model_step3_exp1_old.ckpt about rolo

[paozhuanyinyuba]

@bipenalo Hello，check the code, and correct the path setting in the original code

[bipenalo]

Thanks.. it works now..

[bipenalo]

@paozhuanyinyuba Hello, one more question. I am trying to run experiment 1, however I get this error 👍 IOError: [Errno 2] No such file or directory: 'output/trainging-20-log.txt'

where do I get that file? Do you know? Thanks

[paozhuanyinyuba]

@bipenalo i thing that it's probably a path setting question, though i haven't done the experiment 1.

[SHMCU]

Just download the model provided on the ROLO code webpage, and store it somewhere, and revise the path variable
self.rolo_weights_file , it is around line 261 in the file ROLO_network_test_all.py
to make it point to the model folder.

[sunshinezhihuo]

@SHMCU Hello，do you run MOLO sucessfully?

[wei-1234567]

ValueError：使用无效保存路径调用恢复/u03/Guanghan/dev/ROLO-dev/output/ROLO_model/model_step3_exp1_old.ckpt 如何解决这个问题

[wei-1234567]

output / trainging-20-log.txt 这个有人知道怎么回事吗

[SHMCU]

I think now there are many other networks that can work for multiple object detection in video.There is a paper using a backbone network and LSTM and a spatial transform layer to detection objects in video. I could not find this paper now, but it was somewhere online. The proposed method CNN+spatial transform layer+LSTM is better than CNN+LSTM scheme.If you found this one, could you email me its title? Thank you very much! Hai On Friday, March 30, 2018, 8:06:24 AM EDT, Qi <[email protected]> wrote: @SHMCU Hello，do you run MOLO sucessfully? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.

[SHMCU]

Aha, found it: https://arxiv.org/pdf/1712.06317.pdf Previously I tried SSD + convLSTM, it helps a little, but not significantly, I think the above work solved the problem using spatial transform layer. Discussions are welcome! Best,Hai On Friday, March 30, 2018, 8:06:24 AM EDT, Qi <[email protected]> wrote: @SHMCU Hello，do you run MOLO sucessfully? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.

[yogidosalwar]

@bipenalo
How you solve this problem.
I can't find that file. model_step3_exp1_old.ckpt

Hello,,
I have been trying to run the ROLO_network_test_all.py code. However, I keep getting this error
ValueError: Restore called with invalid save path /u03/Guanghan/dev/ROLO-dev/output/ROLO_model/model_step3_exp1_old.ckpt

Can anyone tell me how to go around this issue?
I would deeply appreciate it.

[bipenalo]

@yogidosalwar
Hello man. You need to change the path u03/Guanghan/dev/ROLO-dev/output/ROLO_model/model_step3_exp1_old.ckpt to a valid path in your computer. I think that is what I did to solve it!

[yogidosalwar]

HI Guanghan, Thank you for the quick response. I don't have model_step3_exp1_old.ckpt file.

…

-- *Thank You...* *With Regards..* *Mr. Yugendra L. Dosalwar* Contact No. 9970642251 Mail id: [email protected] <[email protected]>

On Sat, Jul 13, 2019 at 11:57 PM Boris Penaloza ***@***.***> wrote: @yogidosalwar <https://github.com/yogidosalwar> Hello man. You need to change the path *u03/Guanghan/dev/ROLO-dev/output/ROLO_model/model_step3_exp1_old.ckpt* to a valid path in your computer. I think that is what I did to solve it! — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#6?email_source=notifications&email_token=ALZRXI3LSLL6MUZSYBP2NSDP7INB3A5CNFSM4CXIJZMKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODZ3W7NI#issuecomment-511143861>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ALZRXI6XAII4XK6IVZ4PKITP7INB3ANCNFSM4CXIJZMA> .