bump to v1.0.0rc2 (#1405)

* [Improve] add a version option in docs menu (#1162)
* [Enhance] update dev_scripts for link checking (#1164)
* [Refactoring] decompose the implementations of different metrics into several files (#1161)
* [Fix] Fix PPL bug (#1172)
* [Fix] Fix some known bugs. (#1200)
* [Fix] Benchmark related bugs (#1236)
* [Enhancement] Support rerun failed or canceled jobs in `train_benchmark.py` (#1259)
* [Fix]  Fix bugs in `sr test config`, `realbasicvsr config` and `pconv config` (#1167)
* [Fix] fix test of Vid4 datasets bug (#1293)
* [Feature] Support multi-metrics with different sample-model (#1171)
* [Fix] fix GenerateSegmentIndices ut (#1302)
* [Enhancement] Reduce the randomness in unit test of `stylegan3_utils.py` (#1306)
* [CI] Fix GitHub windows CI (#1320)
* [Fix] fix basicvsr++ mirror sequence bug (#1304)
* [Fix] fix sisr-test psnr config (#1319)
* [Fix] fix vsr models pytorch2onnx (#1300)
* [Bug] Ensure the output type of `GenerateFacialHeatmap` is `np.float32` (#1310)
* [Bug] Fix sampling behavior of `unpaired_dataset.py` and  urls in cyclegan's README  (#1308)
* [README] Fix TTSR's README (#1325)
* [CI] Update `paths-ignore` for GitHub CI (#1327)
* [Bug] Save gt images in PGGAN's `forward` (#1328)
* [Bug] Correct RDN number of channels (#1332)
* [Bug] Revise flip transformation in some conditional gan's setting (#1331)
* [Unit Test] Fix unit test of SNR (#1335)
* [Bug] Revise flavr config (#1336)
* [Fix] fix realesrgan ema (#1341)
* [Fix] Fix bugs find during benchmark running (#1348)
* [Fix] fix liif test config (#1353)
* [Enhancement] Complete save_best in configs (#1349)
* [Config] Revise discriminator's learning rate of TTSR to align with 0.x version (#1352)
* [Fix] fix edsr configs (#1367)
* [Enhancement] Add pixel value clip in visualizer (#1365)
* [Bug] Fix randomness in FixedCrop + add L1 loss in Pix2Pix (#1364)
* [Fix] fix realbasicvsr config (#1358)
* [Enhancement] Fix PESinGAN-inter-pad setting + add SinGAN Dataset + add SinGAN demo (#1363)
* [Fix] fix types of exceptions in demos (#1372)
* [Enhancement] Support deterministic training in benchmark (#1356)
* [Fix] Avoid cast int and float in GenDataPreprocessor (#1385)
* [Config] Update metric config in ggan (#1386)
* [Config] Revise batch size in wang-gp's config (#1384)
* [Fix]: add type and change default number of preprocess_div2k_dataset.py (#1380)
* [Feature] Support qualitative comparison tools (#1303)
* [Docs] Revise docs (change PackGenInputs and GenDataSample to mmediting ones) (#1382)
* [Config] Revise Pix2Pix edges2shoes config (#1391)
* [Bug] fix rdn and srcnn train configs (#1392)
* [Fix] Fix test/val pipeline of pegan configs (#1393)
* [Fix] Modify Readme of S3 (#1398)
* [Fix] Correct fid of ggan (#1397)
* [Feature] support instance_aware_colorization inference (#1370)

Co-authored-by: ruoning <w853133995@outlook.com>
Co-authored-by: Yifei Yang <2744335995@qq.com>
Co-authored-by: LeoXing1996 <xingzn1996@hotmail.com>
Co-authored-by: Z-Fran <49083766+Z-Fran@users.noreply.github.com>
Co-authored-by: Qunliang Xing <ryanxingql@gmail.com>
Co-authored-by: Yang Gao <Gary1546308416AL@gmail.com>
Co-authored-by: ruoning <44117949+ruoningYu@users.noreply.github.com>

