Migrations
/
pytorch-image-models
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Commit Graph

  • ce62f96d4d ByoaNet with bottleneck transformer, lambda resnet, and halo net experiments Ross Wightman 2021-04-12 09:38:02 -0700
  • cd3dc4979f Fix adabelief imports, remove prints, preserve memory format is the default arg for zeros_like Ross Wightman 2021-04-12 08:25:31 -0700
  • 59ca94572b Run all tests Aman Arora 2021-04-12 04:44:08 -0400
  • 607672841b run all tests Aman Arora 2021-04-12 04:17:37 -0400
  • 84fd045e4d only run test_forward_features Aman Arora 2021-04-12 04:09:01 -0400
  • a17ce022e6 back to defaults Aman Arora 2021-04-12 02:56:47 -0400
  • f89bf2c752 Add to exclude ftrs Aman Arora 2021-04-12 02:55:12 -0400
  • 21812d33aa Add prelim efficientnet_v2s weights from 224x224 train, eval 83.3 @ 288. Add eca_nfnet_l1 weights, train at 256, eval 84 @ 320. Ross Wightman 2021-04-11 23:24:58 -0700
  • 094f04d992 Run test_model_forward_features for all Aman Arora 2021-04-12 02:20:21 -0400
  • bf63f0cc7b Add to exclude filter Aman Arora 2021-04-12 02:00:48 -0400
  • 973b8801c4 Try running only resnetv2_101x1_bitm on Linux runner Aman Arora 2021-04-12 00:04:04 -0400
  • 530f8ddb84 Add first conv Aman Arora 2021-04-11 23:53:56 -0400
  • 823162c4af increase time Aman Arora 2021-04-11 22:25:35 -0400
  • 3df71a1625 Update time Aman Arora 2021-04-11 21:37:07 -0400
  • ca2877f176 EOF newline Aman Arora 2021-04-11 20:25:26 -0400
  • 3f12ad3ebb remove whitespace diff Aman Arora 2021-04-11 20:24:35 -0400
  • 206407a0e1 remove whitespace diff Aman Arora 2021-04-11 20:23:52 -0400
  • b117e16128 Only include resnet-rs changes Aman Arora 2021-04-11 20:22:16 -0400
  • 4b2952825e
         Update test_models.py #548 Kai Han 2021-04-11 23:14:31 +0800
  • 0be1fa4793
         Argument description fixed #555 Michael Monashev 2021-04-11 18:08:43 +0300
  • b85650093f Update ghostnet iamhankai 2021-04-11 21:55:27 +0800
  • 7236ba08e2 Update ghostnet iamhankai 2021-04-11 20:32:19 +0800
  • 844993267b Add ResNet-RS models Aman Arora 2021-04-10 05:28:45 -0400
  • 5772c55c57 Make wandb optional #550 Aman Arora 2021-04-10 01:34:20 -0400
  • f54897cc0b make wandb not required but rather optional as huggingface_hub Aman Arora 2021-04-10 01:27:23 -0400
  • f13f7508a9 Keep changes to minimal and use args.experiment as wandb project name if it exists Aman Arora 2021-04-10 00:50:52 -0400
  • f8bb13f640 Default project name to None Aman Arora 2021-04-10 00:44:05 -0400
  • 8db8ff346f add wandb to requirements.txt Aman Arora 2021-04-08 03:52:14 -0400
  • 3f028ebc0f import wandb in summary.py Aman Arora 2021-04-08 03:48:51 -0400
  • a9e5d9e5ad log loss as before Aman Arora 2021-04-08 03:41:40 -0400
  • 624c9b6949 log to wandb only if using using wandb Aman Arora 2021-04-08 03:40:22 -0400
  • 00c8e0b8bd Make use of wandb configurable Aman Arora 2021-04-08 03:35:59 -0400
  • 8e6fb861e4 Add wandb support Aman Arora 2021-04-08 03:22:29 -0400
  • b0bd2884e7 Fix GhostNet bug iamhankai 2021-04-08 08:24:18 +0800
  • 9295f0e85e Add GhostNet iamhankai 2021-04-07 22:51:15 +0800
  • 779107b693
         Merge pull request #542 from juntang-zhuang/adabelief Ross Wightman 2021-04-05 16:34:54 -0700
  • 74366f733c
         Delete distributed_train_adabelief.sh #542 Juntang Zhuang 2021-04-05 15:11:45 -0400
  • 1d848f409a
         Delete args.yaml Juntang Zhuang 2021-04-05 15:11:35 -0400
  • addfc7c1ac adabelief juntang 2021-04-04 23:48:15 -0400
  • fb896c0b26 Update some comments re preliminary EfficientNet-V2 assumptions Ross Wightman 2021-04-03 12:00:25 -0700
