![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F3126050%2F5a56bfb58d2e29820f856d7f486ec5de%2F2019-12-18%20144344.png?generation=1576648067639365&alt=media\")
\n so, there are no doubt that your submission of private version is all zero.\nthe result you attached is public version submission, not the private version submission.My public highest score is 0.9910. but all my private scores are 0.10260? I don't know what happened as a newer. Who anyone can tell me the reason? thanks very much!
\n\nprivate public\n0.10260 0.98200\n0.10260 0.98660\n0.10260 0.98600\n0.10260 0.98400\n0.10260 0.98200\n0.10260 0.99040\n0.10260 0.99100
\n\nsubmit code is here:
\n\nfilename = 'submerage1110'\nparentpath = '../input/' + filename\ndf_sub = pd.read_csv(parentpath + '/' + filename + '.csv')\nsample_submission = pd.read_csv('../input/Kannada-MNIST/sample_submission.csv')\ndf_merged = pd.merge(sample_submission, df_sub, on='id', how='left')\ndf_merged.fillna(0, inplace=True)\ndf_merged['label'] = df_merged['label_y']\ndf_merged = df_merged[['id', 'label']]\ndf_merged['label'] = df_merged['label'].astype('int64')\ndf_merged.to_csv('submission.csv', index=False)\nctime = time.strftime(\"%H:%M:%S\")\nprint(parentpath + '/' + filename + '.csv')\nprint('==={}==='.format(ctime) + '*' * 20 + 'end' + '*' * 20)
", "messages": [ { "id": "697501", "postDate": "12/18/2019 02:32:26", "content": "My public highest score is 0.9910. but all my private scores are 0.10260? I don't know what happened as a newer. Who anyone can tell me the reason? thanks very much!
\n\nprivate public\n0.10260 0.98200\n0.10260 0.98660\n0.10260 0.98600\n0.10260 0.98400\n0.10260 0.98200\n0.10260 0.99040\n0.10260 0.99100
\n\nsubmit code is here:
\n\nfilename = 'submerage1110'\nparentpath = '../input/' + filename\ndf_sub = pd.read_csv(parentpath + '/' + filename + '.csv')\nsample_submission = pd.read_csv('../input/Kannada-MNIST/sample_submission.csv')\ndf_merged = pd.merge(sample_submission, df_sub, on='id', how='left')\ndf_merged.fillna(0, inplace=True)\ndf_merged['label'] = df_merged['label_y']\ndf_merged = df_merged[['id', 'label']]\ndf_merged['label'] = df_merged['label'].astype('int64')\ndf_merged.to_csv('submission.csv', index=False)\nctime = time.strftime(\"%H:%M:%S\")\nprint(parentpath + '/' + filename + '.csv')\nprint('==={}==='.format(ctime) + '*' * 20 + 'end' + '*' * 20)
", "rawMarkdown": "My public highest score is 0.9910. but all my private scores are 0.10260? I don't know what happened as a newer. Who anyone can tell me the reason? thanks very much!\n\nprivate public\n0.10260 0.98200\n0.10260 0.98660\n0.10260 0.98600\n0.10260 0.98400\n0.10260 0.98200\n0.10260 0.99040\n0.10260 0.99100\n\nsubmit code is here:\n\nfilename = 'submerage1110'\nparentpath = '../input/' + filename\ndf_sub = pd.read_csv(parentpath + '/' + filename + '.csv')\nsample_submission = pd.read_csv('../input/Kannada-MNIST/sample_submission.csv')\ndf_merged = pd.merge(sample_submission, df_sub, on='id', how='left')\ndf_merged.fillna(0, inplace=True)\ndf_merged['label'] = df_merged['label_y']\ndf_merged = df_merged[['id', 'label']]\ndf_merged['label'] = df_merged['label'].astype('int64')\ndf_merged.to_csv('submission.csv', index=False)\nctime = time.strftime(\"%H:%M:%S\")\nprint(parentpath + '/' + filename + '.csv')\nprint('==={}==='.format(ctime) + '*' * 20 + 'end' + '*' * 20)", "votes": null }, { "id": "697545", "postDate": "12/18/2019 04:13:59", "content": "It seems like your code generate all 0 submission when submission, not commit.
