![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F2033774%2F03426a6f0ef1ab2adce89e93af016424%2F123.PNG?generation=1565843496758812&alt=media\")
If some people use code like me to install NVIDIA apex in kernel, since the version up of PyTorch last night, NVIDIA Apex will fail to install. \nhttps://www.kaggle.com/chanhu/eye-efficientnet-pytorch-lb-0-777
\n\nif I find someway, I will update this discussion.
\n\n[Updated]\n1. reinstall torch 1.1.0: \nCommand::\n !pip install torch==1.1.0\n2. and then use command below to install NVIDIA apex.\nCommand:\n! pip install -v --no-cache-dir --global-option=\"--cpp_ext\" --global-option=\"--cuda_ext\" ../input/nvidiaapex/repository/NVIDIA-apex-39e153a
", "messages": [ { "id": "599464", "postDate": "08/15/2019 02:20:35", "content": "If some people use code like me to install NVIDIA apex in kernel, since the version up of PyTorch last night, NVIDIA Apex will fail to install. \nhttps://www.kaggle.com/chanhu/eye-efficientnet-pytorch-lb-0-777
\n\nif I find someway, I will update this discussion.
\n\n[Updated]\n1. reinstall torch 1.1.0: \nCommand::\n !pip install torch==1.1.0\n2. and then use command below to install NVIDIA apex.\nCommand:\n! pip install -v --no-cache-dir --global-option=\"--cpp_ext\" --global-option=\"--cuda_ext\" ../input/nvidiaapex/repository/NVIDIA-apex-39e153a
", "rawMarkdown": "If some people use code like me to install NVIDIA apex in kernel, since the version up of PyTorch last night, NVIDIA Apex will fail to install. \nhttps://www.kaggle.com/chanhu/eye-efficientnet-pytorch-lb-0-777\n\nif I find someway, I will update this discussion. \n\n[Updated]\n1. reinstall torch 1.1.0: \nCommand::\n !pip install torch==1.1.0\n2. and then use command below to install NVIDIA apex.\nCommand:\n! pip install -v --no-cache-dir --global-option=\"--cpp_ext\" --global-option=\"--cuda_ext\" ../input/nvidiaapex/repository/NVIDIA-apex-39e153a", "votes": null }, { "id": "599467", "postDate": "08/15/2019 02:26:26", "content": "GG
", "rawMarkdown": "GG", "votes": null }, { "id": "599516", "postDate": "08/15/2019 04:17:31", "content": "With that !pip install torch==1.1.0 command I got:\n\"Failed to establish a new connection: [Errno -3] Temporary failure in name resolution',)': /simple/torch/\"\nI searched it and it's a dns problem, this is likely specific to people from countries like China where the internet is often blocked. I had tried to turn off my vpn but the error is still there. Will appreciate it if anyone can help.
", "rawMarkdown": "With that !pip install torch==1.1.0 command I got:\n\"Failed to establish a new connection: [Errno -3] Temporary failure in name resolution',)': /simple/torch/\"\nI searched it and it's a dns problem, this is likely specific to people from countries like China where the internet is often blocked. I had tried to turn off my vpn but the error is still there. Will appreciate it if anyone can help.", "votes": null }, { "id": "599525", "postDate": "08/15/2019 04:31:39", "content": "I don't think it is your DNS problem. Because you run a kernel in Virtual Machine at Data Center. \nDid you turn on you Internet access in Kernel?
\n\n
No, it's not because of china , it's because of you didn't open your internet in kernel
", "rawMarkdown": "No, it's not because of china , it's because of you didn't open your internet in kernel", "votes": null }, { "id": "599552", "postDate": "08/15/2019 05:42:34", "content": "Thanks, it now works with internet on. And sorry for my ignorance.
", "rawMarkdown": "Thanks, it now works with internet on. And sorry for my ignorance.", "votes": null }, { "id": "599668", "postDate": "08/15/2019 08:35:05", "content": "Thank you for paying attention to this problem. But I still can not install apex. I get 125 different links and writes\n \"Given no hashes to check 125 links for project 'pip': discarding no candidates\"
\n\nIf you help me I will be very grateful, thank you
", "rawMarkdown": "Thank you for paying attention to this problem. But I still can not install apex. I get 125 different links and writes\n \"Given no hashes to check 125 links for project 'pip': discarding no candidates\"\n\nIf you help me I will be very grateful, thank you", "votes": null }, { "id": "599692", "postDate": "08/15/2019 09:05:12", "content": "I revised and recommitted my kernel https://www.kaggle.com/chanhu/eye-efficientnet-pytorch-lb-0-777, you can use same code. \nDon't worry about \"Given no hashes to check 125 links for project 'pip': discarding no candidates\". \nUsing \"from apex import amp\" to check if apex had been installed.
