DeiT (fastforks)
DeiT (Touvron et al. 2021): ViT trained on ImageNet-1k only.
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/deit_base_patch16_224.fb_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 |
|---|---|---|---|---|
deit_base_patch16_224.fb_in1k@1 |
deit_base_patch16_224.fb_in1k.safetensors | cd2da27b74ed |
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-1k (ILSVRC 2012 classification).
Licence evidence: https://huggingface.co/timm/deit_base_patch16_224.fb_in1k/blob/b78cc5532a69df6bcad9c3a8d76653fd20b31ac6/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): deit_base_patch16_224.fb_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 |
|---|---|---|---|---|---|---|---|---|---|---|
deit_base_patch16_224.fb_in1k |
224x224 | 6.39 | 5.46 | 2.98 | 1.64 | 3.3x | 1.8x | 531 / 548 | 203 / 386 | 300 / 370 |
Startup and first call
| model | load (s): original | portable | kernels | first call (s): original | portable | kernels |
|---|---|---|---|---|---|---|
deit_base_patch16_224.fb_in1k |
2.52 | 0.72 | 1.96 | 0.31 | 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) |
|---|---|---|---|---|---|---|---|---|---|---|---|
deit_base_patch16_224.fb_in1k |
224x224 | 2.98 ms | 2.09 ms | 1.12 ms | 1.64 ms | 7.27 s | 28.3 s | 203 / 386 | 238 / 266 | 300 / 370 | 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 |
|---|---|---|---|---|---|---|
deit_base_patch16_224.fb_in1k |
ImageNet-1k val, 5000 random images, top-1 / top-5 % | 83.12 / 95.92 | 83.14 / 95.88 | 83.22 / 95.84 | 83.16 / 95.84 |
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("deit_base_patch16_224.fb_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: deit_base_patch16_224.fb_in1k.
Kernels backend
classification.load("deit_base_patch16_224.fb_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.