jaywongwang / sst-tensorflow Goto Github PK
View Code? Open in Web Editor NEWTensorflow Implementation of the Paper "SST: Single-Stream Temporal Action Proposals" in CVPR 2017.
Tensorflow Implementation of the Paper "SST: Single-Stream Temporal Action Proposals" in CVPR 2017.
Hi,Dr Wang!
I am trying to run this project according to the README. I successed to run train.py,but when I run test.py,there is a NotFoundError message:
Caused by op u'save/RestoreV2_1', defined at:
File "test.py", line 196, in
test(options)
File "test.py", line 89, in test
saver = tf.train.Saver()
NotFoundError (see above for traceback): Unsuccessful TensorSliceReader constructor: Failed to find any matching files for
[[Node: save/RestoreV2_1 = RestoreV2[dtypes=[DT_FLOAT], _device="/job:localhost/replica:0/task:0/cpu:0"](_recv_save/Const_0, save/RestoreV2_1/tensor_names, save/RestoreV2_1/shape_and_slices)]]
[[Node: save/RestoreV2_6/_5 = _Recvclient_terminated=false, recv_device="/job:localhost/replica:0/task:0/gpu:0", send_device="/job:localhost/replica:0/task:0/cpu:0", send_device_incarnation=1, tensor_name="edge_37_save/RestoreV2_6", tensor_type=DT_FLOAT, _device="/job:localhost/replica:0/task:0/gpu:0"]]
How can I solve it?Need I modify the test.py code?Thank u!
Hi Wang, as mentioned Thumos videos size [412 videos (200 val videos + 212 test videos, I found one test video missing)] can be downloaded from Thumos4 webpage https://www.crcv.ucf.edu/THUMOS14/download.html.
However, I can see UCF101 link only and that contains 13320/1010/1574 train/val/test videos. Can you please share the Thumos14 videos link if possible? How much is its train size and overall size?
Thanks
Hello,
I was trying to run the code and when I ran train.py i had this error message:
"OSError: Unable to open file (unable to open file: name = 'dataset/thumos14/features/thumos14_c3d_fc6.hdf5', errno = 2, error message = 'No such file or directory')"
I would like to know how to create that missing file in features folder or if you have a version of it and you could send it to me I would appreciate it.
There's something I don't understand regarding the goundtruth proposals generated, I don't understand how they don't contain the action type of each proposal. I expected that for each proposal, the action class would be included but that wasn't the case. So how does the model know this when it's training?
I'm sorry to post this question here but I couldn't find your email to contact you.
Hi
Thanks for the code.
I try to extract the feature according to the github address you mentioned(https://github.com/yyuanad/Pytorch_C3D_Feature_Extractor). But I got a complete different feature file, which contains huge values. But in your feature file the value is small. Could you please show me how you extract the feature from videos.
Any other ways to access pretrained models? We dont have access to Baidu Drive here.
在get_anchor_weight.py文件中求weight的时候最后除以sum_sample_length,这个sum_sample_length到底代表什么?
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.