", "rawMarkdown": "It seems like your code generate all 0 submission when submission, not commit.", "votes": null }, { "id": "697581", "postDate": "12/18/2019 05:40:01", "content": "Thanks for you reply. But it‘s not like what you say.\nWhen with \"dfmerged.to_csv('submission.csv', index=False)\" run\n My run log: \n id label\n0 0 3\n1 1 0\n2 2 2\n3 3 6\n4 4 7\n5.6s2../input/submerage1110/submerage1110.csv\n===03:31:08===*********end*********
\n\nsubmission.csv in output is:\n1 id label\n2 0 3\n3 1 0\n4 2 2\n5 3 6\n6 4 7\n7 5 7\n8 6 1\n9 7 9\n10 8 3\n11 9 4\n12 10 8\n13 11 8\n14 12 1\n15 13 7\n16 14 8\n17 15 1\n18 16 5\n19 17 1\n20 18 5\n21 19 9\n22 20 3\n23 21 7\n24 22 6\n25 23 0\n26 24 2\n27 25 0\n28 26 8\n29 27 7\n30 28 0\n31 29 0\n32 30 8
\n\nSo not all label is 0.\nThank you again anyway.
", "rawMarkdown": "Thanks for you reply. But it‘s not like what you say.\nWhen with \"dfmerged.to_csv('submission.csv', index=False)\" run\n My run log: \n id label\n0 0 3\n1 1 0\n2 2 2\n3 3 6\n4 4 7\n5.6s2../input/submerage1110/submerage1110.csv\n===03:31:08===********************end********************\n\nsubmission.csv in output is:\n1\tid\tlabel\n2\t0\t3\n3\t1\t0\n4\t2\t2\n5\t3\t6\n6\t4\t7\n7\t5\t7\n8\t6\t1\n9\t7\t9\n10\t8\t3\n11\t9\t4\n12\t10\t8\n13\t11\t8\n14\t12\t1\n15\t13\t7\n16\t14\t8\n17\t15\t1\n18\t16\t5\n19\t17\t1\n20\t18\t5\n21\t19\t9\n22\t20\t3\n23\t21\t7\n24\t22\t6\n25\t23\t0\n26\t24\t2\n27\t25\t0\n28\t26\t8\n29\t27\t7\n30\t28\t0\n31\t29\t0\n32\t30\t8\n\nSo not all label is 0.\nThank you again anyway.", "votes": null }, { "id": "697588", "postDate": "12/18/2019 05:53:52", "content": "I'm sorry, maybe I didn't explain it clearly.
\n\nI remember I had a all zero test for the leader-board, I check it private score just now, the score is:\n\n so, there are no doubt that your submission of private version is all zero.\nthe result you attached is public version submission, not the private version submission.
Maybe you remember that the kaggle need run our code two times, when you press the 'commit' and 'submission', your code will run in different test data. the commit is public 5000 images, and the submission is private 5000 images.
\n\nSo I think there maybe some bugs in your code that result in all zero when press 'submission'.
", "rawMarkdown": "I'm sorry, maybe I didn't explain it clearly. \n\nI remember I had a all zero test for the leader-board, I check it private score just now, the score is:\n\n so, there are no doubt that your submission of private version is all zero.\nthe result you attached is public version submission, not the private version submission.\n\nMaybe you remember that the kaggle need run our code two times, when you press the 'commit' and 'submission', your code will run in different test data. the commit is public 5000 images, and the submission is private 5000 images.\n\nSo I think there maybe some bugs in your code that result in all zero when press 'submission'.", "votes": null }, { "id": "697589", "postDate": "12/18/2019 05:54:29", "content": "I know!\nI only submit a csv file but no more.\na. run train and inference in the kernel.\nb.up load offline model and run inference in the kernel
", "rawMarkdown": "I know!\nI only submit a csv file but no more.\na. run train and inference in the kernel.\nb.up load offline model and run inference in the kernel", "votes": null }, { "id": "697592", "postDate": "12/18/2019 05:57:03", "content": "Thank you for your patience reply.
", "rawMarkdown": "Thank you for your patience reply.", "votes": null } ], "comments": [ { "id": 697545, "author_name": "lhy1997", "author_url": "", "post_date": "12/18/2019 04:13:59", "content": "It seems like your code generate all 0 submission when submission, not commit.
", "votes": null, "replies": [ { "id": 697581, "author_name": "h030162", "author_url": "", "post_date": "12/18/2019 05:40:01", "content": "Thanks for you reply. But it‘s not like what you say.\nWhen with \"dfmerged.to_csv('submission.csv', index=False)\" run\n My run log: \n id label\n0 0 3\n1 1 0\n2 2 2\n3 3 6\n4 4 7\n5.6s2../input/submerage1110/submerage1110.csv\n===03:31:08===*********end*********
\n\nsubmission.csv in output is:\n1 id label\n2 0 3\n3 1 0\n4 2 2\n5 3 6\n6 4 7\n7 5 7\n8 6 1\n9 7 9\n10 8 3\n11 9 4\n12 10 8\n13 11 8\n14 12 1\n15 13 7\n16 14 8\n17 15 1\n18 16 5\n19 17 1\n20 18 5\n21 19 9\n22 20 3\n23 21 7\n24 22 6\n25 23 0\n26 24 2\n27 25 0\n28 26 8\n29 27 7\n30 28 0\n31 29 0\n32 30 8
\n\nSo not all label is 0.\nThank you again anyway.