  • 2b49ab7a36 Fix ResNetV2 pretrained classifier issue. Fixes #540 Ross Wightman 2021-04-03 11:18:12 -0700
  • bcb4879390
         Update num_classes in reset_classifier in resnetv2 #537 Mohamed Al Salti 2021-04-03 04:25:10 +0300
  • de9dff933a EfficientNet-V2S preliminary model def (for experimentation) Ross Wightman 2021-04-02 09:35:03 -0700
  • d5ed58d623
         Merge pull request #533 from rwightman/pit_and_vit_update Ross Wightman 2021-04-01 22:56:59 -0700
  • 37c71a5609 Some further create_optimizer_v2 tweaks, remove some redudnant code, add back safe model str. Benchmark step times per batch. #533 Ross Wightman 2021-04-01 22:34:55 -0700
  • 2bb65bd875 Wrong default_cfg pool_size for L1 Ross Wightman 2021-04-01 20:00:41 -0700
  • bf2ca6bdf4 Merge jax and original weight init Ross Wightman 2021-04-01 18:11:51 -0700
  • acbd698c83 Update README.md with updates. Small tweak to head_dist handling. Ross Wightman 2021-04-01 17:49:05 -0700
  • 9071568f0e Add weights for SE NFNet-L0 model, rename nfnet_l0b -> nfnet_l0. 82.75 top-1 @ 288. Add nfnet_l1 model def for training. Ross Wightman 2021-04-01 17:22:27 -0700
  • c468c47a9c Add regnety_160 weights from DeiT teacher model, update that and my regnety_032 weights to use higher test size. Ross Wightman 2021-04-01 16:41:04 -0700
  • 288682796f Update benchmark script to add precision arg. Fix some downstream (DeiT) compat issues with latest changes. Bump version to 0.4.7 Ross Wightman 2021-04-01 16:40:12 -0700
  • ea9c9550b2 Fully move ViT hybrids to their own file, including embedding module. Remove some extra DeiT models that were for benchmarking only. Ross Wightman 2021-04-01 14:17:38 -0700
  • a5310a3451 Merge remote-tracking branch 'origin/benchmark-fixes-vit_hybrids' into pit_and_vit_update Ross Wightman 2021-04-01 12:15:34 -0700
  • 7953e5d11a Fix pos_embed scaling for ViT and num_classes != 1000 for pretrained distilled deit and pit models. Fix #426 and fix #433 Ross Wightman 2021-03-31 23:11:28 -0700
  • a760a4c3f4 Some ViT cleanup, merge distilled model with main, fixup torchscript support for distilled models Ross Wightman 2021-03-31 18:21:02 -0700
  • 0dfc5a66bb Add PiT model from https://github.com/naver-ai/pit Ross Wightman 2021-03-31 18:20:14 -0700
  • 333af604f8 Get Torchvision and Pytorch Hub models with scripts #429 Csaba Kertesz 2021-02-10 08:21:41 +0200
  • 1ad1645a50 Merge branch 'contrastive-master' v0.1-pit-weights Ross Wightman 2021-03-29 11:33:32 -0700
  • 51febd869b Small tweak to tests for tnt model, reorder model imports. Ross Wightman 2021-03-29 11:33:08 -0700
  • b27a4e0d88 Merge branch 'master' of https://github.com/contrastive/pytorch-image-models into contrastive-master Ross Wightman 2021-03-29 10:37:05 -0700
  • 2319cbbff2
         Merge pull request #525 from amaarora/spp Ross Wightman 2021-03-29 08:36:04 -0700
  • 6b18061773 Add GIST to docstring for quick access #525 Aman Arora 2021-03-29 15:33:31 +1100
  • 809271b0f3
         Update test_models.py #528 contrastive 2021-03-29 09:39:50 +0800
  • de86314655 Update TNT contrastive 2021-03-29 08:23:34 +0800
  • 92b1db9a79 update docstrings and add check on and Aman Arora 2021-03-29 10:04:51 +1100
  • b85be24054 update to work with fnmatch Aman Arora 2021-03-29 09:36:31 +1100
  • cfc15283a4 Update TNT url contrastive 2021-03-28 23:19:15 +0800
  • 4a09bc851e Add TNT model contrastive 2021-03-28 19:53:42 +0800
  • 20626e8387 Add to extract stats for SPP Aman Arora 2021-03-27 05:40:04 +1100
  • 00dfd2cb6d
         Merge bc25e17bed into a2727c1bf7 #240 Kim, Taehoon 2021-03-22 07:15:20 +0000
