I am struggling to find a gs://... bucket address for the training data. I'm sure it must have been provided somewhere on the competition page but I can't find it. Can someone help?
", "messages": [ { "id": 745889, "postDate": "2020-02-14T10:40:49.893Z", "content": "I am struggling to find a gs://... bucket address for the training data. I'm sure it must have been provided somewhere on the competition page but I can't find it. Can someone help?
", "rawMarkdown": "I am struggling to find a gs://... bucket address for the training data. I'm sure it must have been provided somewhere on the competition page but I can't find it. Can someone help?", "votes": 1 }, { "id": 745970, "postDate": "2020-02-14T12:41:46.217Z", "content": "Unfortunately, there is no public bucket. There are a few ways to get the data described here. I think the approach using the Kaggle API is not working anymore. This would have been my favorite.
", "rawMarkdown": "Unfortunately, there is no public bucket. There are a few ways to get the data described [here](https://www.kaggle.com/c/deepfake-detection-challenge/discussion/121194). I think the approach using the Kaggle API is not working anymore. This would have been my favorite.", "votes": 2, "replies": [ { "id": 745995, "postDate": "2020-02-14T13:06:01.410Z", "content": "Aha. Thanks. There was reference to a GCS bucket on the competition page here. I have managed to download the data using the CurlWget Chrome extension described on the page you reference. I will create my own GCS bucket that way but not sure how much this will cost in GCP credits. Thanks for the help @seesee.
", "rawMarkdown": "Aha. Thanks. There was reference to a GCS bucket on the competition page [here](https://www.kaggle.com/c/deepfake-detection-challenge/overview/getting-started). I have managed to download the data using the CurlWget Chrome extension described on the page you reference. I will create my own GCS bucket that way but not sure how much this will cost in GCP credits. Thanks for the help @seesee." }, { "id": 746970, "postDate": "2020-02-15T19:45:48.213Z", "content": "Does it mean that there is no way to directly grab it from Google GCS bucket and put it into my own GCS bucket, I will have to first download it into my own hard disk, only then upload into my own GCS bucket?
", "rawMarkdown": "Does it mean that there is no way to directly grab it from Google GCS bucket and put it into my own GCS bucket, I will have to first download it into my own hard disk, only then upload into my own GCS bucket?" }, { "id": 747324, "postDate": "2020-02-16T09:06:28.650Z", "content": "@chewkokwahibrainai I think that is the only way unfortunately.
", "rawMarkdown": "@chewkokwahibrainai I think that is the only way unfortunately.", "votes": 1 } ] }, { "id": 746368, "postDate": "2020-02-14T22:43:48.957Z", "content": "Have you tried to download data directly to GCP notebook local disk? Just like in AWS SageMaker local disk(S3 bucket is not used)
", "rawMarkdown": "Have you tried to download data directly to GCP notebook local disk? Just like in AWS SageMaker local disk(S3 bucket is not used)", "isDeleted": true, "replies": [ { "id": 746384, "postDate": "2020-02-14T23:16:27.010Z", "content": "I guess you can always wget using the terminal of your notebook.
Thanks ! Just wonder if GCP bucket can be skipped since I see many discussions are about GCP bucket.
", "rawMarkdown": "Thanks ! Just wonder if GCP bucket can be skipped since I see many discussions are about GCP bucket. ", "isDeleted": true }, { "id": 746600, "postDate": "2020-02-15T08:55:03.110Z", "content": "If you want to use TFRecords and TF2 you have to use GCP buckets. Using the local file system is not supported.
", "rawMarkdown": "If you want to use TFRecords and TF2 you have to use GCP buckets. Using the local file system is [not supported](https://cloud.google.com/tpu/docs/troubleshooting#common-errors).", "votes": 1 }, { "id": 746869, "postDate": "2020-02-15T16:54:49.827Z", "content": "I am downloading to a GCP instance local drive. Then unzip and upload files to GCS bucket. TPUs can only read data from a bucket, not from a GCP virtual machine local drive.
