160 Commits (maxxvit_hub)

Author	SHA1	Message	Date
Antoine Broyelle	78fa0772cc	Leverage python hierachical logger ... with this update one can tune the kind of logs generated by timm but training and inference traces are unchanged	4 years ago
Ross Wightman	6441e9cc1b	Fix memory_efficient mode for DenseNets. Add AntiAliasing (Blur) support for DenseNets and create one test model. Add lr cycle/mul params to train args.	4 years ago
AFLALO, Jonathan Isaac	a7f570c9b7	added MultiEpochsDataLoader	4 years ago
Ross Wightman	13cf68850b	Remove poorly named metrics from torch imagenet example origins. Use top1/top5 in csv output for consistency with existing validation results files, acc elsewhere. Fixes #111	4 years ago
Ross Wightman	27b3680d49	Revamp LR noise, move logic to scheduler base. Fixup PlateauLRScheduler and add it as an option.	4 years ago
Ross Wightman	514b0938c4	Experimenting with per-epoch learning rate noise w/ step scheduler	4 years ago
Ross Wightman	43225d110c	Unify drop connect vs drop path under 'drop path' name, switch all EfficientNet/MobilenetV3 refs to 'drop_path'. Update factory to handle new drop args.	4 years ago
Ross Wightman	b3cb5f3275	Working on CutMix impl as per #8, integrating with Mixup, currently experimenting...	4 years ago
Andrew Lavin	b72013def8	Added commandline argument validation-batch-size-multiplier with default set to 1.	4 years ago
Ross Wightman	5b7cc16ac9	Add warning about using sync-bn with zero initialized BN layers. Fixes #54	4 years ago
Ross Wightman	d9a6a9d0af	Merge pull request #74 from rwightman/augmix-jsd ... AugMix, JSD loss, SplitBatchNorm (Auxiliary BN), and more	4 years ago
Ross Wightman	3eb4a96eda	Update AugMix, JSD, etc comments and references	4 years ago
Ross Wightman	7547119891	Add SplitBatchNorm. AugMix, Rand/AutoAugment, Split (Aux) BatchNorm, Jensen-Shannon Divergence, RandomErasing all working together	4 years ago
Ross Wightman	40fea63ebe	Add checkpoint averaging script. Add headers, shebangs, exec perms to all scripts	4 years ago
Ross Wightman	4666cc9aed	Add --pin-mem arg to enable dataloader pin_memory (showing more benefit in some scenarios now), also add --torchscript arg to validate.py for testing models with jit.script	4 years ago
Ross Wightman	232ab7fb12	Working on an implementation of AugMix with JensenShannonDivergence loss that's compatible with my AutoAugment and RandAugment impl	5 years ago
Ross Wightman	5719b493ad	Missed update dist-bn logic for EMA model	5 years ago
Ross Wightman	a435ea1327	Change reduce_bn to distribute_bn, add ability to choose between broadcast and reduce (mean). Add crop_pct arg to allow selecting validation crop while training.	5 years ago
Ross Wightman	3bff2b21dc	Add support for keeping running bn stats the same across distributed training nodes before eval/save	5 years ago
Ross Wightman	1f39d15f15	Allow float decay epochs arg for training, works out with step lr math	5 years ago
Ross Wightman	7b83e67f77	Pass drop connect arg through to EfficientNet models	5 years ago
Ross Wightman	4748c6dff2	Fix non-prefetch variant of Mixup. Fixes #50	5 years ago
Ross Wightman	187ecbafbe	Add support for loading args from yaml file (and saving them with each experiment)	5 years ago
Ross Wightman	b750b76f67	More AutoAugment work. Ready to roll...	5 years ago
Ross Wightman	3d9c8a6489	Add support for new AMP checkpointing support w/ amp.state_dict	5 years ago
Ross Wightman	fac58f609a	Add RAdam, NovoGrad, Lookahead, and AdamW optimizers, a few ResNet tweaks and scheduler factory tweak. ... * Add some of the trendy new optimizers. Decent results but not clearly better than the standards. * Can create a None scheduler for constant LR * ResNet defaults to zero_init of last BN in residual * add resnet50d config	5 years ago
Ross Wightman	66634d2200	Add support to split random erasing blocks into randomly selected number with --recount arg. Fix random selection of aspect ratios.	5 years ago
Ross Wightman	e7c8a37334	Make min-lr and cooldown-epochs cmdline args, change dash in color_jitter arg for consistency	5 years ago
Ross Wightman	c6b32cbe73	A number of tweaks to arguments, epoch handling, config ... * reorganize train args * allow resolve_data_config to be used with dict args, not just arparse * stop incrementing epoch before save, more consistent naming vs csv, etc * update resume and start epoch handling to match above * stop auto-incrementing epoch in scheduler	5 years ago
Ross Wightman	b20bb58284	Distributed tweaks ... * Support PyTorch native DDP as fallback if APEX not present * Support SyncBN for both APEX and Torch native (if torch >= 1.1) * EMA model does not appear to need DDP wrapper, no gradients, updated from wrapped model	5 years ago