  • bc25e17bed post quantization using graph_fx(beta) #240 taehoon.kim 2021-03-22 16:15:12 +0900
  • 2b2e9c77fd post quantization using graph_fx(beta) taehoon.kim 2021-03-22 16:05:27 +0900
  • 15d1dbc896 reset to latest original repo taehoon.kim 2021-03-22 13:42:58 +0900
  • a2727c1bf7
         Merge pull request #510 from rwightman/dependabot/pip/jinja2-2.11.3 Ross Wightman 2021-03-20 16:45:53 -0700
  • e2e3290fbf Add '--experiment' to train args for fixed exp name if desired, 'train' not added to output folder if specified. #450 Ross Wightman 2021-03-20 12:02:17 -0700
  • cf5fec5047 Cleanup experimental vit weight init a bit Ross Wightman 2021-03-20 09:44:24 -0700
  • b36099ace3
         Bump jinja2 from 2.11.2 to 2.11.3 #510 dependabot[bot] 2021-03-20 05:50:25 +0000
  • f42f1df26c Improve evenness of per-worker split for validation set with TFDS Ross Wightman 2021-03-18 23:16:14 -0700
  • cbcb76d72c Should have included Conv2d layers in original weight init. Lets see what the impact is... Ross Wightman 2021-03-18 23:15:48 -0700
  • 4de57ccf01 Add weight init scheme that's closer to JAX impl Ross Wightman 2021-03-18 15:35:22 -0700
  • 14ac4abf74 Change huggingface hub revision delimiter to '@', add hf_hub reference for eca_nfnet_l0 model as an example. Ross Wightman 2021-03-18 11:38:06 -0700
  • 740f32c96a Add ECA-NFNet-L0 weights and update model name. Update README and bump version to 0.4.6 Ross Wightman 2021-03-17 13:55:32 -0700
  • 5e2e4e7fb6
         Merge pull request #502 from khatami-mehrdad/efficientnet_dg Ross Wightman 2021-03-17 13:35:22 -0700
  • 3eac7dc5a3
         Merge pull request #501 from rwightman/hf_hub_revisit Ross Wightman 2021-03-17 13:26:50 -0700
  • 45c048ba13 A few minor fixes and bit more cleanup on the huggingface hub integration. #501 hf_hub_revisit Ross Wightman 2021-03-17 13:18:52 -0700
  • 97f49d163c adding efficientnet_el, efficientnet_es_prunedefficientnet_es_pruned and efficientnet_el_pruned pretrained models to timm/models #502 Mehrdad 2021-03-17 10:51:51 -0700
  • ead80d33c5 Fix typo, naming consistency Ross Wightman 2021-03-16 23:02:51 -0700
  • d584e7f617 Support for huggingface hub via create_model and default_cfgs. * improve consistency of model creation helper fns * add comments to some of the model helpers * support passing external default_cfgs so they can be sourced from hub Ross Wightman 2021-03-16 22:48:26 -0700
  • e5ba5dcbc9 Merge branch 'yoniaflalo-adding_Hardcore_NAS' Ross Wightman 2021-03-15 10:21:18 -0700
  • 1f799afcb3 Remove capital letters in model names Ross Wightman 2021-03-15 10:20:34 -0700
  • 52b5d0ad0a Merge branch 'adding_Hardcore_NAS' of https://github.com/yoniaflalo/pytorch-image-models into yoniaflalo-adding_Hardcore_NAS Ross Wightman 2021-03-15 09:37:25 -0700
  • 82c1fe958a added hardcorenas model #497 AFLALO, Jonathan Isaac 2021-03-14 22:06:33 +0200
  • acb81357c4 Deployed 06aa926 with MkDocs version: 1.1.2 Ross Wightman 2021-03-14 11:45:51 -0800
  • 06aa926fbf Tweak docs model summaries / pages names and layout Ross Wightman 2021-03-14 12:45:20 -0700
  • ef9970fd4d Deployed ac9a2c9 with MkDocs version: 1.1.2 Ross Wightman 2021-03-14 11:27:04 -0800
  • ac9a2c978d Update docs requirements Ross Wightman 2021-03-14 12:26:31 -0700
  • dda8fd80d2 Merge branch 'gcucurull-master' Ross Wightman 2021-03-14 12:25:08 -0700
  • 74fcd4c50e Update finetune detail, rerun model doc gen. Ross Wightman 2021-03-14 12:22:59 -0700
  • e6f5617dcc Fix a few typos and consistency issues in model docs Ross Wightman 2021-03-14 12:20:12 -0700
  • 5392c0a4b3 Rerun modelindex generate script Ross Wightman 2021-03-14 12:08:56 -0700