My submission file is correct but still I getting this error. Can anybody explain it??
", "messages": [ { "id": "915927", "postDate": "07/05/2020 07:58:26", "content": "My submission file is correct but still I getting this error. Can anybody explain it??
", "rawMarkdown": "My submission file is correct but still I getting this error. Can anybody explain it??", "votes": null }, { "id": "915931", "postDate": "07/05/2020 08:02:48", "content": "@ravirajsinh45 \nYou must limit your notebook's execution time in about 9h when using only CPU and 6h for GPU. More details could be referred here
", "rawMarkdown": "ravirajsinh45 \nYou must limit your notebook's execution time in about 9h when using only CPU and 6h for GPU. More details could be referred [*here*](https://www.kaggle.com/c/global-wheat-detection/overview/code-requirements)", "votes": null }, { "id": "915963", "postDate": "07/05/2020 08:43:39", "content": "@phunghieu Thank you for quick response. my notebook runtime is only 58 seconds why still i getting this error??
", "rawMarkdown": "phunghieu Thank you for quick response. my notebook runtime is only 58 seconds why still i getting this error??", "votes": null }, { "id": "916042", "postDate": "07/05/2020 09:51:21", "content": "@ravirajsinh45 Because there are only 10 test images, in the test folder, used during the committing process, yet it could be thousands of them during the submitting one 😉.
@phunghieu So how to submit it?? Any advice??
", "rawMarkdown": "phunghieu So how to submit it?? Any advice??", "votes": null }, { "id": "916284", "postDate": "07/05/2020 14:28:19", "content": "@ravirajsinh45 I assume that all your code logics run correctly without any bugs; in that case, the only solution is optimizing your inference pipeline. But when you mentioned 58 seconds for the committing process, I thought that it could be possible to submit in about less than 3h; so, I think you should recheck your notebook first.
", "rawMarkdown": "ravirajsinh45 I assume that all your code logics run correctly without any bugs; in that case, the only solution is optimizing your inference pipeline. But when you mentioned 58 seconds for the committing process, I thought that it could be possible to submit in about less than 3h; so, I think you should recheck your notebook first.", "votes": null }, { "id": "916331", "postDate": "07/05/2020 15:01:45", "content": "@phunghieu I checked but I can’t find anything. Here’s my notebook . If you can help me out:) Thank you
", "rawMarkdown": "phunghieu I checked but I can’t find anything. Here’s my [notebook](https://www.kaggle.com/ravirajsinh45/global-wheat-tensorflow-object-detection-api) . If you can help me out:) Thank you", "votes": null }, { "id": "916461", "postDate": "07/05/2020 17:16:26", "content": "You're doing the inference on only 10 images included in sample_submission. What you should do is doing the inference on all the test images (they are actually private and we don't have access to them). your code runs correctly but during the submission it runs on the whole test set (private).
\n\nYour mistake starts here:
\n\n\n\n\nsub = pd.read_csv(sample_csv)\n plt.figure(figsize=(18,15))\n for ind in tqdm(range(len(sub))) :
\n
I suggest you check some working kernels and see how it is correctly done.
\n\nGood luck!
", "rawMarkdown": "You're doing the inference on only 10 images included in sample_submission. What you should do is doing the inference on all the test images (they are actually private and we don't have access to them). your code runs correctly but during the submission it runs on the whole test set (private). \n\nYour mistake starts here:\n\n> sub = pd.read_csv(sample_csv)\nplt.figure(figsize=(18,15))\nfor ind in tqdm(range(len(sub))) : \n\n\nI suggest you check some working kernels and see how it is correctly done.\n\nGood luck!", "votes": null }, { "id": "916469", "postDate": "07/05/2020 17:27:54", "content": "@ffares Thank You. I will try what you suggested. I have my own working kernels that I submitted previously. I have to observe carefully that what i doing wrong in this kernel.
", "rawMarkdown": "ffares Thank You. I will try what you suggested. I have my own working kernels that I submitted previously. I have to observe carefully that what i doing wrong in this kernel.", "votes": null }, { "id": "916903", "postDate": "07/06/2020 05:31:56", "content": "@ffares @phunghieu I solved this error and submitted Successfully :) What i am doing wrong is i starting session inside for loop which increasing running time due to starting and closing of session.