", "votes": null, "replies": [] }, { "id": 697588, "author_name": "lhy1997", "author_url": "", "post_date": "12/18/2019 05:53:52", "content": "I'm sorry, maybe I didn't explain it clearly.
\n\nI remember I had a all zero test for the leader-board, I check it private score just now, the score is:\n\n so, there are no doubt that your submission of private version is all zero.\nthe result you attached is public version submission, not the private version submission.
Maybe you remember that the kaggle need run our code two times, when you press the 'commit' and 'submission', your code will run in different test data. the commit is public 5000 images, and the submission is private 5000 images.
\n\nSo I think there maybe some bugs in your code that result in all zero when press 'submission'.
", "votes": null, "replies": [] }, { "id": 697592, "author_name": "h030162", "author_url": "", "post_date": "12/18/2019 05:57:03", "content": "Thank you for your patience reply.
", "votes": null, "replies": [] } ] }, { "id": 697589, "author_name": "h030162", "author_url": "", "post_date": "12/18/2019 05:54:29", "content": "I know!\nI only submit a csv file but no more.\na. run train and inference in the kernel.\nb.up load offline model and run inference in the kernel
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "697501": "My public highest score is 0.9910. but all my private scores are 0.10260? I don't know what happened as a newer. Who anyone can tell me the reason? thanks very much!\n\nprivate public\n0.10260 0.98200\n0.10260 0.98660\n0.10260 0.98600\n0.10260 0.98400\n0.10260 0.98200\n0.10260 0.99040\n0.10260 0.99100\n\nsubmit code is here:\n\nfilename = 'submerage1110'\nparentpath = '../input/' + filename\ndf_sub = pd.read_csv(parentpath + '/' + filename + '.csv')\nsample_submission = pd.read_csv('../input/Kannada-MNIST/sample_submission.csv')\ndf_merged = pd.merge(sample_submission, df_sub, on='id', how='left')\ndf_merged.fillna(0, inplace=True)\ndf_merged['label'] = df_merged['label_y']\ndf_merged = df_merged[['id', 'label']]\ndf_merged['label'] = df_merged['label'].astype('int64')\ndf_merged.to_csv('submission.csv', index=False)\nctime = time.strftime(\"%H:%M:%S\")\nprint(parentpath + '/' + filename + '.csv')\nprint('==={}==='.format(ctime) + '*' * 20 + 'end' + '*' * 20)", "697545": "It seems like your code generate all 0 submission when submission, not commit.", "697581": "Thanks for you reply. But it‘s not like what you say.\nWhen with \"dfmerged.to_csv('submission.csv', index=False)\" run\n My run log: \n id label\n0 0 3\n1 1 0\n2 2 2\n3 3 6\n4 4 7\n5.6s2../input/submerage1110/submerage1110.csv\n===03:31:08===********************end********************\n\nsubmission.csv in output is:\n1\tid\tlabel\n2\t0\t3\n3\t1\t0\n4\t2\t2\n5\t3\t6\n6\t4\t7\n7\t5\t7\n8\t6\t1\n9\t7\t9\n10\t8\t3\n11\t9\t4\n12\t10\t8\n13\t11\t8\n14\t12\t1\n15\t13\t7\n16\t14\t8\n17\t15\t1\n18\t16\t5\n19\t17\t1\n20\t18\t5\n21\t19\t9\n22\t20\t3\n23\t21\t7\n24\t22\t6\n25\t23\t0\n26\t24\t2\n27\t25\t0\n28\t26\t8\n29\t27\t7\n30\t28\t0\n31\t29\t0\n32\t30\t8\n\nSo not all label is 0.\nThank you again anyway.", "697588": "I'm sorry, maybe I didn't explain it clearly. \n\nI remember I had a all zero test for the leader-board, I check it private score just now, the score is:\n\n so, there are no doubt that your submission of private version is all zero.\nthe result you attached is public version submission, not the private version submission.\n\nMaybe you remember that the kaggle need run our code two times, when you press the 'commit' and 'submission', your code will run in different test data. the commit is public 5000 images, and the submission is private 5000 images.\n\nSo I think there maybe some bugs in your code that result in all zero when press 'submission'.", "697589": "I know!\nI only submit a csv file but no more.\na. run train and inference in the kernel.\nb.up load offline model and run inference in the kernel", "697592": "Thank you for your patience reply." }, "source": "meta" }