From b01ad75700c9cc2848edfb310ad24ceb9b06aacf Mon Sep 17 00:00:00 2001 From: Jirka Borovec Date: Sun, 2 Aug 2020 12:34:36 +0200 Subject: [PATCH] missing chlogs (#2672) MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit * missing * miss * Apply suggestions from code review Co-authored-by: Adrian Wälchli * miss * note * notes * update CI testing with pip upgrade (#2380) * try pt1.5 * cpu * upgrade * tpu * user * [blocked by #2380] freeze GPU PT 1.4 (#2780) * freeze * user Co-authored-by: Adrian Wälchli --- CHANGELOG.md | 37 ++++++++++++++++++++++++++ README.md | 10 ++++--- pytorch_lightning/trainer/lr_finder.py | 24 ++++++++--------- 3 files changed, 55 insertions(+), 16 deletions(-) diff --git a/CHANGELOG.md b/CHANGELOG.md index ac509ae5b5a88..04eef634a678f 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -10,29 +10,54 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/). ### Added - Added SSIM metrics ([#2671](https://github.com/PyTorchLightning/pytorch-lightning/pull/2671)) + - Added BLEU metrics ([#2535](https://github.com/PyTorchLightning/pytorch-lightning/pull/2535)) - Added support to export a model to ONNX format ([#2596](https://github.com/PyTorchLightning/pytorch-lightning/pull/2596)) - Added support for `Trainer(num_sanity_val_steps=-1)` to check all validation data before training ([#2246](https://github.com/PyTorchLightning/pytorch-lightning/pull/2246)) +- Added struct. output: + * tests for val loop flow ([#2605](https://github.com/PyTorchLightning/pytorch-lightning/pull/2605)) + * `EvalResult` support for train and val. loop ([#2615](https://github.com/PyTorchLightning/pytorch-lightning/pull/2615), [#2651](https://github.com/PyTorchLightning/pytorch-lightning/pull/2651)) + +- Added class `LightningDataModule` ([#2668](https://github.com/PyTorchLightning/pytorch-lightning/pull/2668)) + - Added support for PyTorch 1.6 ([#2745](https://github.com/PyTorchLightning/pytorch-lightning/pull/2745)) +- Added call DataModule hooks implicitly in trainer ([#2755](https://github.com/PyTorchLightning/pytorch-lightning/pull/2755)) + ### Changed - Truncated long version numbers in progress bar ([#2594](https://github.com/PyTorchLightning/pytorch-lightning/pull/2594)) +- Enable val/test loop disabling ([#2692](https://github.com/PyTorchLightning/pytorch-lightning/pull/2692)) + +- Refactor into `accelerator` module: + * GPU training ([#2704](https://github.com/PyTorchLightning/pytorch-lightning/pull/2704)) + * TPU training ([#2708](https://github.com/PyTorchLightning/pytorch-lightning/pull/2708)) + ### Deprecated - Deprecated Trainer attribute `ckpt_path`, which will now be set by `weights_save_path` ([#2681](https://github.com/PyTorchLightning/pytorch-lightning/pull/2681)) ### Removed +- Removed deprecated: ([#2760](https://github.com/PyTorchLightning/pytorch-lightning/pull/2760)) + * core decorator `data_loader` + * Module hook `on_sanity_check_start` and loading `load_from_metrics` + * package `pytorch_lightning.logging` + * Trainer arguments: `show_progress_bar`, `num_tpu_cores`, `use_amp`, `print_nan_grads` + * LR Finder argument `num_accumulation_steps` ### Fixed - Fixed setup call while testing ([#2624](https://github.com/PyTorchLightning/pytorch-lightning/pull/2624)) +- Fixed local rank zero casting ([#2640](https://github.com/PyTorchLightning/pytorch-lightning/pull/2640)) + +- Fixed single scalar return from training ([#2587](https://github.com/PyTorchLightning/pytorch-lightning/pull/2587)) + - Fixed Horovod backend to scale LR schedlers with the optimizer ([#2626](https://github.com/PyTorchLightning/pytorch-lightning/pull/2626)) - Fixed `dtype` and `device` properties not getting updated in submodules ([#2657](https://github.com/PyTorchLightning/pytorch-lightning/pull/2657)) @@ -51,6 +76,18 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/). - Fixed shuffle argument for distributed sampler ([#2789](https://github.com/PyTorchLightning/pytorch-lightning/pull/2789)) +- Fixed logging interval ([#2694](https://github.com/PyTorchLightning/pytorch-lightning/pull/2694)) + +- Fixed loss value in the progress bar is wrong when `accumulate_grad_batches > 1` ([#2738](https://github.com/PyTorchLightning/pytorch-lightning/pull/2738)) + +- Fixed correct CWD for ddp sub-processes when using Hydra ([#2719](https://github.com/PyTorchLightning/pytorch-lightning/pull/2719)) + +- Fixed selecting GPUs using `CUDA_VISIBLE_DEVICES` ([#2739](https://github.com/PyTorchLightning/pytorch-lightning/pull/2739)) + +- Fixed false `num_classes` warning in metrics ([#2781](https://github.com/PyTorchLightning/pytorch-lightning/pull/2781)) + +- Fixed shell injection vulnerability in subprocess call ([#2786](https://github.com/PyTorchLightning/pytorch-lightning/pull/2786)) + ## [0.8.5] - 2020-07-09 ### Added diff --git a/README.md b/README.md index 2b2132799965d..08a3a519bac1a 100644 --- a/README.md +++ b/README.md @@ -38,15 +38,19 @@ ## Continuous Integration