\n\n```\nfor ind,image_name in tqdm(enumerate(os.listdir(test_images_path))) :
\n\n img_name = image_name\n with tf.compat.v1.Session() as sess:\n sess.graph.as_default()\n tf.import_graph_def(graph_def, name='')\n\nthan i changed it to this\n\nwith tf.compat.v1.Session() as sess:\n # Restore session\n sess.graph.as_default()\n tf.import_graph_def(graph_def, name='')
for ind,image_name in tqdm(enumerate(os.listdir(test_images_path))) :\n\n img_name = image_name\n```\nwhich worked. Thank you for your response:)
", "rawMarkdown": "ffares @phunghieu I solved this error and submitted Successfully :) What i am doing wrong is i starting session inside **for** loop which increasing running time due to starting and closing of session.\n\n```\nfor ind,image_name in tqdm(enumerate(os.listdir(test_images_path))) :\n\n img_name = image_name\n with tf.compat.v1.Session() as sess:\n sess.graph.as_default()\n tf.import_graph_def(graph_def, name='')\n```\nthan i changed it to this\n```\nwith tf.compat.v1.Session() as sess:\n # Restore session\n sess.graph.as_default()\n tf.import_graph_def(graph_def, name='')\n\n for ind,image_name in tqdm(enumerate(os.listdir(test_images_path))) :\n\n img_name = image_name\n```\nwhich worked. Thank you for your response:)", "votes": null }, { "id": "917151", "postDate": "07/06/2020 09:19:03", "content": "You're welcome and congrats for solving the issue :)
", "rawMarkdown": "You're welcome and congrats for solving the issue :)", "votes": null } ], "comments": [ { "id": 915931, "author_name": "phunghieu", "author_url": "", "post_date": "07/05/2020 08:02:48", "content": "@ravirajsinh45 \nYou must limit your notebook's execution time in about 9h when using only CPU and 6h for GPU. More details could be referred here
", "votes": null, "replies": [ { "id": 915963, "author_name": "ravirajsinh45", "author_url": "", "post_date": "07/05/2020 08:43:39", "content": "@phunghieu Thank you for quick response. my notebook runtime is only 58 seconds why still i getting this error??
", "votes": null, "replies": [] }, { "id": 916042, "author_name": "phunghieu", "author_url": "", "post_date": "07/05/2020 09:51:21", "content": "@ravirajsinh45 Because there are only 10 test images, in the test folder, used during the committing process, yet it could be thousands of them during the submitting one 😉.
@phunghieu So how to submit it?? Any advice??
", "votes": null, "replies": [] }, { "id": 916284, "author_name": "phunghieu", "author_url": "", "post_date": "07/05/2020 14:28:19", "content": "@ravirajsinh45 I assume that all your code logics run correctly without any bugs; in that case, the only solution is optimizing your inference pipeline. But when you mentioned 58 seconds for the committing process, I thought that it could be possible to submit in about less than 3h; so, I think you should recheck your notebook first.
", "votes": null, "replies": [] }, { "id": 916331, "author_name": "ravirajsinh45", "author_url": "", "post_date": "07/05/2020 15:01:45", "content": "@phunghieu I checked but I can’t find anything. Here’s my notebook . If you can help me out:) Thank you
", "votes": null, "replies": [] }, { "id": 916461, "author_name": "ffares", "author_url": "", "post_date": "07/05/2020 17:16:26", "content": "You're doing the inference on only 10 images included in sample_submission. What you should do is doing the inference on all the test images (they are actually private and we don't have access to them). your code runs correctly but during the submission it runs on the whole test set (private).
\n\nYour mistake starts here:
\n\n\n\n\nsub = pd.read_csv(sample_csv)\n plt.figure(figsize=(18,15))\n for ind in tqdm(range(len(sub))) :
\n
I suggest you check some working kernels and see how it is correctly done.
\n\nGood luck!