", "rawMarkdown": "I revised and recommitted my kernel https://www.kaggle.com/chanhu/eye-efficientnet-pytorch-lb-0-777, you can use same code. \nDon't worry about \"Given no hashes to check 125 links for project 'pip': discarding no candidates\". \nUsing \"from apex import amp\" to check if apex had been installed.", "votes": null }, { "id": "599700", "postDate": "08/15/2019 09:14:39", "content": "Thanks, this is really helpful. I tried about 100 method to install apex but all failed earlier.
", "rawMarkdown": "Thanks, this is really helpful. I tried about 100 method to install apex but all failed earlier.", "votes": null }, { "id": "599703", "postDate": "08/15/2019 09:20:03", "content": "Thank you very much, sorry for my stupidity
", "rawMarkdown": "Thank you very much, sorry for my stupidity", "votes": null }, { "id": "599705", "postDate": "08/15/2019 09:21:54", "content": "Still, kaggle is unusable in China without a VPN.
", "rawMarkdown": "Still, kaggle is unusable in China without a VPN.", "votes": null }, { "id": "599753", "postDate": "08/15/2019 10:12:12", "content": "No, I can always access kaggle without a VPN , downloading dataset with 10MB/second , maybe you need a tool to change your DNS . Don't blame my China
", "rawMarkdown": "No, I can always access kaggle without a VPN , downloading dataset with 10MB/second , maybe you need a tool to change your DNS . Don't blame my China", "votes": null }, { "id": "599792", "postDate": "08/15/2019 11:06:01", "content": "The problem is with one of the CPP files (they are missing ;)\nThe issue was fixed in their GitHub repostry \nI managed to overcome this with the following lines
\n\n\n! git clone https://github.com/NVIDIA/apex\n!pip install -v --no-cache-dir --global-option=\"--cpp_ext\" --global-option=\"--cuda_ext\" ./apex\n\nI have spent few hours this morning on this
Update - \nThis option runs when you run the notebook, but my commit failed with this error :\nOutput path '/apex/.git/logs/refs/remotes/origin/HEAD' contains too many nested subdirectories (max 6)\nI will update if I find a solution
", "rawMarkdown": "The problem is with one of the CPP files (they are missing ;)\nThe issue was fixed in their GitHub repostry \nI managed to overcome this with the following lines \n\n```\n! git clone https://github.com/NVIDIA/apex\n!pip install -v --no-cache-dir --global-option=\"--cpp_ext\" --global-option=\"--cuda_ext\" ./apex\n```\nI have spent few hours this morning on this \n\n\nUpdate - \nThis option runs when you run the notebook, but my commit failed with this error :\nOutput path '/apex/.git/logs/refs/remotes/origin/HEAD' contains too many nested subdirectories (max 6)\nI will update if I find a solution", "votes": null }, { "id": "599911", "postDate": "08/15/2019 12:46:48", "content": "Thank you!
", "rawMarkdown": "Thank you!", "votes": null }, { "id": "600016", "postDate": "08/15/2019 14:36:05", "content": "Thanks a lot!
", "rawMarkdown": "Thanks a lot!", "votes": null }, { "id": "600118", "postDate": "08/15/2019 17:08:24", "content": "How about uploading datasets? I cannot upload datasets without my VPN on in China.
", "rawMarkdown": "How about uploading datasets? I cannot upload datasets without my VPN on in China.", "votes": null }, { "id": "600331", "postDate": "08/16/2019 01:56:30", "content": "Not always. Maybe now is good. The situation was different sereval month ago. It took years to load one webpage.
", "rawMarkdown": "Not always. Maybe now is good. The situation was different sereval month ago. It took years to load one webpage.", "votes": null }, { "id": "601129", "postDate": "08/17/2019 04:05:31", "content": "In edit mode it's fine, but when commiting, I always get this error:\n\"Output path '/apex/.git/logs/refs/remotes/origin/HEAD' contains too many nested subdirectories (max 6)\".\nAnyone has the same problem?
", "rawMarkdown": "In edit mode it's fine, but when commiting, I always get this error:\n\"Output path '/apex/.git/logs/refs/remotes/origin/HEAD' contains too many nested subdirectories (max 6)\".\nAnyone has the same problem?", "votes": null }, { "id": "601848", "postDate": "08/18/2019 08:31:58", "content": "hi chunhu\nHow about the submission kernel..\nDo we need installation there ?
