【实习日志】TSN 网络学习与实验记录/Constructor returned NULL && protobuf import error

这篇实习日志详述了在复现TSN网络过程中遇到的错误,包括constructor返回NULL和protobuf导入错误。通过查阅资料和解决依赖问题,如更新libboost、安装protobuf和openmpi,最终成功解决问题并继续实验。

摘要生成于 C知道 ,由 DeepSeek-R1 满血版支持, 前往体验 >

 

实验目标:复现TSN网络结果

代码:https://github.com/yjxiong/temporal-segment-networks

TSN介绍:https://zhuanlan.zhihu.com/p/32777430

到video-level testing这一步时:

python tools/eval_net.py ucf101 1 rgb FRAME_PATH \
 models/ucf101/tsn_bn_inception_rgb_deploy.prototxt models/ucf101_split_1_tsn_rgb_reference_bn_inception.caffemodel \
 --num_worker 4 --save_scores SCORE_FILE

出错:一、constructor returns null

libdc1394 error: Failed to initialize libdc1394
/usr/local/lib/python2.7/dist-packages/scipy/spatial/__init__.py:95: RuntimeWarning: numpy.dtype size changed, may indicate binary incompatibility. Expected 96, got 88
  from .ckdtree import *
/usr/local/lib/python2.7/dist-packages/scipy/spatial/__init__.py:96: RuntimeWarning: numpy.dtype size changed, may indicate binary incompatibility. Expected 96, got 88
  from .qhull import *
Namespace(caffe_path='./lib/caffe-action/', dataset='ucf101', flow_x_prefix='flow_x_', flow_y_prefix='flow_y_', frame_path='OutputImages', gpus=None, modality='rgb', net_proto='models/ucf101/tsn_bn_inception_rgb_deploy.prototxt', net_weights='models/ucf101_split_1_tsn_rgb_reference_bn_inception.caffemodel', num_frame_per_video=25, num_worker=4, rgb_prefix='img_', save_scores='Scores', split=1)
error: XDG_RUNTIME_DIR not set in the environment.
error: XDG_RUNTIME_DIR not set in the environment.

(eval_net.py:23248): Gdk-CRITICAL **: gdk_cursor_new_for_display: assertion 'GDK_IS_DISPLAY (display)' failed
Traceback (most recent call last):
  File "tools/eval_net.py", line 38, in <module>
    from pyActionRecog.action_caffe import CaffeNet
  File "./pyActionRecog/action_caffe.py", line 4, in <module>
    import caffe
  File "/root/caffe/python/caffe/__init__.py", line 1, in <module>
    from .pycaffe import Net, SGDSolver, NesterovSolver, AdaGradSolver, RMSPropSolver, AdaDeltaSolver, AdamSolver, NCCL, Timer
  File "/root/caffe/py
评论 2
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值