search

thuiar / MMSA-FET

A Tool for extracting multimodal features from videos.
GNU General Public License v3.0
141 stars 21 forks source link

IndexError: list index out of range #9

Open xiaoyouzi12 opened 2 years ago

xiaoyouzi12 commented 2 years ago

您好，我在特征提取过程中可以正常运行，但是到最后保存特征时会报错：

Attempting to read from directory: C:\Users\10376.MMSA-FET/tmp\video_0001_0010 Device or file opened Starting tracking Reading the MTCNN face detector from: E:\openface\OpenFace_2.2.0_win_x64\model/mtcnn_detector/MTCNN_detector.txt Reading the PNet module from: E:\openface\OpenFace_2.2.0_win_x64\model/mtcnn_detector\PNet.dat Reading the RNet module from: E:\openface\OpenFace_2.2.0_win_x64\model/mtcnn_detector\RNet.dat Reading the ONet module from: E:\openface\OpenFace_2.2.0_win_x64\model/mtcnn_detector\ONet.dat 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Closing output recorder Closing input reader Closed successfully Postprocessing the Action Unit predictions 100%|██████████| 1/1 [00:25<00:00, 25.23s/it] MMSA-FET - An Error Occured: Traceback (most recent call last): File "E:/code/MMSA-FET-master/src/MSA_FET/main.py", line 449, in run_dataset final_data[mode][item] = np.array([data[item][v] for v in indexes]) File "E:/code/MMSA-FET-master/src/MSA_FET/main.py", line 449, in final_data[mode][item] = np.array([data[item][v] for v in indexes]) IndexError: list index out of range MMSA-FET - Removing temporary files. Traceback (most recent call last): File "E:/code/MMSA-FET-master/src/MSA_FET/main.py", line 487, in feature = fet.run_dataset(dataset_dir="E:\code\MMSA-FET-master\src\MSA_FET\SIMS", out_file="E:\code\MMSA-FET-master\src\output/feature.pkl") File "E:/code/MMSA-FET-master/src/MSA_FET/main.py", line 483, in run_dataset raise e File "E:/code/MMSA-FET-master/src/MSA_FET/main.py", line 449, in run_dataset final_data[mode][item] = np.array([data[item][v] for v in indexes]) File "E:/code/MMSA-FET-master/src/MSA_FET/main.py", line 449, in final_data[mode][item] = np.array([data[item][v] for v in indexes]) IndexError: list index out of range

我是在windows上运行的

FlameSky-S commented 2 years ago

您好，我们无法复现这个问题。请问您是抽取的自定义数据集的特征吗？该数据集只有一条数据吗？

AIzld commented 1 year ago

您好，我在特征提取过程中可以正常运行，但是到最后保存特征时会报错：

Attempting to read from directory: C:\Users\10376.MMSA-FET/tmp\video_0001_0010 Device or file opened Starting tracking Reading the MTCNN face detector from: E:\openface\OpenFace_2.2.0_win_x64\model/mtcnn_detector/MTCNN_detector.txt Reading the PNet module from: E:\openface\OpenFace_2.2.0_win_x64\model/mtcnn_detector\PNet.dat Reading the RNet module from: E:\openface\OpenFace_2.2.0_win_x64\model/mtcnn_detector\RNet.dat Reading the ONet module from: E:\openface\OpenFace_2.2.0_win_x64\model/mtcnn_detector\ONet.dat 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Closing output recorder Closing input reader Closed successfully Postprocessing the Action Unit predictions 100%|██████████| 1/1 [00:25<00:00, 25.23s/it] MMSA-FET - An Error Occured: Traceback (most recent call last): File "E:/code/MMSA-FET-master/src/MSA_FET/main.py", line 449, in run_dataset final_data[mode][item] = np.array([data[item][v] for v in indexes]) File "E:/code/MMSA-FET-master/src/MSA_FET/main.py", line 449, in final_data[mode][item] = np.array([data[item][v] for v in indexes]) IndexError: list index out of range MMSA-FET - Removing temporary files. Traceback (most recent call last): File "E:/code/MMSA-FET-master/src/MSA_FET/main.py", line 487, in feature = fet.run_dataset(dataset_dir="E:\code\MMSA-FET-master\src\MSA_FET\SIMS", out_file="E:\code\MMSA-FET-master\src\output/feature.pkl") File "E:/code/MMSA-FET-master/src/MSA_FET/main.py", line 483, in run_dataset raise e File "E:/code/MMSA-FET-master/src/MSA_FET/main.py", line 449, in run_dataset final_data[mode][item] = np.array([data[item][v] for v in indexes]) File "E:/code/MMSA-FET-master/src/MSA_FET/main.py", line 449, in final_data[mode][item] = np.array([data[item][v] for v in indexes]) IndexError: list index out of range

