Commit Graph

Select branches

Hide Pull Requests

0.5.x

0.6.x

attn_update

bits_and_tpu

cleanup_xla_model_fixes

clip_laion2b

convnext_and_copyright

convnextv2

dataset_info

edgenext_csp_and_more

efficientnet_attn

eva

fix_tests

fix_window_trace

fixes-syncbn_pretrain_cfg_resolve

fixes_bce_regnet

focalnet_and_swin_refactor

freeze_norm_revisit

gh-pages

hf_hub_revisit

levit_efficientformer_redux

main

maxxvit_hub

more_attn

more_datasets

more_vit

multi-weight

multi-weight_effnet_convnext

norm_mlp_classifier

norm_norm_norm

parse_kwargs

parse_kwargs2

perceiver

refactor-imports

scaling_vit

script_cleanup

swin_v2

update-test-pytorch

vit_and_bit_test_fixes

vit_relpos_refactor

xsepconv

#1

#1007

#101

#1012

#1014

#1034

#1037

#105

#1055

#1056

#1056

#1058

#1061

#1063

#1063

#1069

#1083

#1091

#1094

#1105

#1112

#1113

#115

#1150

#117

#1187

#1190

#1194

#12

#1208

#1210

#1210

#1212

#1213

#1213

#122

#1222

#1222

#1229

#123

#1230

#1233

#1239

#1245

#1249

#125

#1251

#1253

#1253

#1256

#1259

#1265

#1266

#1270

#1271

#1271

#1273

#1274

#1276

#1277

#1294

#13

#1317

#132

#1322

#1327

#1330

#1330

#1336

#1340

#1345

#1349

#1351

#1354

#136

#1363

#1365

#137

#1381

#1383

#1385

#140

#141

#1414

#1415

#142

#1420

#143

#1439

#1439

#145

#146

#1467

#1471

#1473

#1476

#1479

#148

#1483

#1493

#1494

#1498

#150

#1500

#1502

#1520

#1522

#1525

#1527

#1532

#1538

#154

#1545

#1547

#155

#1552

#1553

#1553

#156

#1562

#1564

#1565

#1575

#1578

#1578

#1581

#1582

#1583

#1586

#1590

#1590

#1592

#1593

#16

#1606

#1612

#1614

#1617

#1617

#1621

#1622

#1624

#1627

#1628

#1628

#163

#1630

#1633

#1641

#1643

#1643

#1645

#1647

#1647

#1649

#1654

#1654

#1655

#1659

#1659

#1662

#1673

#1674

#1677

#1679

#1679

#1680

#1683

#1685

#1687

#1688

#1688

#175

#179

#183

#19

#191

#195

#199

#2

#208

#212

#212

#213

#213

#214

#216

#218

#227

#228

#23

#233

#237

#24

#240

#240

#244

#250

#255

#256

#263

#268

#282

#284

#286

#29

#290

#290

#297

#30

#300

#302

#31

#315

#32

#322

#323

#33

#334

#335

#35

#352

#367

#382

#389

#401

#413

#418

#419

#421

#425

#427

#429

#434

#437

#440

#447

#450

#459

#475

#489

#494

#497

#501

#502

#510

#52

#525

#528

#53

#533

#537

#542

#548

#550

#554

#555

#556

#561

#571

#580

#581

#595

#6

#603

#603

#609

#612

#613

#62

#625

#626

#636

#637

#645

#647

#65

#651

#652

#653

#659

#659

#66

#660

#668

#669

#679

#681

#685

#693

#7

#702

#714

#729

#731

#738

#74

#746

#747

#750

#768

#768

#771

#772

#800

#801

#804

#804

#805

#808

#812

#813

#816

#817

#82

#821

#828

#83

#841

#842

#853

#853

#858

#87

#874

#875

#876

#88

#880

#882

#891

#898

#910

#913

#914

#915

#933

#94

#95

#964

#967

#968

#986

#989

#99

#993

v0.1-attn-weights

v0.1-cadene

v0.1-coat-weights

v0.1-dnf-weights

v0.1-effv2-weights

v0.1-ger-weights

v0.1-hrnet

v0.1-mvit-weights

v0.1-pit-weights

v0.1-regnet

v0.1-repvgg-weights

v0.1-res2net

v0.1-resnest