.circleci/config.yml

@@ -26,6 +26,7 @@ workflows:
             tools/.* lint_only false
             configs/.* lint_only false
             .circleci/.* lint_only false
+            .dev_scripts/.* lint_only true
           base-revision: 1.x
           # this is the path of the configuration we should trigger once
           # path filtering and pipeline parameter value updates are

.circleci/test.yml

@@ -62,6 +62,7 @@ jobs:
             pip install git+https://github.com/open-mmlab/mmengine.git@main
             pip install -U openmim
             mim install 'mmcv >= 2.0.0rc1'
+            mim install 'mmdet >= 3.0.0rc2'
             pip install -r requirements/tests.txt
       - run:
           name: Build and install
@@ -98,13 +99,14 @@ jobs:
           name: Build Docker image
           command: |
             docker build .circleci/docker -t mmedit:gpu --build-arg PYTORCH=<< parameters.torch >> --build-arg CUDA=<< parameters.cuda >> --build-arg CUDNN=<< parameters.cudnn >>
-            docker run --gpus all -t -d -v /home/circleci/project:/mmedit -v /home/circleci/mmengine:/mmengine -w /mmedit --name mmedit mmedit:gpu
+            docker run --gpus all -t -d -v /home/circleci/project:/mmedit -v /home/circleci/mmengine:/mmengine -v /home/circleci/mmdetection:/mmdetection -w /mmedit --name mmedit mmedit:gpu
       - run:
           name: Install mmedit dependencies
           command: |
             docker exec mmedit pip install -e /mmengine
             docker exec mmedit pip install -U openmim
             docker exec mmedit mim install 'mmcv >= 2.0.0rc1'
+            docker exec mmedit mim install 'mmdet >= 3.0.0rc2'
             docker exec mmedit pip install -r requirements/tests.txt
       - run:
           name: Build and install

.dev_scripts/README.md

@@ -9,6 +9,7 @@
 - [4. Monitor your training](#4-monitor-your-training)
 - [5. Train with a list of models](#5-train-with-a-list-of-models)
 - [6. Train with skipping a list of models](#6-train-with-skipping-a-list-of-models)
+- [7. Automatically check links](#automatically-check-links)
 
 ## 1. Check UT
 
@@ -128,7 +129,7 @@ python .dev_scripts/train_benchmark.py mm_lol \
     --quotatype=auto
 ```
 
-# 4. Monitor your training
+## 4. Monitor your training
 
 After you submitting jobs following [3-Train-all-the-models](#3-train-all-the-models), you will find a `xxx.log` file.
 This log file list all the job name of job id you have submitted. With this log file, you can monitor your training by running `.dev_scripts/job_watcher.py`.
@@ -141,7 +142,7 @@ python .dev_scripts/job_watcher.py --work-dir work_dirs/benchmark_fp32/ --log 20
 
 Then, you will find `20220923-140317.csv`, which reports the status and recent log of each job.
 
-# 5. Train with a list of models
+## 5. Train with a list of models
 
 If you only need to run some of the models, you can list all the models' name in a file, and specify the models when using `train_benchmark.py`.
 
@@ -162,7 +163,7 @@ python .dev_scripts/train_benchmark.py mm_lol \
 
 Specifically, you need to enable `--rerun`, and specify the list of models to rerun by `--rerun-list`
 
-# 6. Train with skipping a list of models
+## 6. Train with skipping a list of models
 
 If you want to train all the models while skipping some models, you can also list all the models' name in a file, and specify the models when running `train_benchmark.py`.
 
@@ -182,3 +183,53 @@ python .dev_scripts/train_benchmark.py mm_lol \
 ```
 