", "rawMarkdown": "I am downloading to a GCP instance local drive. Then unzip and upload files to GCS bucket. TPUs can only read data from a bucket, not from a GCP virtual machine local drive.", "votes": 1 }, { "id": 746989, "postDate": "2020-02-15T20:26:45.373Z", "content": "Thanks !! Now I see. Upvoted your answer !!
", "rawMarkdown": "Thanks !! Now I see. Upvoted your answer !!", "isDeleted": true } ] } ], "comments": [ { "id": 745970, "author_name": "See--", "author_url": "", "post_date": "2020-02-14T12:41:46.217000", "content": "Unfortunately, there is no public bucket. There are a few ways to get the data described here. I think the approach using the Kaggle API is not working anymore. This would have been my favorite.
", "votes": 2, "replies": [ { "id": 745995, "author_name": "Ken Krige", "author_url": "", "post_date": "2020-02-14T13:06:01.410000", "content": "Aha. Thanks. There was reference to a GCS bucket on the competition page here. I have managed to download the data using the CurlWget Chrome extension described on the page you reference. I will create my own GCS bucket that way but not sure how much this will cost in GCP credits. Thanks for the help @seesee.
", "votes": 0, "replies": [] }, { "id": 746970, "author_name": "Chew Kok Wah", "author_url": "", "post_date": "2020-02-15T19:45:48.213000", "content": "Does it mean that there is no way to directly grab it from Google GCS bucket and put it into my own GCS bucket, I will have to first download it into my own hard disk, only then upload into my own GCS bucket?
", "votes": 0, "replies": [] }, { "id": 747324, "author_name": "Ken Krige", "author_url": "", "post_date": "2020-02-16T09:06:28.650000", "content": "@chewkokwahibrainai I think that is the only way unfortunately.
", "votes": 1, "replies": [] } ] }, { "id": 746368, "author_name": "", "author_url": "", "post_date": "2020-02-14T22:43:48.957000", "content": "Have you tried to download data directly to GCP notebook local disk? Just like in AWS SageMaker local disk(S3 bucket is not used)
", "votes": 0, "replies": [ { "id": 746384, "author_name": "Debanga Raj Neog", "author_url": "", "post_date": "2020-02-14T23:16:27.010000", "content": "I guess you can always wget using the terminal of your notebook.
Thanks ! Just wonder if GCP bucket can be skipped since I see many discussions are about GCP bucket.
", "votes": 0, "replies": [] }, { "id": 746600, "author_name": "See--", "author_url": "", "post_date": "2020-02-15T08:55:03.110000", "content": "If you want to use TFRecords and TF2 you have to use GCP buckets. Using the local file system is not supported.
", "votes": 1, "replies": [] }, { "id": 746869, "author_name": "Ken Krige", "author_url": "", "post_date": "2020-02-15T16:54:49.827000", "content": "I am downloading to a GCP instance local drive. Then unzip and upload files to GCS bucket. TPUs can only read data from a bucket, not from a GCP virtual machine local drive.
", "votes": 1, "replies": [] }, { "id": 746989, "author_name": "", "author_url": "", "post_date": "2020-02-15T20:26:45.373000", "content": "Thanks !! Now I see. Upvoted your answer !!
", "votes": 0, "replies": [] } ] } ], "raw_markdown_by_id": { "745889": "I am struggling to find a gs://... bucket address for the training data. I'm sure it must have been provided somewhere on the competition page but I can't find it. Can someone help?", "745970": "Unfortunately, there is no public bucket. There are a few ways to get the data described [here](https://www.kaggle.com/c/deepfake-detection-challenge/discussion/121194). I think the approach using the Kaggle API is not working anymore. This would have been my favorite.", "746368": "Have you tried to download data directly to GCP notebook local disk? Just like in AWS SageMaker local disk(S3 bucket is not used)" } }