-| System / PyTorch ver. | 1.3 (min. req.) [w/o py3.8] | 1.4 | 1.5 | 1.6 (latest) | +| System / PyTorch ver. | 1.3 (min. req.)* | 1.4 | 1.5 | 1.6 (latest) | | :---: | :---: | :---: | :---: | :---: | | Conda py3.7 [linux] | ![PyTorch & Conda](https://github.com/PyTorchLightning/pytorch-lightning/workflows/PyTorch%20&%20Conda/badge.svg) | ![PyTorch & Conda](https://github.com/PyTorchLightning/pytorch-lightning/workflows/PyTorch%20&%20Conda/badge.svg) | ![PyTorch & Conda](https://github.com/PyTorchLightning/pytorch-lightning/workflows/PyTorch%20&%20Conda/badge.svg) | ![PyTorch & Conda](https://github.com/PyTorchLightning/pytorch-lightning/workflows/PyTorch%20&%20Conda/badge.svg) | -| Linux py3.7 [GPU] | - | - | - | [![Build Status](http://35.192.60.23/api/badges/PyTorchLightning/pytorch-lightning/status.svg)](http://35.192.60.23/PyTorchLightning/pytorch-lightning) | -| Linux py3.7 [TPU] | - | - | - | ![TPU tests](https://github.com/PyTorchLightning/pytorch-lightning/workflows/TPU%20tests/badge.svg) | +| Linux py3.7 [GPUs**] | - | - | - | [![Build Status](http://35.192.60.23/api/badges/PyTorchLightning/pytorch-lightning/status.svg)](http://35.192.60.23/PyTorchLightning/pytorch-lightning) | +| Linux py3.7 [TPUs***] | - | - | - | ![TPU tests](https://github.com/PyTorchLightning/pytorch-lightning/workflows/TPU%20tests/badge.svg) | | Linux py3.6 / py3.7 / py3.8 | [![CI testing](https://github.com/PyTorchLightning/pytorch-lightning/workflows/CI%20testing/badge.svg?event=push)](https://github.com/PyTorchLightning/pytorch-lightning/actions?query=workflow%3A%22CI+testing%22) | - | - | [![CI testing](https://github.com/PyTorchLightning/pytorch-lightning/workflows/CI%20testing/badge.svg?event=push)](https://github.com/PyTorchLightning/pytorch-lightning/actions?query=workflow%3A%22CI+testing%22) | | OSX py3.6 / py3.7 | - | [![CI testing](https://github.com/PyTorchLightning/pytorch-lightning/workflows/CI%20testing/badge.svg?event=push)](https://github.com/PyTorchLightning/pytorch-lightning/actions?query=workflow%3A%22CI+testing%22) | - | [![CI testing](https://github.com/PyTorchLightning/pytorch-lightning/workflows/CI%20testing/badge.svg?event=push)](https://github.com/PyTorchLightning/pytorch-lightning/actions?query=workflow%3A%22CI+testing%22) | | Windows py3.6 / py3.7 / py3.8 | [![CI testing](https://github.com/PyTorchLightning/pytorch-lightning/workflows/CI%20testing/badge.svg?event=push)](https://github.com/PyTorchLightning/pytorch-lightning/actions?query=workflow%3A%22CI+testing%22) | - | - | [![CI testing](https://github.com/PyTorchLightning/pytorch-lightning/workflows/CI%20testing/badge.svg?event=push)](https://github.com/PyTorchLightning/pytorch-lightning/actions?query=workflow%3A%22CI+testing%22) +- _\* `torch>=1.4` is the minimal pytorch version for Python 3.8_ +- _\** tests run on two NVIDIA K80_ +- _\*** tests run on Google GKE TPUv2/3_ +
Simple installation from PyPI diff --git a/pytorch_lightning/trainer/lr_finder.py b/pytorch_lightning/trainer/lr_finder.py index db6863a60cf1c..3b2778d24071c 100755 --- a/pytorch_lightning/trainer/lr_finder.py +++ b/pytorch_lightning/trainer/lr_finder.py @@ -77,16 +77,17 @@ def _run_lr_finder_internally(self, model: LightningModule): ' either has field `lr` or `learning_rate` that can overridden') log.info(f'Learning rate set to {lr}') - def lr_find(self, - model: LightningModule, - train_dataloader: Optional[DataLoader] = None, - val_dataloaders: Optional[Union[DataLoader, List[DataLoader]]] = None, - min_lr: float = 1e-8, - max_lr: float = 1, - num_training: int = 100, - mode: str = 'exponential', - early_stop_threshold: float = 4.0, - num_accumulation_steps=None): + def lr_find( + self, + model: LightningModule, + train_dataloader: Optional[DataLoader] = None, + val_dataloaders: Optional[Union[DataLoader, List[DataLoader]]] = None, + min_lr: float = 1e-8, + max_lr: float = 1, + num_training: int = 100, + mode: str = 'exponential', + early_stop_threshold: float = 4.0, + ): r""" lr_find enables the user to do a range test of good initial learning rates, to reduce the amount of guesswork in picking a good starting learning rate. @@ -113,9 +114,6 @@ def lr_find(self, loss at any point is larger than early_stop_threshold*best_loss then the search is stopped. To disable, set to None. - num_accumulation_steps: deprepecated, number of batches to calculate loss over. - Set trainer argument ``accumulate_grad_batches`` instead. - Example:: # Setup model and trainer