 Specifically, you need to enable `--skip`, and specify the list of models to skip by `--skip-list`
+
+## 7. Train failed or canceled jobs
+
+If you want to rerun failed or canceled jobs in the last run, you can combine `--rerun` flag with `--rerun-failure` and `--rerun-cancel` flags.
+
+For example, the log file of the last run is `train-20221009-211904.log`, and now you want to rerun the failed jobs. You can use the following command:
+
+```bash
+python .dev_scripts/train_benchmark.py mm_lol \
+    --job-name RERUN \
+    --rerun train-20221009-211904.log \
+    --rerun-fail \
+    --run
+```
+
+We can combine `--rerun-fail` and `--rerun-cancel` with flag `---models` to rerun a **subset** of failed or canceled model.
+
+```bash
+python .dev_scripts/train_benchmark.py mm_lol \
+    --job-name RERUN \
+    --rerun train-20221009-211904.log \
+    --rerun-fail \
+    --models sagan \  # only rerun 'sagan' models in all failed tasks
+    --run
+```
+
+Specifically, `--rerun-fail` and `--rerun-cancel` can be used together to rerun both failed and cancaled jobs.
+
+## 8. `deterministic` training
+
+Set `torch.backends.cudnn.deterministic = True` and `torch.backends.cudnn.benchmark = False` can remove randomness operation in Pytorch training. You can add `--deterministic` flag when start your benchmark training to remove the influence of randomness operation.
+
+```shell
+python .dev_scripts/train_benchmark.py mm_lol --job-name xzn --models pix2pix --cpus-per-job 16 --run --deterministic
+```
+
+## 9. Automatically check links
+
+Use the following script to check whether the links in documentations are valid:
+
+```shell
+python3 .github/scripts/doc_link_checker.py --target docs/zh_cn
+python3 .github/scripts/doc_link_checker.py --target README_zh-CN.md
+python3 .github/scripts/doc_link_checker.py --target docs/en
+python3 .github/scripts/doc_link_checker.py --target README.md
+```
+
+You can specify the `--target` by a file or a directory.
+
+**Notes:** DO NOT use it in CI, because requiring too many http requirements by CI will cause 503 and CI will propabaly fail.

.dev_scripts/create_ceph_configs.py

@@ -40,7 +40,7 @@ def convert_data_config(data_cfg):
     dataset: dict = data_cfg['dataset']
 
     dataset_type: str = dataset['type']
-    if 'mmcls' in dataset_type:
+    if dataset_type in ['ImageNet', 'CIFAR10']:
         repo_name = 'classification'
     else:
         repo_name = 'editing'
@@ -112,8 +112,6 @@ def convert_data_config(data_cfg):
                     bg_dir_path = bg_dir_path.replace(dataroot_prefix,
                                                       ceph_dataroot_prefix)
                     bg_dir_path = bg_dir_path.replace(repo_name, 'detection')
-                    bg_dir_path = bg_dir_path.replace('openmmlab:',
-                                                      'sproject:')
                 pipeline['bg_dir'] = bg_dir_path
             elif type_ == 'CompositeFg':
                 fg_dir_path = pipeline['fg_dirs']
@@ -188,9 +186,10 @@ def update_ceph_config(filename, args, dry_run=False):
 
         # 2. change visualizer
         if hasattr(config, 'vis_backends'):
-            for vis_cfg in config['vis_backends']:
-                if vis_cfg['type'] == 'GenVisBackend':
-                    vis_cfg['ceph_path'] = work_dir
+            # TODO: support upload to ceph
+            # for vis_cfg in config['vis_backends']:
+            #     if vis_cfg['type'] == 'GenVisBackend':
+            #         vis_cfg['ceph_path'] = work_dir
 