", "rawMarkdown": "hi chunhu\nHow about the submission kernel..\nDo we need installation there ?", "votes": null }, { "id": "642541", "postDate": "10/06/2019 09:02:33", "content": "This worked for me, thanks!
", "rawMarkdown": "This worked for me, thanks!", "votes": null } ], "comments": [ { "id": 599467, "author_name": "xujingzhao", "author_url": "", "post_date": "08/15/2019 02:26:26", "content": "GG
", "votes": null, "replies": [] }, { "id": 599516, "author_name": "homoalways", "author_url": "", "post_date": "08/15/2019 04:17:31", "content": "With that !pip install torch==1.1.0 command I got:\n\"Failed to establish a new connection: [Errno -3] Temporary failure in name resolution',)': /simple/torch/\"\nI searched it and it's a dns problem, this is likely specific to people from countries like China where the internet is often blocked. I had tried to turn off my vpn but the error is still there. Will appreciate it if anyone can help.
", "votes": null, "replies": [ { "id": 599525, "author_name": "chanhu", "author_url": "", "post_date": "08/15/2019 04:31:39", "content": "I don't think it is your DNS problem. Because you run a kernel in Virtual Machine at Data Center. \nDid you turn on you Internet access in Kernel?
\n\n
No, it's not because of china , it's because of you didn't open your internet in kernel
", "votes": null, "replies": [] }, { "id": 599552, "author_name": "homoalways", "author_url": "", "post_date": "08/15/2019 05:42:34", "content": "Thanks, it now works with internet on. And sorry for my ignorance.
", "votes": null, "replies": [] }, { "id": 599705, "author_name": "ys19931006", "author_url": "", "post_date": "08/15/2019 09:21:54", "content": "Still, kaggle is unusable in China without a VPN.
", "votes": null, "replies": [] }, { "id": 599753, "author_name": "xujingzhao", "author_url": "", "post_date": "08/15/2019 10:12:12", "content": "No, I can always access kaggle without a VPN , downloading dataset with 10MB/second , maybe you need a tool to change your DNS . Don't blame my China
", "votes": null, "replies": [] }, { "id": 600118, "author_name": "xiaojiu1414", "author_url": "", "post_date": "08/15/2019 17:08:24", "content": "How about uploading datasets? I cannot upload datasets without my VPN on in China.
", "votes": null, "replies": [] }, { "id": 600331, "author_name": "ys19931006", "author_url": "", "post_date": "08/16/2019 01:56:30", "content": "Not always. Maybe now is good. The situation was different sereval month ago. It took years to load one webpage.
", "votes": null, "replies": [] } ] }, { "id": 599668, "author_name": "boykoaa", "author_url": "", "post_date": "08/15/2019 08:35:05", "content": "Thank you for paying attention to this problem. But I still can not install apex. I get 125 different links and writes\n \"Given no hashes to check 125 links for project 'pip': discarding no candidates\"
\n\nIf you help me I will be very grateful, thank you
", "votes": null, "replies": [ { "id": 599692, "author_name": "chanhu", "author_url": "", "post_date": "08/15/2019 09:05:12", "content": "I revised and recommitted my kernel https://www.kaggle.com/chanhu/eye-efficientnet-pytorch-lb-0-777, you can use same code. \nDon't worry about \"Given no hashes to check 125 links for project 'pip': discarding no candidates\". \nUsing \"from apex import amp\" to check if apex had been installed.
", "votes": null, "replies": [] }, { "id": 599703, "author_name": "boykoaa", "author_url": "", "post_date": "08/15/2019 09:20:03", "content": "Thank you very much, sorry for my stupidity
", "votes": null, "replies": [] }, { "id": 601848, "author_name": "jaideepvalani", "author_url": "", "post_date": "08/18/2019 08:31:58", "content": "hi chunhu\nHow about the submission kernel..\nDo we need installation there ?
", "votes": null, "replies": [] }, { "id": 642541, "author_name": "pai1995", "author_url": "", "post_date": "10/06/2019 09:02:33", "content": "This worked for me, thanks!
", "votes": null, "replies": [] } ] }, { "id": 599700, "author_name": "ys19931006", "author_url": "", "post_date": "08/15/2019 09:14:39", "content": "Thanks, this is really helpful. I tried about 100 method to install apex but all failed earlier.
