS_NONCOMMERCIAL=1 in the environment
model = classification.load("convnextv2_base.fcmae_ft_in22k_in1k")  # mode="auto": the preferred backend from recorded measurements
result = model("photo.jpg")
result.top(5)  # top-5 classes
result.backend  # what ran, e.g. Backend(mode='portable', device='cuda:0', dtype='float16')

Install with pip install fastforks (see https://github.com/kleinicke/fastforks). Other variants: convnextv2_base.fcmae_ft_in22k_in1k.

Kernels backend

classification.load("convnextv2_base.fcmae_ft_in22k_in1k", mode="kernels") runs our CUDA kernels (cudnn_hacc, fvm_conv, fvm_extra, fvm_mb, fvm_r3). It needs a prebuilt kernel for your GPU architecture (downloaded from https://huggingface.co/fastforks/kernels on first use) and a recent NVIDIA driver; if there is none it raises BackendUnavailableError with the reason and never falls back silently. mode="auto" picks kernels when it can run and logs the choice.