@makeitworkjml Just to clarify, are we allowed to use https://github.com/switchablenorms/DeepFashion2 as pretraining weights or multi-task ?\nDoes this violate the rule?
\n\nThe annotations of DeepFashion2 are different and cannot be used for inferencing target labels in this competition.
\n\n> \\3. EXTERNAL DATA PROVISIONS\nThe following provisions augment General Rules Section 7.C. below: \"Pre-trained models are allowed in the competition, but are restricted to train their algorithms on iMaterialist 2020 train and validation images. Collecting additional data for the target labels is not allowed. Collecting additional unlabeled data for pretraining is ok. Please specify any and all external data used for training in the specified discussion post. Only image content is allowed for training (e.g. URLs in the provided dataset are not allowed for training). Additional annotation on the provided train and validation data is fine (e.g. bounding boxes, keypoints, etc..). Teams should specify that they collected additional annotations when submitting results.”
", "messages": [ { "id": "807941", "postDate": "04/15/2020 03:55:13", "content": "@makeitworkjml Just to clarify, are we allowed to use https://github.com/switchablenorms/DeepFashion2 as pretraining weights or multi-task ?\nDoes this violate the rule?
\n\nThe annotations of DeepFashion2 are different and cannot be used for inferencing target labels in this competition.
\n\n> \\3. EXTERNAL DATA PROVISIONS\nThe following provisions augment General Rules Section 7.C. below: \"Pre-trained models are allowed in the competition, but are restricted to train their algorithms on iMaterialist 2020 train and validation images. Collecting additional data for the target labels is not allowed. Collecting additional unlabeled data for pretraining is ok. Please specify any and all external data used for training in the specified discussion post. Only image content is allowed for training (e.g. URLs in the provided dataset are not allowed for training). Additional annotation on the provided train and validation data is fine (e.g. bounding boxes, keypoints, etc..). Teams should specify that they collected additional annotations when submitting results.”
", "rawMarkdown": "makeitworkjml Just to clarify, are we allowed to use https://github.com/switchablenorms/DeepFashion2 as pretraining weights or multi-task ?\nDoes this violate the rule?\n\nThe annotations of DeepFashion2 are different and cannot be used for inferencing target labels in this competition.\n\n\n> \\3. EXTERNAL DATA PROVISIONS\nThe following provisions augment General Rules Section 7.C. below: \"Pre-trained models are allowed in the competition, but are restricted to train their algorithms on iMaterialist 2020 train and validation images. Collecting additional data for the target labels is not allowed. Collecting additional unlabeled data for pretraining is ok. Please specify any and all external data used for training in the specified discussion post. Only image content is allowed for training (e.g. URLs in the provided dataset are not allowed for training). Additional annotation on the provided train and validation data is fine (e.g. bounding boxes, keypoints, etc..). Teams should specify that they collected additional annotations when submitting results.”", "votes": null }, { "id": "808034", "postDate": "04/15/2020 05:30:07", "content": "When you want to use someone else work, you must always tell where it came from. If you mentioned its original source then you have no problem. If they don't want you to use it, you won't be able to download it.
", "rawMarkdown": "When you want to use someone else work, you must always tell where it came from. If you mentioned its original source then you have no problem. If they don't want you to use it, you won't be able to download it.", "votes": null }, { "id": "808814", "postDate": "04/15/2020 16:26:29", "content": "Sure, maybe you have misunderstood my question?
\n\nMy question is whether we are able to use external data (e.g. DeepFashion2) in this imat-fashion competition, and whether this is violating the rules of this imat-fashion competition.
", "rawMarkdown": "Sure, maybe you have misunderstood my question?\n\nMy question is whether we are able to use external data (e.g. DeepFashion2) in this imat-fashion competition, and whether this is violating the rules of this imat-fashion competition.", "votes": null }, { "id": "852000", "postDate": "05/18/2020 05:18:03", "content": "If you fine-tune a model pre-trained on DeepFashion2, that is ok.
", "rawMarkdown": "If you fine-tune a model pre-trained on DeepFashion2, that is ok.", "votes": null } ], "comments": [ { "id": 808034, "author_name": "masteronepiece", "author_url": "", "post_date": "04/15/2020 05:30:07", "content": "When you want to use someone else work, you must always tell where it came from. If you mentioned its original source then you have no problem. If they don't want you to use it, you won't be able to download it.
", "votes": null, "replies": [ { "id": 808814, "author_name": "lzhbrian", "author_url": "", "post_date": "04/15/2020 16:26:29", "content": "Sure, maybe you have misunderstood my question?
\n\nMy question is whether we are able to use external data (e.g. DeepFashion2) in this imat-fashion competition, and whether this is violating the rules of this imat-fashion competition.
", "votes": null, "replies": [] } ] }, { "id": 852000, "author_name": "makeitworkjml", "author_url": "", "post_date": "05/18/2020 05:18:03", "content": "If you fine-tune a model pre-trained on DeepFashion2, that is ok.
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "807941": "makeitworkjml Just to clarify, are we allowed to use https://github.com/switchablenorms/DeepFashion2 as pretraining weights or multi-task ?\nDoes this violate the rule?\n\nThe annotations of DeepFashion2 are different and cannot be used for inferencing target labels in this competition.\n\n\n> \\3. EXTERNAL DATA PROVISIONS\nThe following provisions augment General Rules Section 7.C. below: \"Pre-trained models are allowed in the competition, but are restricted to train their algorithms on iMaterialist 2020 train and validation images. Collecting additional data for the target labels is not allowed. Collecting additional unlabeled data for pretraining is ok. Please specify any and all external data used for training in the specified discussion post. Only image content is allowed for training (e.g. URLs in the provided dataset are not allowed for training). Additional annotation on the provided train and validation data is fine (e.g. bounding boxes, keypoints, etc..). Teams should specify that they collected additional annotations when submitting results.”", "808034": "When you want to use someone else work, you must always tell where it came from. If you mentioned its original source then you have no problem. If they don't want you to use it, you won't be able to download it.", "808814": "Sure, maybe you have misunderstood my question?\n\nMy question is whether we are able to use external data (e.g. DeepFashion2) in this imat-fashion competition, and whether this is violating the rules of this imat-fashion competition.", "852000": "If you fine-tune a model pre-trained on DeepFashion2, that is ok." }, "source": "meta" }