", "votes": null, "replies": [] }, { "id": 599792, "author_name": "omershect", "author_url": "", "post_date": "08/15/2019 11:06:01", "content": "The problem is with one of the CPP files (they are missing ;)\nThe issue was fixed in their GitHub repostry \nI managed to overcome this with the following lines
\n\n\n! git clone https://github.com/NVIDIA/apex\n!pip install -v --no-cache-dir --global-option=\"--cpp_ext\" --global-option=\"--cuda_ext\" ./apex\n\nI have spent few hours this morning on this
Update - \nThis option runs when you run the notebook, but my commit failed with this error :\nOutput path '/apex/.git/logs/refs/remotes/origin/HEAD' contains too many nested subdirectories (max 6)\nI will update if I find a solution
", "votes": null, "replies": [ { "id": 599911, "author_name": "chanhu", "author_url": "", "post_date": "08/15/2019 12:46:48", "content": "Thank you!
", "votes": null, "replies": [] }, { "id": 600016, "author_name": "bamps53", "author_url": "", "post_date": "08/15/2019 14:36:05", "content": "Thanks a lot!
", "votes": null, "replies": [] }, { "id": 601129, "author_name": "homoalways", "author_url": "", "post_date": "08/17/2019 04:05:31", "content": "In edit mode it's fine, but when commiting, I always get this error:\n\"Output path '/apex/.git/logs/refs/remotes/origin/HEAD' contains too many nested subdirectories (max 6)\".\nAnyone has the same problem?
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "599464": "If some people use code like me to install NVIDIA apex in kernel, since the version up of PyTorch last night, NVIDIA Apex will fail to install. \nhttps://www.kaggle.com/chanhu/eye-efficientnet-pytorch-lb-0-777\n\nif I find someway, I will update this discussion. \n\n[Updated]\n1. reinstall torch 1.1.0: \nCommand::\n !pip install torch==1.1.0\n2. and then use command below to install NVIDIA apex.\nCommand:\n! pip install -v --no-cache-dir --global-option=\"--cpp_ext\" --global-option=\"--cuda_ext\" ../input/nvidiaapex/repository/NVIDIA-apex-39e153a", "599467": "GG", "599516": "With that !pip install torch==1.1.0 command I got:\n\"Failed to establish a new connection: [Errno -3] Temporary failure in name resolution',)': /simple/torch/\"\nI searched it and it's a dns problem, this is likely specific to people from countries like China where the internet is often blocked. I had tried to turn off my vpn but the error is still there. Will appreciate it if anyone can help.", "599525": "I don't think it is your DNS problem. Because you run a kernel in Virtual Machine at Data Center. \nDid you turn on you Internet access in Kernel?\n\n", "599527": "No, it's not because of china , it's because of you didn't open your internet in kernel", "599552": "Thanks, it now works with internet on. And sorry for my ignorance.", "599668": "Thank you for paying attention to this problem. But I still can not install apex. I get 125 different links and writes\n \"Given no hashes to check 125 links for project 'pip': discarding no candidates\"\n\nIf you help me I will be very grateful, thank you", "599692": "I revised and recommitted my kernel https://www.kaggle.com/chanhu/eye-efficientnet-pytorch-lb-0-777, you can use same code. \nDon't worry about \"Given no hashes to check 125 links for project 'pip': discarding no candidates\". \nUsing \"from apex import amp\" to check if apex had been installed.", "599700": "Thanks, this is really helpful. I tried about 100 method to install apex but all failed earlier.", "599703": "Thank you very much, sorry for my stupidity", "599705": "Still, kaggle is unusable in China without a VPN.", "599753": "No, I can always access kaggle without a VPN , downloading dataset with 10MB/second , maybe you need a tool to change your DNS . Don't blame my China", "599792": "The problem is with one of the CPP files (they are missing ;)\nThe issue was fixed in their GitHub repostry \nI managed to overcome this with the following lines \n\n```\n! git clone https://github.com/NVIDIA/apex\n!pip install -v --no-cache-dir --global-option=\"--cpp_ext\" --global-option=\"--cuda_ext\" ./apex\n```\nI have spent few hours this morning on this \n\n\nUpdate - \nThis option runs when you run the notebook, but my commit failed with this error :\nOutput path '/apex/.git/logs/refs/remotes/origin/HEAD' contains too many nested subdirectories (max 6)\nI will update if I find a solution", "599911": "Thank you!", "600016": "Thanks a lot!", "600118": "How about uploading datasets? I cannot upload datasets without my VPN on in China.", "600331": "Not always. Maybe now is good. The situation was different sereval month ago. It took years to load one webpage.", "601129": "In edit mode it's fine, but when commiting, I always get this error:\n\"Output path '/apex/.git/logs/refs/remotes/origin/HEAD' contains too many nested subdirectories (max 6)\".\nAnyone has the same problem?", "601848": "hi chunhu\nHow about the submission kernel..\nDo we need installation there ?", "642541": "This worked for me, thanks!" }, "source": "meta" }