Check this question in my notebook π
\nhttps://www.kaggle.com/blueboy97/install-detectron2-offline
\nAND
\nThe datasets link: https://www.kaggle.com/blueboy97/detectron2-whl-file
Check this question in my notebook π
\nhttps://www.kaggle.com/blueboy97/install-detectron2-offline
\nAND
\nThe datasets link: https://www.kaggle.com/blueboy97/detectron2-whl-file
Thanks for this!!
\nBut I think you have made the detectron a private dataset right?
\nYou can make it public so all of us can use itπ
π
U can download it from my notebook INPUT.
", "rawMarkdown": "U can download it from my notebook INPUT.", "votes": null }, { "id": "1337023", "postDate": "06/05/2021 11:05:40", "content": "Yeah, it's showing Private there, You have to make it public first, else, except you, we cannot see it.
", "rawMarkdown": "Yeah, it's showing Private there, You have to make it public first, else, except you, we cannot see it.", "votes": null }, { "id": "1337040", "postDate": "06/05/2021 11:19:30", "content": "I don't know it. Sorry for that. AND now I update my TOPIC(The Dataset LINK)->PUBLIC
", "rawMarkdown": "I don't know it. Sorry for that. AND now I update my TOPIC(The Dataset LINK)->PUBLIC", "votes": null }, { "id": "1337041", "postDate": "06/05/2021 11:22:09", "content": "Thank you!!
", "rawMarkdown": "Thank you!!", "votes": null }, { "id": "1345165", "postDate": "06/11/2021 11:14:56", "content": "When I use detectrron2 in offline mode the import doesn't work, can someone tell why?
\nfrom detectron2.structures import BoxMode\n\n---------------------------------------------------------------------------\nImportError Traceback (most recent call last)\n<ipython-input-5-f4fa7c3ba98d> in <module>\n----> 1 from detectron2.structures import BoxMode\n\n/opt/conda/lib/python3.7/site-packages/detectron2/structures/__init__.py in <module>\n 4 \n 5 from .instances import Instances\n----> 6 from .keypoints import Keypoints, heatmaps_to_keypoints\n 7 from .masks import BitMasks, PolygonMasks, rasterize_polygons_within_box, polygons_to_bitmask\n 8 from .rotated_boxes import RotatedBoxes\n\n/opt/conda/lib/python3.7/site-packages/detectron2/structures/keypoints.py in <module>\n 4 import torch\n 5 \n----> 6 from detectron2.layers import interpolate\n 7 \n 8 \n\n/opt/conda/lib/python3.7/site-packages/detectron2/layers/__init__.py in <module>\n 1 # Copyright (c) Facebook, Inc. and its affiliates.\n 2 from .batch_norm import FrozenBatchNorm2d, get_norm, NaiveSyncBatchNorm\n----> 3 from .deform_conv import DeformConv, ModulatedDeformConv\n 4 from .mask_ops import paste_masks_in_image\n 5 from .nms import batched_nms, batched_nms_rotated, nms, nms_rotated\n\n/opt/conda/lib/python3.7/site-packages/detectron2/layers/deform_conv.py in <module>\n 8 from torch.nn.modules.utils import _pair\n 9 \n---> 10 from detectron2 import _C\n 11 \n 12 from .wrappers import _NewEmptyTensorOp\n\nImportError: libcudart.so.11.0: cannot open shared object file: No such file or directory\nHi @mrinath , I used Detectron2 offline, and worked well.
\nsee my kernel
\nSIIM COVID-19 Detectron2 Inferance
\nversion23
Hi everyone, I used Detectron2 offline in this competition, and worked well.
\nSee my kernel
\nSIIM COVID-19 Detectron2 Inferance
\nversion23
Finally!! I was waiting for it, hope you get a good score with detection 2.
\nI think I got why it was not working, You have to set GPU as accelerator, Thanks
According to the error reminding, U haven't add the detecron2 wheel files in correct way, Or the wheel file dir path needs to modify correctly.
", "rawMarkdown": "According to the error reminding, U haven't add the detecron2 wheel files in correct way, Or the wheel file dir path needs to modify correctly.", "votes": null }, { "id": "1346190", "postDate": "06/12/2021 07:26:57", "content": "Wow, it looks good.But it seems many redundant boxes in the opacity prediction.
", "rawMarkdown": "Wow, it looks good.But it seems many redundant boxes in the opacity prediction.", "votes": null }, { "id": "1346221", "postDate": "06/12/2021 08:02:27", "content": "@blueboy97 Absolutely,We need some effort to solve this issue
", "rawMarkdown": "blueboy97 Absolutely,We need some effort to solve this issue", "votes": null }, { "id": "1346234", "postDate": "06/12/2021 08:09:27", "content": "@mrinath, I am still working on the inference kernel, the problem is not using GPU but I think the problem is the way to calculate the scores.
", "rawMarkdown": "mrinath, I am still working on the inference kernel, the problem is not using GPU but I think the problem is the way to calculate the scores.", "votes": null } ], "comments": [ { "id": 1335279, "author_name": "mrinath", "author_url": "", "post_date": "06/04/2021 06:05:53", "content": "Thanks for this!!
\nBut I think you have made the detectron a private dataset right?
