train c3d on ucf101-nori

README.md

@@ -80,4 +80,3 @@ accelerate launch --config_file=tools/single_acc.yml --num_processes=8 main.py -
 ## Thanks
 
 部分代码和组织结构参考[BasicSR](https://github.com/XPixelGroup/BasicSR), [MMAction2](https://github.com/open-mmlab/mmaction2)以及其他卓越的工作。
-

mmaction/datasets/transforms/loading.py

@@ -1406,8 +1406,6 @@ def transform(self, results: dict) -> dict:
 
         act_name, video_name = directory.split('/')[-2:]
 
-        print(act_name, video_name)
-
         if self.file_client is None:
             self.file_client = FileClient(self.io_backend, **self.kwargs)
 
@@ -1430,12 +1428,10 @@ def transform(self, results: dict) -> dict:
             frame_idx += offset
             if modality == 'RGB':
                 filepath = osp.join(directory, filename_tmpl.format(frame_idx))
-                # print(directory)
                 img_bytes = self.file_client.get(filepath)
                 # Get frame with channel order RGB directly.
                 cur_frame = mmcv.imfrombytes(img_bytes, channel_order='rgb')
                 imgs.append(cur_frame)
-                print(cur_frame.shape, cur_frame.dtype)
             elif modality == 'Flow':
                 x_filepath = osp.join(directory,
                                       filename_tmpl.format('x', frame_idx))

mmaction/datasets/transforms/nori_loading.py

@@ -39,7 +39,7 @@ def __init__(self,
         self.file_client = None
 
         with open(nori_file, 'r') as file:
-            self.nids = json.load(nori_file)
+            self.nids = json.load(file)
 
     def transform(self, results):
         mmcv.use_backend(self.decoding_backend)

usage/nori2_video.md

@@ -83,7 +83,7 @@ train_pipeline = [
 
 ```python
 file_client_args = dict(
-    io_backend='oss',
+    io_backend='disk',
     nori_file = 'data/ucf101/ucf101_train_split_1_nid.json',
     dtype = 'uint8',
     retry = 60
@@ -101,4 +101,65 @@ train_pipeline = [
     dict(type='FormatShape', input_format='NCTHW'),
     dict(type='PackActionInputs')
 ]
+```
+
+### C3D-UCF101-Nori2
+
+```shell
+bash tools/dist_train.sh work_dir/c3d/c3d_pretrained_ncf101_nori.py 1 --work-dir results/c3d_ucf101_nori --seed=0
+```
+相关的配置文件在[work_dir/c3d](../work_dir/c3d/)目录下，这面展示了本地抽帧和nori抽帧配置文件的差异
+
+```python
+--- c3d_pretrained_ncf101_nori.py	2024-03-24 15:46:20.950437836 +0800
++++ c3d_pretrained_ucf101_rgb.py	2024-02-23 10:22:25.396394506 +0800
+@@ -21,19 +21,7 @@
+ ann_file_test = f'/data/dataset/ucf101/ucf101_val_split_{split}_rawframes.txt'
+ ann_file_val = f'/data/dataset/ucf101/ucf101_val_split_{split}_rawframes.txt'
+
+-file_client_args_train = dict(
+-    io_backend='disk',
+-    nori_file = 'data/ucf101/ucf101_train_split_1_nid.json',
+-    dtype = 'uint8',
+-    retry = 60
+-)
+-
+-file_client_args_eval = dict(
+-    io_backend='disk',
+-    nori_file = 'data/ucf101/ucf101_val_split_1_nid.json',
+-    dtype = 'uint8',
+-    retry = 60
+-)
++file_client_args = dict(io_backend='disk')
+
+ # dataset pipeline
+ train_pipeline = [
+@@ -41,7 +29,7 @@
+     # dict(type='DecordInit', **file_client_args),
+     dict(type='SampleFrames', clip_len=16, frame_interval=1, num_clips=1),
+     # dict(type='DecordDecode'),
+-    dict(type='RawFrameDecodeNoir2', **file_client_args_train),
++    dict(type='RawFrameDecode', **file_client_args),
+     dict(type='Resize', scale=(-1, 128)),
+     dict(type='RandomCrop', size=112),
+     dict(type='Flip', flip_ratio=0.5),
+@@ -58,7 +46,7 @@
+         num_clips=1,
+         test_mode=True),
+     # dict(type='DecordDecode'),
+-    dict(type='RawFrameDecodeNoir2', **file_client_args_eval),
++    dict(type='RawFrameDecode', **file_client_args),
+     dict(type='Resize', scale=(-1, 128)),
+     dict(type='CenterCrop', crop_size=112),
+     dict(type='FormatShape', input_format='NCTHW'),
+@@ -74,7 +62,7 @@
+         num_clips=10,
+         test_mode=True),
+     # dict(type='DecordDecode'),
+-    dict(type='RawFrameDecodeNoir2', **file_client_args_eval),
++    dict(type='RawFrameDecode', **file_client_args),
+     dict(type='Resize', scale=(-1, 128)),
+     dict(type='CenterCrop', crop_size=112),
+     dict(type='FormatShape', input_format='NCTHW'),
+
 ```