v0.1-rexnet

v0.1-rs-weights

v0.1-rsb-weights

v0.1-selecsls

v0.1-tpu-weights

v0.1-tresnet

v0.1-vitjx

v0.1-vt3p-weights

v0.1-weights

v0.1-weights-maxx

v0.1-weights-morevit

v0.1-weights-swinv2

v0.1.1

v0.2.1

v0.3.3

v0.3.4

v0.4.12

v0.4.5

v0.4.9

v0.5.4

v0.6.11

v0.6.12

v0.6.5

v0.6.7

v0.8.10dev0

v0.8.13dev0

v0.8.2dev0

v0.8.6dev0

Mono Color

• a85df34993 Update lambda_resnet26rpt weights to 78.9, add better halonet26t weights at 79.1 with tweak to attention dim Ross Wightman 2021-10-08 17:44:13 -0700
• 38804c721b Checkpoint clean fn useable stand alone Ross Wightman 2021-10-08 17:43:53 -0700
• b544ad4d3f regnetz model default cfg tweaks Ross Wightman 2021-10-06 21:14:59 -0700
• d80653cb99 Merge branch 'alexander-soare-freeze-functionality' Ross Wightman 2021-10-06 17:01:41 -0700
• e5da481073 Small post-merge tweak for freeze/unfreeze, add to __init__ for utils Ross Wightman 2021-10-06 17:00:27 -0700
• 5ca72dcc75 Merge branch 'freeze-functionality' of https://github.com/alexander-soare/pytorch-image-models into alexander-soare-freeze-functionality Ross Wightman 2021-10-06 16:51:03 -0700
• e2b8d44ff0 Halo, bottleneck attn, lambda layer additions and cleanup along w/ experimental model defs * align interfaces of halo, bottleneck attn and lambda layer * add qk_ratio to all of above, control q/k dim relative to output dim * add experimental haloregnetz, and trionet (lambda + halo + bottle) models Ross Wightman 2021-10-06 16:29:33 -0700
• e0b3a3fab3 Make test-pooling flag for validate.py opt in Ross Wightman 2021-10-06 16:12:05 -0700
• 431e60c83f Add acknowledgements for freeze_batch_norm inspiration #876 Alexander Soare 2021-10-06 14:28:49 +0100
• fbf59c04ee Change crop ratio on correct resnet50 variant. Ross Wightman 2021-10-04 22:31:08 -0700
• ae1ff5792f Clean a1/a2/3 rsb _0 checkpoints properly, fix v2 loading. Ross Wightman 2021-10-04 16:46:00 -0700
• d123042605
  Update README.md Ross Wightman 2021-10-03 21:38:47 -0700
• cd638d50a5
  Merge pull request #880 from rwightman/fixes_bce_regnet Ross Wightman 2021-10-03 19:37:01 -0700
• 93901e992f Version bump to 0.5.0 for pending release post RSB and ATTN updates #880 fixes_bce_regnet Ross Wightman 2021-10-03 17:34:57 -0700
• da0d39bedd Update default crop_pct for byoanet Ross Wightman 2021-10-03 17:33:16 -0700
• cc9bedf373 Add initial ResNet Strikes Back weights for ResNet50 and ResNetV2-50 models Ross Wightman 2021-10-03 17:32:02 -0700
• 64495505b7 Add updated lambda resnet26 and botnet26 checkpoints with fixes applied Ross Wightman 2021-10-03 17:31:39 -0700
• b2094f4ee8 support bits checkpoints in avg/load Ross Wightman 2021-10-03 17:31:22 -0700
• 007bc39323 Some halo and bottleneck attn code cleanup, add halonet50ts weights, use optimal crop ratios Ross Wightman 2021-10-02 15:51:42 -0700
• 6d2acec1bb Fix ordering of tests Alexander Soare 2021-10-02 16:10:11 +0100
• 65c3d78b96 Freeze unfreeze functionality finalized. Tests added Alexander Soare 2021-10-02 15:54:14 +0100
• 0cb8ea432c wip Alexander Soare 2021-09-21 12:46:42 +0100
• 7b223af6f1
  Merge 417b8a7bb7 into b5bf4dce98 #213 Santiago Castro 2021-10-01 21:46:58 -0700
