{ "id": 230541, "title": "Can we use the pre-training model？", "url": "/competitions/plant-pathology-2021-fgvc8/discussion/230541", "author_name": "", "post_date": "2021-04-04T11:16:11.169390300Z", "votes": 3, "comment_count": 4, "views": 0, "content": "

Use models that have been trained

", "messages": [ { "id": "1262487", "postDate": "04/04/2021 11:16:11", "content": "

Use models that have been trained

", "rawMarkdown": "Use models that have been trained", "votes": null }, { "id": "1262915", "postDate": "04/04/2021 21:44:00", "content": "

Hi @dlidli Please read in the competition rules here that 'Pre-trained models are allowed'. You might want to check these kernels to see a way how pretrained models can be used: kernel1, kernel2.

", "rawMarkdown": "Hi @dlidli Please read in the competition rules [here](https://www.kaggle.com/c/plant-pathology-2021-fgvc8/overview/code-requirements) that 'Pre-trained models are allowed'. You might want to check these kernels to see a way how pretrained models can be used: [kernel1](https://www.kaggle.com/pegasos/plant2021-pytorch-lightning-starter-training), [kernel2](https://www.kaggle.com/pegasos/plant2021-pytorch-lightning-starter-inference).", "votes": null }, { "id": "1263198", "postDate": "04/05/2021 07:02:42", "content": "

@buinyi Thank you for your answer

", "rawMarkdown": "buinyi Thank you for your answer", "votes": null }, { "id": "1266543", "postDate": "04/07/2021 21:32:35", "content": "

If you select 'Add Data' and search for the model you want, then you can add the weights (if the model is on kaggle) to your data set, then use that as the path to the weights, like

\n

from tensorflow.keras.applications.vgg16 import VGG16
\nvgg16 = VGG16(weights = '/kaggle/input/vgg16/vgg16_weights_tf_dim_ordering_tf_kernels_notop.h5', include_top=False)

\n

or you could upload the model weights after downloading them elsewhere, since internet is not allowed in the code.

", "rawMarkdown": "If you select 'Add Data' and search for the model you want, then you can add the weights (if the model is on kaggle) to your data set, then use that as the path to the weights, like \n\nfrom tensorflow.keras.applications.vgg16 import VGG16\nvgg16 = VGG16(weights = '/kaggle/input/vgg16/vgg16_weights_tf_dim_ordering_tf_kernels_notop.h5', include_top=False)\n\nor you could upload the model weights after downloading them elsewhere, since internet is not allowed in the code.", "votes": null }, { "id": "1266648", "postDate": "04/08/2021 00:47:08", "content": "

@gb00000 Thank

", "rawMarkdown": "gb00000 Thank", "votes": null } ], "comments": [ { "id": 1262915, "author_name": "buinyi", "author_url": "", "post_date": "04/04/2021 21:44:00", "content": "

Hi @dlidli Please read in the competition rules here that 'Pre-trained models are allowed'. You might want to check these kernels to see a way how pretrained models can be used: kernel1, kernel2.

", "votes": null, "replies": [] }, { "id": 1263198, "author_name": "dlidli", "author_url": "", "post_date": "04/05/2021 07:02:42", "content": "

@buinyi Thank you for your answer

", "votes": null, "replies": [] }, { "id": 1266543, "author_name": "gb00000", "author_url": "", "post_date": "04/07/2021 21:32:35", "content": "

If you select 'Add Data' and search for the model you want, then you can add the weights (if the model is on kaggle) to your data set, then use that as the path to the weights, like

\n

from tensorflow.keras.applications.vgg16 import VGG16
\nvgg16 = VGG16(weights = '/kaggle/input/vgg16/vgg16_weights_tf_dim_ordering_tf_kernels_notop.h5', include_top=False)

\n

or you could upload the model weights after downloading them elsewhere, since internet is not allowed in the code.

", "votes": null, "replies": [] }, { "id": 1266648, "author_name": "dlidli", "author_url": "", "post_date": "04/08/2021 00:47:08", "content": "

@gb00000 Thank

", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "1262487": "Use models that have been trained", "1262915": "Hi @dlidli Please read in the competition rules [here](https://www.kaggle.com/c/plant-pathology-2021-fgvc8/overview/code-requirements) that 'Pre-trained models are allowed'. You might want to check these kernels to see a way how pretrained models can be used: [kernel1](https://www.kaggle.com/pegasos/plant2021-pytorch-lightning-starter-training), [kernel2](https://www.kaggle.com/pegasos/plant2021-pytorch-lightning-starter-inference).", "1263198": "buinyi Thank you for your answer", "1266543": "If you select 'Add Data' and search for the model you want, then you can add the weights (if the model is on kaggle) to your data set, then use that as the path to the weights, like \n\nfrom tensorflow.keras.applications.vgg16 import VGG16\nvgg16 = VGG16(weights = '/kaggle/input/vgg16/vgg16_weights_tf_dim_ordering_tf_kernels_notop.h5', include_top=False)\n\nor you could upload the model weights after downloading them elsewhere, since internet is not allowed in the code.", "1266648": "gb00000 Thank" }, "source": "meta" }