MORPH
SIGN IN SIGN UP
huggingface / pytorch-image-models UNCLAIMED

The largest collection of PyTorch image encoders / backbones. Including train, eval, inference, export scripts, and pretrained weights -- ResNet, ResNeXT, EfficientNet, NFNet, Vision Transformer (ViT), MobileNetV4, MobileNet-V3 & V2, RegNet, DPN, CSPNet, Swin Transformer, MaxViT, CoAtNet, ConvNeXt, and more

0 0 0 Python

COMMITS

/ benchmark.py
diff_parallel
December 1, 2025
R
Remove apex AMP use from scripts
Ross Wightman committed
fee045e
July 3, 2025
G
Update no_grad usage to inference_mode if possible
GuillaumeErhard committed
8630947
January 20, 2025
J
Fix metavar for --input-size
Josua Rieder committed
634b68a
October 18, 2024
R
Cleanup some amp related behaviour to better support different (non-cuda) devices
Ross Wightman committed
1766a01
October 2, 2024
R
Add --torchcompile-mode args to train, validation, inference, benchmark scripts
Ross Wightman committed
4d4bdd6
August 23, 2023
R
MobileOne and FastViT weights on HF hub, more code cleanup and tweaks, features_only working. Add reparam flag to validate and benchmark, support reparm of all models with fuse(), reparameterize() or switch_to_deploy() methods on modules
Ross Wightman committed
5242ba6
January 15, 2023
F
Davit update formatting and fix grad checkpointing (#7)
Fredo Guan committed
81ca323
December 9, 2022
R
Merge remote-tracking branch 'origin/main' into refactor-imports
Ross Wightman committed
d5e7d6b
R
Add a deprecation phase to module re-org
Ross Wightman committed
cda39b3
December 7, 2022
R
Add EVA FT results, hopefully fix BEiT test failures
Ross Wightman committed
98047ef
December 2, 2022
R
Update scripts to support torch.compile(). Make --results_file arg more consistent across benchmark/validate/inference. Fix #1570
Ross Wightman committed
dbe7531
November 22, 2022
R
Add crop_mode for pretraind config / image transforms. Add support for dynamo compilation to benchmark/train/validate
Ross Wightman committed
9da7e3a
September 23, 2022
R
Refactor device handling in scripts, distributed init to be less 'cuda' centric. More device args passed through where needed.
Ross Wightman committed
87939e6
August 26, 2022
R
Add --fast-norm arg to benchmark.py, train.py, validate.py
Ross Wightman committed
ff6a919
July 19, 2022
R
Add bulk_runner script and updates to benchmark.py and validate.py for better error handling in bulk runs (used for benchmark and validation result runs). Improved batch size decay stepping on retry...
Ross Wightman committed
0dbd935
July 7, 2022
R
Reorg benchmark.py import
Ross Wightman committed
4670d37
R
Add --no-retry flag to benchmark.py to skip batch_size decay and retry on error. Fix #1226. Update deepspeed profile usage for latest DS releases. Fix # 1333
Ross Wightman committed
28e0152
July 1, 2022
R
move dataconfig before script, scripting killing metadata now (PyTorch 1.12? just nvfuser?)
Ross Wightman committed
34f382f
June 9, 2022
R
Move aot-autograd opt after model metadata used to setup data config in benchmark.py
Ross Wightman committed
2d7ab06
June 8, 2022
X
add --aot-autograd
Xiao Wang committed
ca991c1
February 28, 2022
R
Significant model refactor and additions:
Ross Wightman committed
372ad5f
January 26, 2022
R
Merge remote-tracking branch 'origin/master' into norm_norm_norm
Ross Wightman committed
95cfc9b
January 24, 2022
R
Update benchmark and validate scripts to output results in JSON with a fixed delimiter for use in multi-process launcher
Ross Wightman committed
cf43343
January 21, 2022
K
refactor: remove if-condition
kozistr committed
56a6b38
January 17, 2022
R
Add --fuser arg to train/validate/benchmark scripts to select jit fuser type
Ross Wightman committed
f0f9ecc
December 14, 2021
R
Add drop args to benchmark.py
Ross Wightman committed
683fba7
October 28, 2021
R
Add new 50ts attn models to benchmark/meta csv files
Ross Wightman committed
aaff2d8
R
Fixed botne*t26 model results, add some 50ts self-attn variants
Ross Wightman committed
1e17863
October 25, 2021
R
Don't run profile if model is torchscripted
Ross Wightman committed
71f00bf
October 24, 2021
R
Add activation count to fvcore based profiling in benchmark.py
Ross Wightman committed
5882e62
October 20, 2021
R
Support either deepspeed or fvcore for flop profiling
Ross Wightman committed
f7325c7
October 19, 2021
R
Add `--bench profile` mode for benchmark.py to just run deepspeed detailed profile on model
Ross Wightman committed
6625379
R
Add train size override and deepspeed GMACs counter (if deepspeed installed) to benchmark.py
Ross Wightman committed
13a8bf7
August 18, 2021
R
Optimizer improvements, additions, cleanup
Ross Wightman committed
ac469b5
May 4, 2021
R
Merge pull request #555 from MichaelMonashev/patch-1
Ross Wightman committed
137a374
April 16, 2021
R
Fix #566, summary.csv writing to pwd on local_rank != 0. Tweak benchmark mem handling to see if it reduces likelihood of 'bad' exceptions on OOM.
Ross Wightman committed
e15e68d
April 11, 2021
M
Argument description fixed
Michael Monashev committed
0be1fa4
April 2, 2021
R
Some further create_optimizer_v2 tweaks, remove some redudnant code, add back safe model str. Benchmark step times per batch.
Ross Wightman committed
37c71a5
April 1, 2021
R
Update benchmark script to add precision arg. Fix some downstream (DeiT) compat issues with latest changes. Bump version to 0.4.7
Ross Wightman committed
2886827
March 6, 2021
R
Add img_size to benchmark output
Ross Wightman committed
4445eaa
March 1, 2021
R
Benchmark models listed in txt file. Add more hybrid vit variants for testing
Ross Wightman committed
0706d05
February 23, 2021
R
Add benchmark.py script, and update optimizer factory to be more friendly to use outside of argparse interface.
Ross Wightman committed
0e16d4e