我是在windows上运行的

请问这个问题解决了么 我遇到了这个问题

xlggzzz commented 1 year ago

您好，我在Linux系统上提取MOSI数据集定制特征时多次遇到了这个错误，但程序还在继续运行！我不知道如何修复这个错误，这个越界是否会影响特征的抽取所以向您请教下！ 我的原始代码如下： from MSA_FET import run_dataset if name == 'main':

Extract audio features for MOSI using default aligned feature config

run_dataset(
    config='aligned',
    dataset_dir='/data1/lx/data/Raw/CMU-MOSI/',
    out_file='/data1/lx/FET/feature_aligned.pkl',
    # num_workers = 1
)
print("finished")

报错信息如下： 2023-06-30 18:03:14 - FET-Subprocess - PID:1768915 [ERROR] Ignore error and continue, see log for details. 2023-06-30 18:03:14 - FET-Subprocess - PID:1768915 [ERROR] index 32 is out of bounds for axis 0 with size 32 Traceback (most recent call last): File "/data/lx/anaconda3/envs/msa/lib/python3.10/site-packages/MSA_FET/dataset.py", line 118, in extract_one align_result = aligner.align_with_transcript(video_path, text) File "/data/lx/anaconda3/envs/msa/lib/python3.10/site-packages/MSA_FET/aligners/wav2vec2.py", line 67, in align_with_transcript segments = ctc_segmentation.determine_utterance_segments(config, utt_begin_indices, char_probs, timings, transcripts) File "/data/lx/anaconda3/envs/msa/lib/python3.10/site-packages/ctc_segmentation/ctc_segmentation.py", line 391, in determine_utterance_segments start = compute_time(utt_begin_indices[i], "begin") File "/data/lx/anaconda3/envs/msa/lib/python3.10/site-packages/ctc_segmentation/ctc_segmentation.py", line 384, in compute_time return max(timings[index + 1] - 0.5, middle) IndexError: index 32 is out of bounds for axis 0 with size 32 程序执行完结果： 2023-06-30 18:21:49 - FET-Subprocess - PID:1765176 [ERROR] Ignore error and continue, see log for details. 2023-06-30 18:21:49 - FET-Subprocess - PID:1765176 [ERROR] index 53 is out of bounds for axis 0 with size 53 Traceback (most recent call last): File "/data/lx/anaconda3/envs/msa/lib/python3.10/site-packages/MSA_FET/dataset.py", line 118, in extract_one align_result = aligner.align_with_transcript(video_path, text) File "/data/lx/anaconda3/envs/msa/lib/python3.10/site-packages/MSA_FET/aligners/wav2vec2.py", line 67, in align_with_transcript segments = ctc_segmentation.determine_utterance_segments(config, utt_begin_indices, char_probs, timings, transcripts) File "/data/lx/anaconda3/envs/msa/lib/python3.10/site-packages/ctc_segmentation/ctc_segmentation.py", line 391, in determine_utterance_segments start = compute_time(utt_begin_indices[i], "begin") File "/data/lx/anaconda3/envs/msa/lib/python3.10/site-packages/ctc_segmentation/ctc_segmentation.py", line 384, in compute_time return max(timings[index + 1] - 0.5, middle) IndexError: index 53 is out of bounds for axis 0 with size 53 100%|███████████████████████████████████████| 2199/2199 [42:00<00:00, 1.15s/it] 2023-06-30 18:27:49 - FET-Dataset - PID:1763448 [INFO] Feature file saved: '/data1/lx/FET/feature_aligned.pkl'. 2023-06-30 18:27:49 - FET-Dataset - PID:1763448 [INFO] Feature extraction complete! Process finished with exit code 0

请教一下这个IndexError影响特征的提取吗？期待和感谢您的回复!