Ross Wightman	6fc886acaf	Remove all prints, change most to logging calls, tweak alignment of batch logs, improve setup.py	5 years ago
Ross Wightman	aa4354f466	Big re-org, working towards making pip/module as 'timm'	5 years ago
Ross Wightman	7dab6d1ec7	Default to img_size in model default_cfg, defer output folder creation until later in the init sequence	5 years ago
Ross Wightman	9bcd65181b	Add exponential moving average for model weights + few other additions and cleanup ... * ModelEma class added to track an EMA set of weights for the model being trained * EMA handling added to train, validation and clean_checkpoint scripts * Add multi checkpoint or multi-model validation support to validate.py * Add syncbn option (APEX) to train script for experimentation * Cleanup interface of CheckpointSaver while adding ema functionality	5 years ago
Ross Wightman	e6c14427c0	More appropriate/correct loss name	5 years ago
Zhun Zhong	127487369f	Fix bug for prefetcher ... Set input and target to Cuda when without using prefetcher.	5 years ago
Ross Wightman	4d2056722a	Mixup and prefetcher improvements ... * Do mixup in custom collate fn if prefetcher enabled, reduces performance impact * Move mixup code to own file * Add arg to disable prefetcher * Fix no cuda transfer when prefetcher off * Random erasing when prefetcher off wasn't changed to match new args, fixed * Default random erasing to off (prob = 0.) for train	5 years ago
Ross Wightman	780c0a96a4	Change args for RandomErasing so only one required for pixel/color mode	5 years ago
Ross Wightman	76539d905e	Some transform/data/loader refactoring, hopefully didn't break things ... * factor out data related constants to own file * move data related config helpers to own file * add a variant of RandomResizeCrop that randomizes interpolation method * remove old Numpy version of RandomErasing * cleanup torch version of RandomErasing and use it in either GPU loader batch mode or single image cpu Transform	5 years ago
Ross Wightman	fee607edf6	Mixup implemention in progress ... * initial impl w/ label smoothing converging, but needs more testing	5 years ago
Ross Wightman	8fbd62a169	Exclude batchnorm and bias params from weight_decay by default	5 years ago
Ross Wightman	bc264269c9	Morph mnasnet impl into a generic mobilenet that covers Mnasnet, MobileNetV1/V2, ChamNet, FBNet, and related ... * add an alternate RMSprop opt that applies eps like TF * add bn params for passing through alternates and changing defaults to TF style	5 years ago
Ross Wightman	e9c7961efc	Fix pooling in mnasnet, more sensible default for AMP opt level	5 years ago
Ross Wightman	0562b91c38	Add per model crop pct, interpolation defaults, tie it all together ... * create one resolve fn to pull together model defaults + cmd line args * update attribution comments in some models * test update train/validation/inference scripts	5 years ago
Ross Wightman	c328b155e9	Random erasing crash fix and args pass through	5 years ago
Ross Wightman	9c3859fb9c	Uniform pretrained model handling. ... * All models have 'default_cfgs' dict * load/resume/pretrained helpers factored out * pretrained load operates on state_dict based on default_cfg * test all models in validate * schedule, optim factor factored out * test time pool wrapper applied based on default_cfg	5 years ago
Ross Wightman	f1cd1a5ce3	Cleanup CheckpointSaver, add support for increasing or decreasing metric, switch to prec1 metric in train loop	5 years ago
Ross Wightman	5180f94c7e	Distributed (multi-process) train, multi-gpu single process train, and NVIDIA AMP support	5 years ago
Ross Wightman	45cde6f0c7	Improve creation of data pipeline with prefetch enabled vs disabled, fixup inception_res_v2 and dpn models	5 years ago
Ross Wightman	2295cf56c2	Add some Nvidia performance enhancements (prefetch loader, fast collate), and refactor some of training and model fact/transforms	5 years ago
Ross Wightman	9d927a389a	Add adabound, random erasing	5 years ago
Ross Wightman	1577c52976	Resnext added, changes to bring it and seresnet in line with rest of models	5 years ago
Ross Wightman	31055466fc	Fixup validate/inference script args, fix senet init for better test accuracy	5 years ago
Ross Wightman	b1a5a71151	Update schedulers	5 years ago
Ross Wightman	b5255960d9	Tweaking tanh scheduler, senet weight init (for BN), transform defaults	5 years ago
Ross Wightman	a336e5bff3	Minor updates	5 years ago
Ross Wightman	cf0c280e1b	Cleanup tranforms, add custom schedulers, tweak senet34 model	5 years ago
Ross Wightman	c57717d325	Fix tta train bug, improve logging	5 years ago
Ross Wightman	72b4d162a2	Increase training performance	5 years ago
Ross Wightman	5855b07ae0	Initial commit, puting some ol pieces together	5 years ago