• 037bd1a057
  Merge 008f25430b into b5bf4dce98 #853 Alexander Soare 2021-10-01 21:46:38 -0700
• 1fdc7af8fd Merge remote-tracking branch 'origin/fixes_bce_regnet' into bits_and_tpu Ross Wightman 2021-10-01 16:32:04 -0700
• d9abfa48df Make broadcast_buffers disable its own flag for now (needs more testing on interaction with dist_bn) Ross Wightman 2021-10-01 13:43:55 -0700
• b1c2e3eb92 Match rel_pos_indices attr rename in conv branch Ross Wightman 2021-09-30 23:19:05 -0700
• b49630a138 Add relative pos embed option to LambdaLayer, fix last transpose/reshape. Ross Wightman 2021-09-30 22:45:09 -0700
• d657e2cc0b Remove dead code line from efficientnet Ross Wightman 2021-09-30 21:54:42 -0700
• 0ca687f224 Make 'regnetz' model experiments closer to actual RegNetZ, bottleneck expansion, expand from in_chs, no shortcut on stride 2, tweak model sizes Ross Wightman 2021-09-30 21:49:38 -0700
• b5bf4dce98
  Merge pull request #898 from leondgarse/master v0.1-rsb-weights Ross Wightman 2021-09-30 13:32:15 -0700
• 51eaf9360d
  Remove a duplicate layer creation in byobnet.py #898 leondgarse 2021-09-30 18:30:48 +0800
• b81e79aae9 Fix bottleneck attn transpose typo, hopefully these train better now.. Ross Wightman 2021-09-28 16:38:41 -0700
• 80075b0b8a Add worker_seeding arg to allow selecting old vs updated data loader worker seed for (old) experiment repeatability Ross Wightman 2021-09-28 16:37:45 -0700
• e21d91a317 Merge branch 'ISSUE-1_Add_training_script_that_uses_our_data' of https://github.com/coreadai/pytorch-image-models into ISSUE-1_Add_training_script_that_uses_our_data #891 mansikataria 2021-09-28 20:26:35 +0530
• ede65f5fac
  Update README.md Mansi 2021-09-28 20:26:12 +0530
• d7afc24ed9
  Update README.md Mansi 2021-09-28 20:24:22 +0530
• f78ab556ec Initial commit for: Add training script that uses our data mansikataria 2021-09-28 19:25:57 +0530
• 9ea8242729 Initial commit for: Add training script that uses our data mansikataria 2021-09-28 19:23:34 +0530
• 52c481ea8e Merge remote-tracking branch 'origin/fixes_bce_regnet' into bits_and_tpu Ross Wightman 2021-09-27 16:32:55 -0700
• 6478bcd02c Fix regnetz_d conv layer name, use inception mean/std Ross Wightman 2021-09-26 14:54:17 -0700
• 3f9959cdd2
  Merge pull request #882 from ShoufaChen/master Ross Wightman 2021-09-25 21:37:44 -0700
• 908563d060
  fix `use_amp` #882 Shoufa Chen 2021-09-26 12:32:22 +0800
• 25d52ea71d Merge remote-tracking branch 'origin/fixes_bce_regnet' into bits_and_tpu Ross Wightman 2021-09-24 22:55:38 -0700
• 0387e6057e Update binary cross ent impl to use thresholding as an option (convert soft targets from mixup/cutmix to 0, 1) Ross Wightman 2021-09-23 15:45:39 -0700
• 5d6983c462 Batch validate a list of files if model is a text file with model per line Ross Wightman 2021-09-23 15:45:17 -0700
• f8a63a3b71 Add worker_init_fn to loader for numpy seed per worker Ross Wightman 2021-09-23 15:44:38 -0700
• 515121cca1 Use reshape instead of view in std_conv, causing issues in recent PyTorch in channels_last Ross Wightman 2021-09-23 15:43:48 -0700
