Vision Transformer (fastforks)

Vision Transformer (Dosovitskiy et al. 2020): 16x16 patches, pre-norm transformer encoder, class token.

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/vit_base_patch16_224.augreg2_in21k_ft_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
vit_base_patch16_224.augreg2_in21k_ft_in1k@1 vit_base_patch16_224.augreg2_in21k_ft_in1k.safetensors 32aa17d6e17b Apache-2.0 yes

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/vit_base_patch16_224.augreg2_in21k_ft_in1k/blob/063c6c38a5d8510b2e57df480445e94b231dad2c/README.md

Input size

fixed 224x224: the input is resized (short side 248) and centre-cropped to it Tested up to (portable, fp16, RTX 5060 Ti 16 GB): vit_base_patch16_224.augreg2_in21k_ft_in1k 224x224 (0.2 GiB peak allocated, 0.3 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
vit_base_patch16_224.augreg2_in21k_ft_in1k 224x224 6.33 5.48 3.02 1.66 3.3x 1.8x 531 / 548 205 / 390 200 / 248

Startup and first call

model load (s): original portable kernels first call (s): original portable kernels
vit_base_patch16_224.augreg2_in21k_ft_in1k 2.49 0.73 2.05 0.28 0.00 0.07

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)
vit_base_patch16_224.augreg2_in21k_ft_in1k 224x224 3.02 ms 2.10 ms 1.14 ms 1.66 ms 7.26 s 28.6 s 205 / 390 238 / 266 200 / 248 394 / 420

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
vit_base_patch16_224.augreg2_in21k_ft_in1k ImageNet-1k val, 5000 random images, top-1 / top-5 % 85.88 / 97.70 85.90 / 97.70 85.84 / 97.70 85.88 / 97.68

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

model = classification.load("vit_base_patch16_224.augreg2_in21k_ft_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: vit_base_patch16_224.augreg2_in21k_ft_in1k.

Kernels backend

classification.load("vit_base_patch16_224.augreg2_in21k_ft_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.

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Collection including fastforks/vit