\nYou can make it public so all of us can use itπ
π
U can download it from my notebook INPUT.
", "votes": null, "replies": [] }, { "id": 1337023, "author_name": "mrinath", "author_url": "", "post_date": "06/05/2021 11:05:40", "content": "Yeah, it's showing Private there, You have to make it public first, else, except you, we cannot see it.
", "votes": null, "replies": [] }, { "id": 1337040, "author_name": "blueboy97", "author_url": "", "post_date": "06/05/2021 11:19:30", "content": "I don't know it. Sorry for that. AND now I update my TOPIC(The Dataset LINK)->PUBLIC
", "votes": null, "replies": [] }, { "id": 1337041, "author_name": "mrinath", "author_url": "", "post_date": "06/05/2021 11:22:09", "content": "Thank you!!
", "votes": null, "replies": [] } ] }, { "id": 1345165, "author_name": "mrinath", "author_url": "", "post_date": "06/11/2021 11:14:56", "content": "When I use detectrron2 in offline mode the import doesn't work, can someone tell why?
\nfrom detectron2.structures import BoxMode\n\n---------------------------------------------------------------------------\nImportError Traceback (most recent call last)\n<ipython-input-5-f4fa7c3ba98d> in <module>\n----> 1 from detectron2.structures import BoxMode\n\n/opt/conda/lib/python3.7/site-packages/detectron2/structures/__init__.py in <module>\n 4 \n 5 from .instances import Instances\n----> 6 from .keypoints import Keypoints, heatmaps_to_keypoints\n 7 from .masks import BitMasks, PolygonMasks, rasterize_polygons_within_box, polygons_to_bitmask\n 8 from .rotated_boxes import RotatedBoxes\n\n/opt/conda/lib/python3.7/site-packages/detectron2/structures/keypoints.py in <module>\n 4 import torch\n 5 \n----> 6 from detectron2.layers import interpolate\n 7 \n 8 \n\n/opt/conda/lib/python3.7/site-packages/detectron2/layers/__init__.py in <module>\n 1 # Copyright (c) Facebook, Inc. and its affiliates.\n 2 from .batch_norm import FrozenBatchNorm2d, get_norm, NaiveSyncBatchNorm\n----> 3 from .deform_conv import DeformConv, ModulatedDeformConv\n 4 from .mask_ops import paste_masks_in_image\n 5 from .nms import batched_nms, batched_nms_rotated, nms, nms_rotated\n\n/opt/conda/lib/python3.7/site-packages/detectron2/layers/deform_conv.py in <module>\n 8 from torch.nn.modules.utils import _pair\n 9 \n---> 10 from detectron2 import _C\n 11 \n 12 from .wrappers import _NewEmptyTensorOp\n\nImportError: libcudart.so.11.0: cannot open shared object file: No such file or directory\nHi @mrinath , I used Detectron2 offline, and worked well.
\nsee my kernel
\nSIIM COVID-19 Detectron2 Inferance
\nversion23
Finally!! I was waiting for it, hope you get a good score with detection 2.
\nI think I got why it was not working, You have to set GPU as accelerator, Thanks
According to the error reminding, U haven't add the detecron2 wheel files in correct way, Or the wheel file dir path needs to modify correctly.
", "votes": null, "replies": [] }, { "id": 1346234, "author_name": "ammarnassanalhajali", "author_url": "", "post_date": "06/12/2021 08:09:27", "content": "@mrinath, I am still working on the inference kernel, the problem is not using GPU but I think the problem is the way to calculate the scores.
", "votes": null, "replies": [] } ] }, { "id": 1346065, "author_name": "ammarnassanalhajali", "author_url": "", "post_date": "06/12/2021 05:14:20", "content": "Hi everyone, I used Detectron2 offline in this competition, and worked well.
\nSee my kernel
\nSIIM COVID-19 Detectron2 Inferance
\nversion23
Wow, it looks good.But it seems many redundant boxes in the opacity prediction.
", "votes": null, "replies": [] }, { "id": 1346221, "author_name": "ammarnassanalhajali", "author_url": "", "post_date": "06/12/2021 08:02:27", "content": "@blueboy97 Absolutely,We need some effort to solve this issue
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "1334582": "Check this question in my notebook π\n[https://www.kaggle.com/blueboy97/install-detectron2-offline](https://www.kaggle.com/blueboy97/install-detectron2-offline)\nAND\nThe datasets link: [https://www.kaggle.com/blueboy97/detectron2-whl-file](https://www.kaggle.com/blueboy97/detectron2-whl-file)", "1335279": "Thanks for this!!\nBut I think you have made the detectron a private dataset right?\nYou can make it public so all of us can use itπ π ", "1336954": "U can download it from my notebook INPUT.", "1337023": "Yeah, it's showing Private there, You have to make it public first, else, except you, we cannot see it.", "1337040": "I don't know it. Sorry for that. AND now I update my TOPIC(The Dataset LINK)->PUBLIC", "1337041": "Thank you!!", "1345165": "When I use detectrron2 in offline mode the import doesn't work, can someone tell why?\n```\nfrom detectron2.structures import BoxMode\n\n---------------------------------------------------------------------------\nImportError Traceback (most recent call last)\n