• da06cc61d4 ResNetV2 seems to work best without zero_init residual Ross Wightman 2021-09-23 15:43:22 -0700
• 8e11da0ce3 Add experimental RegNetZ(ish) models for training / perf trials. Ross Wightman 2021-09-23 15:42:57 -0700
• 3d9c23af87
  Merge pull request #875 from alexander-soare/effnets-norm-layer Ross Wightman 2021-09-21 07:17:52 -0700
• 6bbc50beb4 make it possible to provide norm_layer via create_model #875 Alexander Soare 2021-09-21 10:19:04 +0100
• 9de8192dee
  remove args.aug_repeats #874 YangYang 2021-09-19 00:52:54 +0800
• 3529a41266
  setup num_aug_repeats YangYang 2021-09-19 00:49:27 +0800
• 83a60f55fc
  setup torch.distributed for slurm platform and set num_aug_splits YangYang 2021-09-19 00:45:23 +0800
• adcb74f87f 🎨 Import load_state_dict_from_url directly nateraw 2021-09-14 01:11:40 -0400
• e65a2cba3d 🎨 cleanup and add a couple comments nateraw 2021-09-14 01:07:04 -0400
• 2b6ade24b3 🎨 write model card to enable inference nateraw 2021-09-13 23:31:28 -0400
• a6e8598aaf
  Merge pull request #821 from rwightman/attn_update Ross Wightman 2021-09-13 17:49:34 -0700
• cf5ac2800c BotNet models were still off, remove weights for bad configs. Add good SE-HaloNet33-TS weights. #821 attn_update Ross Wightman 2021-09-13 17:18:59 -0700
• 24720abe3b Merge branch 'master' into attn_update Ross Wightman 2021-09-13 16:51:10 -0700
• 1c9284c640 Add BeiT 'finetuned' 1k weights and pretrained 22k weights, pretraining specific (masked) model excluded for now Ross Wightman 2021-09-13 16:38:23 -0700
• f8a215cfe6 A few more crossvit tweaks, fix training w/ no_weight_decay names, add crop option for scaling, adjust default crop_pct for large img size to 1.0 for better results Ross Wightman 2021-09-13 14:17:34 -0700
• 7ab2491ab7 Better handling of crossvit for tests / forward_features, fix torchscript regression in my changes Ross Wightman 2021-09-13 13:01:05 -0700
• 702982d8af Merge branch 'chunfuchen-feature/crossvit' Ross Wightman 2021-09-13 11:50:58 -0700
• f1808e0970 Post crossvit merge cleanup, change model names to reflect input size, cleanup img size vs scale handling, fix tests Ross Wightman 2021-09-13 11:49:54 -0700
• a897e0ebcc Merge branch 'feature/crossvit' of https://github.com/chunfuchen/pytorch-image-models into chunfuchen-feature/crossvit Ross Wightman 2021-09-10 17:38:37 -0700
• 4027412757 Add resnet33ts weights, update resnext26ts baseline weights Ross Wightman 2021-09-09 14:46:41 -0700
• 9fe5798bee fix bug for reset classifier and fix for validating the dimension #841 Richard Chen 2021-09-08 21:58:17 -0400
• 3718c5a5bd fix loading pretrained model Richard Chen 2021-09-08 11:53:05 -0400
• bb50b69a57 fix for torch script Richard Chen 2021-09-08 11:20:59 -0400
• abf9d51bc3 🚧 wip nateraw 2021-09-07 18:39:26 -0600
• 93bfed11c8
  Rename pr to pr.py #858 han 2021-09-07 20:25:50 +0900
• fb3e3dbd5e
  Create pr han 2021-09-07 20:24:45 +0900
• 008f25430b add deterministic flag + functionality #853 Alexander Soare 2021-09-06 19:06:26 +0100
• 3581affb77 Update train.py with some flags related to scheduler tweaks, fix best checkpoint bug. Ross Wightman 2021-09-05 16:05:31 -0700
