README.md

metadata

library_name: fastforks
pipeline_tag: image-classification
license: other
license_name: bsd-3-clause-code-weights-unstated
license_link: >-
  https://github.com/pytorch/vision/blob/1f1a920d685d46dfbe187934b785b7ecb7a182f2/LICENSE
tags:
  - classification
  - image-classification
  - cuda
  - pytorch
  - fastforks

EfficientNet (fastforks)

EfficientNet (Tan and Le 2019): compound-scaled MBConv networks with squeeze-excite.

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/pytorch/vision at 1f1a920d685d.

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
efficientnet_b0.tv_in1k@1	efficientnet_b0_rwightman-7f5810bc.pth	7f5810bc96de	BSD-3-Clause (assumed from code)	yes
efficientnet_b4.tv_in1k@1	efficientnet_b4_rwightman-23ab8bcd.pth	23ab8bcd5bdb	BSD-3-Clause (assumed from code)	yes

Licence

Original-code strings, recorded as they are (research-use package; check them before any other use). Code: BSD-3-Clause. Files: .

Code licence: BSD-3-Clause; the authors do not state a separate weights licence.

Training data: ImageNet-1k (ILSVRC 2012 classification).

Licence evidence: https://github.com/pytorch/vision/blob/1f1a920d685d46dfbe187934b785b7ecb7a182f2/LICENSE

Input size

• efficientnet_b0.tv_in1k: fixed 224x224: the input is resized (short side 256) and centre-cropped to it
• efficientnet_b4.tv_in1k: fixed 380x380: the input is resized (short side 384) and centre-cropped to it Tested up to (portable, fp16, RTX 5060 Ti 16 GB): efficientnet_b0.tv_in1k 224x224 (0.0 GiB peak allocated, 0.1 reserved); efficientnet_b4.tv_in1k 380x380 (0.1 GiB peak allocated, 0.1 reserved)

Speed and memory

Classification, synthetic, 224x224

model	time (ms): original fp32	original	portable	kernels	kernels speed-up: vs original	vs portable	peak memory (MiB, allocated / reserved): original	portable	kernels
efficientnet_b0.tv_in1k	5.64	7.91	4.98	0.769	7.3x	6.5x	64 / 84	48 / 96	78 / 118
efficientnet_b4.tv_in1k	11.09	15.66	10.0	1.999	5.5x	5.0x	147 / 174	89 / 114	111 / 180

Startup and first call

model	load (s): original	portable	kernels	first call (s): original	portable	kernels
efficientnet_b0.tv_in1k	1.25	0.17	1.53	2.86	0.38	0.06
efficientnet_b4.tv_in1k	0.41	0.37	1.94	1.72	0.06	0.07

torch.compile and TensorRT

model	input	portable ms	compile ms	compile s (cold)	TensorRT fp16 ms	ONNX + build s	kernels ms	peak MiB (alloc / reserved): portable	compile	TensorRT (device)
efficientnet_b0.tv_in1k	224x224	4.98	2.9	9.1	0.5	26	0.769	48 / 96	80 / 96	38

Fixed-size speed and peak memory, all models

model	mode	224x224
efficientnet_b0.tv_in1k	portable	5.0 ms / 48 / 84
efficientnet_b0.tv_in1k	kernels	0.9 ms / 81 / 164
efficientnet_b4.tv_in1k	portable	9.9 ms / 92 / 120
efficientnet_b4.tv_in1k	kernels	2.2 ms / 125 / 200

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 authors' code 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

Parity with the original code

Every variant is checked against golden 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("efficientnet_b0.tv_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: efficientnet_b0.tv_in1k, efficientnet_b4.tv_in1k.

Kernels backend

classification.load("efficientnet_b0.tv_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.