", "votes": null, "replies": [] }, { "id": 916469, "author_name": "ravirajsinh45", "author_url": "", "post_date": "07/05/2020 17:27:54", "content": "@ffares Thank You. I will try what you suggested. I have my own working kernels that I submitted previously. I have to observe carefully that what i doing wrong in this kernel.
", "votes": null, "replies": [] }, { "id": 916903, "author_name": "ravirajsinh45", "author_url": "", "post_date": "07/06/2020 05:31:56", "content": "@ffares @phunghieu I solved this error and submitted Successfully :) What i am doing wrong is i starting session inside for loop which increasing running time due to starting and closing of session.
\n\n```\nfor ind,image_name in tqdm(enumerate(os.listdir(test_images_path))) :
\n\n img_name = image_name\n with tf.compat.v1.Session() as sess:\n sess.graph.as_default()\n tf.import_graph_def(graph_def, name='')\n\nthan i changed it to this\n\nwith tf.compat.v1.Session() as sess:\n # Restore session\n sess.graph.as_default()\n tf.import_graph_def(graph_def, name='')
for ind,image_name in tqdm(enumerate(os.listdir(test_images_path))) :\n\n img_name = image_name\n```\nwhich worked. Thank you for your response:)
", "votes": null, "replies": [] }, { "id": 917151, "author_name": "ffares", "author_url": "", "post_date": "07/06/2020 09:19:03", "content": "You're welcome and congrats for solving the issue :)
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "915927": "My submission file is correct but still I getting this error. Can anybody explain it??", "915931": "ravirajsinh45 \nYou must limit your notebook's execution time in about 9h when using only CPU and 6h for GPU. More details could be referred [*here*](https://www.kaggle.com/c/global-wheat-detection/overview/code-requirements)", "915963": "phunghieu Thank you for quick response. my notebook runtime is only 58 seconds why still i getting this error??", "916042": "ravirajsinh45 Because there are only 10 test images, in the `test` folder, used during the committing process, yet it could be thousands of them during the submitting one 😉.", "916264": "phunghieu So how to submit it?? Any advice??", "916284": "ravirajsinh45 I assume that all your code logics run correctly without any bugs; in that case, the only solution is optimizing your inference pipeline. But when you mentioned 58 seconds for the committing process, I thought that it could be possible to submit in about less than 3h; so, I think you should recheck your notebook first.", "916331": "phunghieu I checked but I can’t find anything. Here’s my [notebook](https://www.kaggle.com/ravirajsinh45/global-wheat-tensorflow-object-detection-api) . If you can help me out:) Thank you", "916461": "You're doing the inference on only 10 images included in sample_submission. What you should do is doing the inference on all the test images (they are actually private and we don't have access to them). your code runs correctly but during the submission it runs on the whole test set (private). \n\nYour mistake starts here:\n\n> sub = pd.read_csv(sample_csv)\nplt.figure(figsize=(18,15))\nfor ind in tqdm(range(len(sub))) : \n\n\nI suggest you check some working kernels and see how it is correctly done.\n\nGood luck!", "916469": "ffares Thank You. I will try what you suggested. I have my own working kernels that I submitted previously. I have to observe carefully that what i doing wrong in this kernel.", "916903": "ffares @phunghieu I solved this error and submitted Successfully :) What i am doing wrong is i starting session inside **for** loop which increasing running time due to starting and closing of session.\n\n```\nfor ind,image_name in tqdm(enumerate(os.listdir(test_images_path))) :\n\n img_name = image_name\n with tf.compat.v1.Session() as sess:\n sess.graph.as_default()\n tf.import_graph_def(graph_def, name='')\n```\nthan i changed it to this\n```\nwith tf.compat.v1.Session() as sess:\n # Restore session\n sess.graph.as_default()\n tf.import_graph_def(graph_def, name='')\n\n for ind,image_name in tqdm(enumerate(os.listdir(test_images_path))) :\n\n img_name = image_name\n```\nwhich worked. Thank you for your response:)", "917151": "You're welcome and congrats for solving the issue :)" }, "source": "meta" }