• c2f02b08b8 Merge remote-tracking branch 'origin/attn_update' into bits_and_tpu Ross Wightman 2021-09-05 16:02:50 -0700
• 5bd04714e4 Cleanup weight init for byob/byoanet and related Ross Wightman 2021-09-05 15:34:05 -0700
• 8642401e88 Swap botnet 26/50 weights/models after realizing a mistake in arch def, now figuring out why they were so low... Ross Wightman 2021-09-05 15:17:19 -0700
• 5f12de4875 Add initial AttentionPool2d that's being trialed. Fix comment and still trying to improve reliability of sgd test. Ross Wightman 2021-09-05 12:29:36 -0700
• 76881d207b Add baseline resnet26t @ 256x256 weights. Add 33ts variant of halonet with at least one halo in stage 2,3,4 Ross Wightman 2021-09-04 14:52:54 -0700
• 54e90e82a5 Another attempt at sgd momentum test passing... Ross Wightman 2021-08-27 10:39:31 -0700
• 484e61648d Adding the attn series weights, tweaking model names, comments... Ross Wightman 2021-09-03 18:09:42 -0700
• 0639d9a591 Fix updated validation_batch_size fallback Ross Wightman 2021-09-02 14:44:53 -0700
• 5db057dca0 Fix misnamed arg, tweak other train script args for better defaults. Ross Wightman 2021-09-02 14:15:49 -0700
• fb94350896 Update training script and loader factory to allow use of scheduler updates, repeat augment, and bce loss Ross Wightman 2021-09-01 17:46:40 -0700
• f262137ff2 Add RepeatAugSampler as per DeiT RASampler impl, showing promise for current (distributed) training experiments. Ross Wightman 2021-09-01 17:40:53 -0700
• ba9c1108a1 Add a BCE loss impl that converts dense targets to sparse /w smoothing as an alternate to CE w/ smoothing. For training experiments. Ross Wightman 2021-09-01 17:39:28 -0700
• 29a37e23ee LR scheduler update: * add polynomial decay 'poly' * cleanup cycle specific args for cosine, poly, and tanh sched, t_mul -> cycle_mul, decay -> cycle_decay, default cycle_limit to 1 in each opt * add k-decay for cosine and poly sched as per https://arxiv.org/abs/2004.05909 * change default tanh ub/lb to push inflection to later epochs Ross Wightman 2021-09-01 17:33:11 -0700
• 28d2841acf 💄 apply isort nateraw 2021-09-01 18:15:08 -0600
• 492c0a4e20 Update HaloAttn comment Ross Wightman 2021-09-01 17:14:31 -0700
• e72c989973 ✨ add ability to push to hf hub nateraw 2021-09-01 18:14:28 -0600
• 7ab9d4555c add crossvit Richard Chen 2021-09-01 17:13:12 -0400
• 3b9032ea48 Use Tensor.unfold().unfold() for HaloAttn, fast like as_strided but more clarity Ross Wightman 2021-08-27 12:45:53 -0700
• fc894c375c Another attempt at sgd momentum test passing... Ross Wightman 2021-08-27 10:39:31 -0700
• 78933122c9 Fix silly typo v0.1-attn-weights Ross Wightman 2021-08-27 09:22:20 -0700
• 2568ffc5ef Merge branch 'master' into attn_update Ross Wightman 2021-08-27 09:21:22 -0700
• 708d87a813 Fix ViT SAM weight compat as weights at URL changed to not use repr layer. Fix #825. Tweak optim test. Ross Wightman 2021-08-27 09:20:13 -0700
• 8449ba210c Improve performance of HaloAttn, change default dim calc. Some cleanup / fixes for byoanet. Rename resnet26ts to tfs to distinguish (extra fc). Ross Wightman 2021-08-26 21:56:44 -0700
• e6cb7a194b
  fixed typo #828 Thijs Vogels 2021-08-25 18:48:47 +0200