I'm working on this competition on my local machine but unable to run from kaggle_datasets import KaggleDatasets as I'm getting ModuleNotFoundError: No module named 'kaggle.datasets'.
I tried pip install kaggle but that didn't help.
I'm working on this competition on my local machine but unable to run from kaggle_datasets import KaggleDatasets as I'm getting ModuleNotFoundError: No module named 'kaggle.datasets'.
I tried pip install kaggle but that didn't help.
If you are working on your local machine, download the dataset to your local machine. You will not be able to use a TPU though.
\n\nIf you want to work with a TPU-enabled Notebook VM in GCP (here is how to provision one: bit.ly/keras-tpu-tf21), then I recommend you download the dataset and upload it to a GCS bucket. TPUs can only read data from GCS.
\n\nThere is a way of working directly with the GCS bucket Kaggle uses for the dataset: call KaggleDataset().get_gcs_path() on Kaggle, record the GCS path it gives you. You can then use it from anywhere. Two caveats: the Kaggle API gets you a bucket in the same region as your Kaggle TPU which might not be the region you want. Also, The Kaggle API treats this GCS bucket as a cache. It can erase and recreate it under a different name at any time. It should stay good for a couple of days though.
", "rawMarkdown": "If you are working on your local machine, download the dataset to your local machine. You will not be able to use a TPU though.\n\nIf you want to work with a TPU-enabled Notebook VM in GCP (here is how to provision one: [bit.ly/keras-tpu-tf21](bit.ly/keras-tpu-tf21)), then I recommend you download the dataset and upload it to a GCS bucket. TPUs can only read data from GCS.\n\nThere is a way of working directly with the GCS bucket Kaggle uses for the dataset: call KaggleDataset().get_gcs_path() on Kaggle, record the GCS path it gives you. You can then use it from anywhere. Two caveats: the Kaggle API gets you a bucket in the same region as your Kaggle TPU which might not be the region you want. Also, The Kaggle API treats this GCS bucket as a cache. It can erase and recreate it under a different name at any time. It should stay good for a couple of days though.", "votes": null }, { "id": "780887", "postDate": "03/20/2020 17:54:15", "content": "Got it. Thanks a lot!
", "rawMarkdown": "Got it. Thanks a lot!", "votes": null }, { "id": "804105", "postDate": "04/11/2020 08:27:52", "content": "What if I want to use the dataset in Google Colab?
", "rawMarkdown": "What if I want to use the dataset in Google Colab?", "votes": null }, { "id": "806332", "postDate": "04/13/2020 16:17:24", "content": "the answer is above:
\n\n\n", "rawMarkdown": "the answer is above:\n\n> I recommend you download the dataset and upload it to a GCS bucket.\n\n> There is a way of working directly with the GCS bucket Kaggle uses for the dataset: call KaggleDataset().getgcspath() on Kaggle, record the GCS path it gives you. You can then use it from anywhere. Two caveats: the Kaggle API gets you a bucket in the same region as your Kaggle TPU which might not be the region you want. Also, The Kaggle API treats this GCS bucket as a cache. It can erase and recreate it under a different name at any time. It should stay good for a couple of days though.", "votes": null } ], "comments": [ { "id": 780880, "author_name": "mgorner", "author_url": "", "post_date": "03/20/2020 17:46:07", "content": "I recommend you download the dataset and upload it to a GCS bucket.
\n \nThere is a way of working directly with the GCS bucket Kaggle uses for the dataset: call KaggleDataset().getgcspath() on Kaggle, record the GCS path it gives you. You can then use it from anywhere. Two caveats: the Kaggle API gets you a bucket in the same region as your Kaggle TPU which might not be the region you want. Also, The Kaggle API treats this GCS bucket as a cache. It can erase and recreate it under a different name at any time. It should stay good for a couple of days though.
\n
If you are working on your local machine, download the dataset to your local machine. You will not be able to use a TPU though.
\n\nIf you want to work with a TPU-enabled Notebook VM in GCP (here is how to provision one: bit.ly/keras-tpu-tf21), then I recommend you download the dataset and upload it to a GCS bucket. TPUs can only read data from GCS.
\n\nThere is a way of working directly with the GCS bucket Kaggle uses for the dataset: call KaggleDataset().get_gcs_path() on Kaggle, record the GCS path it gives you. You can then use it from anywhere. Two caveats: the Kaggle API gets you a bucket in the same region as your Kaggle TPU which might not be the region you want. Also, The Kaggle API treats this GCS bucket as a cache. It can erase and recreate it under a different name at any time. It should stay good for a couple of days though.
", "votes": null, "replies": [ { "id": 780887, "author_name": "alexandersoare", "author_url": "", "post_date": "03/20/2020 17:54:15", "content": "Got it. Thanks a lot!
", "votes": null, "replies": [] } ] }, { "id": 804105, "author_name": "rhtsingh", "author_url": "", "post_date": "04/11/2020 08:27:52", "content": "What if I want to use the dataset in Google Colab?
", "votes": null, "replies": [ { "id": 806332, "author_name": "mgorner", "author_url": "", "post_date": "04/13/2020 16:17:24", "content": "the answer is above:
\n\n\n", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "780873": "I'm working on this competition on my local machine but unable to run `from kaggle_datasets import KaggleDatasets` as I'm getting `ModuleNotFoundError: No module named 'kaggle.datasets'`.\n\nI tried `pip install kaggle` but that didn't help.", "780880": "If you are working on your local machine, download the dataset to your local machine. You will not be able to use a TPU though.\n\nIf you want to work with a TPU-enabled Notebook VM in GCP (here is how to provision one: [bit.ly/keras-tpu-tf21](bit.ly/keras-tpu-tf21)), then I recommend you download the dataset and upload it to a GCS bucket. TPUs can only read data from GCS.\n\nThere is a way of working directly with the GCS bucket Kaggle uses for the dataset: call KaggleDataset().get_gcs_path() on Kaggle, record the GCS path it gives you. You can then use it from anywhere. Two caveats: the Kaggle API gets you a bucket in the same region as your Kaggle TPU which might not be the region you want. Also, The Kaggle API treats this GCS bucket as a cache. It can erase and recreate it under a different name at any time. It should stay good for a couple of days though.", "780887": "Got it. Thanks a lot!", "804105": "What if I want to use the dataset in Google Colab?", "806332": "the answer is above:\n\n> I recommend you download the dataset and upload it to a GCS bucket.\n\n> There is a way of working directly with the GCS bucket Kaggle uses for the dataset: call KaggleDataset().getgcspath() on Kaggle, record the GCS path it gives you. You can then use it from anywhere. Two caveats: the Kaggle API gets you a bucket in the same region as your Kaggle TPU which might not be the region you want. Also, The Kaggle API treats this GCS bucket as a cache. It can erase and recreate it under a different name at any time. It should stay good for a couple of days though." }, "source": "meta" }I recommend you download the dataset and upload it to a GCS bucket.
\n \nThere is a way of working directly with the GCS bucket Kaggle uses for the dataset: call KaggleDataset().getgcspath() on Kaggle, record the GCS path it gives you. You can then use it from anywhere. Two caveats: the Kaggle API gets you a bucket in the same region as your Kaggle TPU which might not be the region you want. Also, The Kaggle API treats this GCS bucket as a cache. It can erase and recreate it under a different name at any time. It should stay good for a couple of days though.
\n