Is there any way to see the log of the failed evaluation?

[Anothorld]

My docker runs well locally but failed on the server.

[ljj898]

metoo!! how did you solve it ?

[Anothorld]

metoo!! how did you solve it ?

I found it was a TypeError in my code, I changed the default output type of detector from float to torch.Tensor. but it failed again even after I fix the TypeError

[ljj898]

ME too!!! thanks for you reply very much I also failed first time because of this problem but when I fixed it  failed again  so by now did you solve the second problem????

------------------ 原始邮件 ------------------ 发件人: "Drdestiny/DeeperForensicsChallengeSubmissionExample" <notifications@github.com>; 发送时间: 2020年9月19日(星期六) 下午5:30 收件人: "Drdestiny/DeeperForensicsChallengeSubmissionExample"<DeeperForensicsChallengeSubmissionExample@noreply.github.com>; 抄送: "Rice & Shine"<1140801877@qq.com>;"Comment"<comment@noreply.github.com>; 主题: Re: [Drdestiny/DeeperForensicsChallengeSubmissionExample] Is there any way to see the log of the failed evaluation? (#2)

metoo!! how did you solve it ?

I found it was a TypeError in my code, I used changed the default output type of detector from float to torch.Tensor. but it failed again even after I fix the TypeError

— You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

[Drdestiny]

I found it was a TypeError in my code, I changed the default output type of detector from float to torch.Tensor. but it failed again even after I fix the TypeError

It seems you haven't built a new Docker image after you changed your code?

[ljj898]

emm，but I did make new docker images several times. do you mean make new images with a different name from first one？

---Original--- From: "Drdestiny"<notifications@github.com> Date: Sat, Sep 19, 2020 19:34 PM To: "Drdestiny/DeeperForensicsChallengeSubmissionExample"<DeeperForensicsChallengeSubmissionExample@noreply.github.com>; Cc: "Comment"<comment@noreply.github.com>;"ljj898"<1140801877@qq.com>; Subject: Re: [Drdestiny/DeeperForensicsChallengeSubmissionExample] Is there any way to see the log of the failed evaluation? (#2)

I found it was a TypeError in my code, I changed the default output type of detector from float to torch.Tensor. but it failed again even after I fix the TypeError

It seems you haven't built a new Docker image after you change your code?

— You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

[Anothorld]

I did rebuild my Docker after I changed my code, but it still didn't work. I'm checking my frame extractor and the face detector. maybe there' some bugs inside them. I'm wondering if there is any test video contained more than one face? My extractor only works when there is only one face in the video.

[Drdestiny]

Have you tried docker run -it --rm --gpus all <image-name> python local_test.py? Running in local environment is different from running in Docker image because you might not install dependencies in Dockerfile.

[ljj898]

A little difference. I tried like this: docker run -it --gpus all <image-name> python local_test.py?

without  ——rm

---Original--- From: "Drdestiny"<notifications@github.com> Date: Sat, Sep 19, 2020 20:06 PM To: "Drdestiny/DeeperForensicsChallengeSubmissionExample"<DeeperForensicsChallengeSubmissionExample@noreply.github.com>; Cc: "Comment"<comment@noreply.github.com>;"ljj898"<1140801877@qq.com>; Subject: Re: [Drdestiny/DeeperForensicsChallengeSubmissionExample] Is there any way to see the log of the failed evaluation? (#2)

Have you tried docker run -it --rm --gpus all <image-name> python local_test.py?

— You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

[Drdestiny]

A little difference. I tried like this: docker run -it --gpus all <image-name> python local_test.py?

without  ——rm

That's okay. --rm just deletes this container after running.

[EndlessSora]

I did rebuild my Docker after I changed my code, but it still didn't work. I'm checking my frame extractor and the face detector. maybe there' some bugs inside them. I'm wondering if there is any test video contained more than one face? My extractor only works when there is only one face in the video.

I think making your extractor robust to diverse situations is a good choice. :-)

[Drdestiny]

Please give us your email account.

[ljj898]

030849093930

---Original--- From: "Drdestiny"<notifications@github.com> Date: Sat, Sep 19, 2020 20:25 PM To: "Drdestiny/DeeperForensicsChallengeSubmissionExample"<DeeperForensicsChallengeSubmissionExample@noreply.github.com>; Cc: "Comment"<comment@noreply.github.com>;"ljj898"<1140801877@qq.com>; Subject: Re: [Drdestiny/DeeperForensicsChallengeSubmissionExample] Is there any way to see the log of the failed evaluation? (#2)

Please give us your AWS ID.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

[Anothorld]

I did rebuild my Docker after I changed my code, but it still didn't work. I'm checking my frame extractor and the face detector. maybe there' some bugs inside them. I'm wondering if there is any test video contained more than one face? My extractor only works when there is only one face in the video.

I think making your extractor robust to diverse situations is a good choice. :-)

thank you, I'll try to make it more robust.

[Anothorld]

811370425750

[Drdestiny]

811370425750

Anothorld, the reason is timeout, not bug.

[Drdestiny]

030849093930

ljj898, the log of your submission is like this:

Traceback (most recent call last):
File "run_evaluation.py", line 80, in <module>
evaluate_runtime(start_time, DeeperForensicsDetector, deeper_forensics_frames_iter, job_name)
File "run_evaluation.py", line 53, in evaluate_runtime
prob = detector.predict(frames)
File "model/toy_predict.py", line 12, in predict
result = super(FakePredictor, self).predict(video_frames)
File "/workspace/DeeperForensics_Eval/eval_kit/detector.py", line 84, in predict
probability = self.process_video(video_frames)
File "/workspace/DeeperForensics_Eval/eval_kit/detector.py", line 58, in process_video
return prob
UnboundLocalError: local variable 'prob' referenced before assignment

[xstju]

Have you tried docker run -it --rm --gpus all <image-name> python local_test.py? Running in local environment is different from running in Docker image because you might not install dependencies in Dockerfile.

I have meet the same problem, My docker runs well locally but failed on the server. I have tried docker run -it --rm --gpus all python local_test.py . It is well. So can you provide the log of my submission？ My aws-id 334889259927 Thank you