             # add pavi config
             if args.add_pavi:

.dev_scripts/doc_link_checker.py

@@ -0,0 +1,85 @@
+# Copyright (c) MegFlow. All rights reserved.
+# /bin/python3
+
+import argparse
+import os
+import re
+
+
+def make_parser():
+    parser = argparse.ArgumentParser('Doc link checker')
+    parser.add_argument(
+        '--http', default=False, type=bool, help='check http or not ')
+    parser.add_argument(
+        '--target',
+        default='./docs',
+        type=str,
+        help='the directory or file to check')
+    return parser
+
+
+pattern = re.compile(r'\[.*?\]\(.*?\)')
+
+
+def analyze_doc(home, path):
+    print('analyze {}'.format(path))
+    problem_list = []
+    code_block = 0
+    with open(path) as f:
+        lines = f.readlines()
+        for line in lines:
+            line = line.strip()
+            if line.startswith('```'):
+                code_block = 1 - code_block
+
+            if code_block > 0:
+                continue
+
+            if '[' in line and ']' in line and '(' in line and ')' in line:
+                all = pattern.findall(line)
+                for item in all:
+                    # skip  ![]()
+                    if item.find('[') == item.find(']') - 1:
+                        continue
+
+                    # process the case [text()]()
+                    offset = item.find('](')
+                    if offset == -1:
+                        continue
+                    item = item[offset:]
+                    start = item.find('(')
+                    end = item.find(')')
+                    ref = item[start + 1:end]
+
+                    if ref.startswith('http') or ref.startswith('#'):
+                        continue
+                    if '.md#' in ref:
+                        ref = ref[ref.find('#'):]
+                    fullpath = os.path.join(home, ref)
+                    if not os.path.exists(fullpath):
+                        problem_list.append(ref)
+            else:
+                continue
+    if len(problem_list) > 0:
+        print(f'{path}:')
+        for item in problem_list:
+            print(f'\t {item}')
+        print('\n')
+        raise Exception('found link error')
+
+
+def traverse(target):
+    if os.path.isfile(target):
+        analyze_doc(os.path.dirname(target), target)
+        return
+    for home, dirs, files in os.walk(target):
+        for filename in files:
+            if filename.endswith('.md'):
+                path = os.path.join(home, filename)
+                if os.path.islink(path) is False:
+                    analyze_doc(home, path)
+
+
+if __name__ == '__main__':
+    args = make_parser().parse_args()
+    traverse(args.target)

.dev_scripts/download_models.py

@@ -76,6 +76,7 @@ def download(args):
 
     http_prefix_long = 'https://openmmlab-share.oss-cn-hangzhou.aliyuncs.com/mmediting/'  # noqa
     http_prefix_short = 'https://download.openmmlab.com/mmediting/'
+    http_prefix_gen = 'https://download.openmmlab.com/mmgen/'
 
     # load model list
     if args.model_list:
@@ -112,6 +113,11 @@ def download(args):
             model_name = model_weight_url[len(http_prefix_long):]
         elif model_weight_url.startswith(http_prefix_short):
             model_name = model_weight_url[len(http_prefix_short):]
+        elif model_weight_url.startswith(http_prefix_gen):
+            model_name = model_weight_url[len(http_prefix_gen):]
+        elif model_weight_url == '':
+            print(f'{model_info.Name} weight is missing')
+            return None
         else:
             raise ValueError(f'Unknown url prefix. \'{model_weight_url}\'')
 

.dev_scripts/job_watcher.py

@@ -9,7 +9,7 @@
 from pygments.util import ClassNotFound
 from simple_term_menu import TerminalMenu
 
-CACHE_DIR = '~/.task_watcher'
+CACHE_DIR = osp.join(osp.abspath('~'), '.task_watcher')
 
 
 def show_job_out(name, root, job_name_list):

.dev_scripts/metric_mapping.py

@@ -1,5 +1,10 @@
 # key-in-metafile: key-in-results.pkl
 METRICS_MAPPING = {
+    'FID': {
+        'keys': ['FID-Full-50k/fid'],
+        'tolerance': 0.5,
+        'rule': 'less'
+    },
     'PSNR': {
         'keys': ['PSNR'],
         'tolerance': 0.1,

.dev_scripts/test_benchmark.py

@@ -100,12 +100,18 @@ def create_test_job_batch(commands, model_info, args, port, script_name):
 
     http_prefix_short = 'https://download.openmmlab.com/mmediting/'
     http_prefix_long = 'https://openmmlab-share.oss-cn-hangzhou.aliyuncs.com/mmediting/'  # noqa
+    http_prefix_gen = 'https://download.openmmlab.com/mmgen/'
     model_weight_url = model_info.weights
 
     if model_weight_url.startswith(http_prefix_long):
         model_name = model_weight_url[len(http_prefix_long):]
     elif model_weight_url.startswith(http_prefix_short):
         model_name = model_weight_url[len(http_prefix_short):]
+    elif model_weight_url.startswith(http_prefix_gen):
+        model_name = model_weight_url[len(http_prefix_gen):]
+    elif model_weight_url == '':
+        print(f'{fname} weight is missing')
+        return None
     else:
         raise ValueError(f'Unknown url prefix. \'{model_weight_url}\'')