Hi All,
\nI was just trying to understand what exactly the model needs to predict and therefore searching for the prediction data format. I could not find any csv/json file (ex: sample_submission.csv) in this competition. I would really appreciate any suggestions/thoughts on this. Thanks in advance!
Hi All,
\nI was just trying to understand what exactly the model needs to predict and therefore searching for the prediction data format. I could not find any csv/json file (ex: sample_submission.csv) in this competition. I would really appreciate any suggestions/thoughts on this. Thanks in advance!
Hi @nanditab35!
\nThe submission format is explained in the Evaluation page: https://www.kaggle.com/competitions/iwildcam2022-fgvc9/overview/evaluation Let me know if that's not clear enough.
\nȘtefan
", "rawMarkdown": "Hi @nanditab35!\n\nThe submission format is explained in the Evaluation page: https://www.kaggle.com/competitions/iwildcam2022-fgvc9/overview/evaluation Let me know if that's not clear enough.\n\nȘtefan", "votes": null }, { "id": "1746380", "postDate": "04/05/2022 17:49:29", "content": "I got it. Thank you for your reply.
", "rawMarkdown": "I got it. Thank you for your reply.", "votes": null }, { "id": "1746388", "postDate": "04/05/2022 17:58:09", "content": "Hi @nanditab35.
\nHere's how I created my sample_submission.csv file. Let me know if anything isn't clear.
\nimport json\nimport pandas as pd\nfrom pathlib import Path\n\nDATA_ROOT = Path(\"your path\")\nJSON_FILENAME = DATA_ROOT / \"iwildcam2022_test_information.json\"\n\nwith open(JSON_FILENAME) as f:\n data = json.load(f)\n\n# extract unique sequence ids from list of test images\nim_sequence_ids = [im['seq_id'] for im in data['images']]\nunique_seqs = set(im_sequence_ids)\n\n# we know total number of sequences from Kaggle \"Submit Predictions\" page\nassert len(unique_seqs) == 11028, \"expected 11028 sequence ids\"\n\nsequences = sorted(list(unique_seqs))\n\n# use Pandas to create dataframe and output as csv\ndata_dict = {\"Id\":sequences, \"Predicted\": [0] * len(sequences)}\n\ndf = pd.DataFrame(data_dict)\nprint(df.head())\n\ndf.to_csv(\"sample.submission.csv\", header=True, index=False)\nThis is really helpful, Thank you!
", "rawMarkdown": "This is really helpful, Thank you!", "votes": null } ], "comments": [ { "id": 1746363, "author_name": "stefanistrate", "author_url": "", "post_date": "04/05/2022 17:34:58", "content": "Hi @nanditab35!
\nThe submission format is explained in the Evaluation page: https://www.kaggle.com/competitions/iwildcam2022-fgvc9/overview/evaluation Let me know if that's not clear enough.
\nȘtefan
", "votes": null, "replies": [ { "id": 1746380, "author_name": "nanditab35", "author_url": "", "post_date": "04/05/2022 17:49:29", "content": "I got it. Thank you for your reply.
", "votes": null, "replies": [] } ] }, { "id": 1746388, "author_name": "cjturner", "author_url": "", "post_date": "04/05/2022 17:58:09", "content": "Hi @nanditab35.
\nHere's how I created my sample_submission.csv file. Let me know if anything isn't clear.
\nimport json\nimport pandas as pd\nfrom pathlib import Path\n\nDATA_ROOT = Path(\"your path\")\nJSON_FILENAME = DATA_ROOT / \"iwildcam2022_test_information.json\"\n\nwith open(JSON_FILENAME) as f:\n data = json.load(f)\n\n# extract unique sequence ids from list of test images\nim_sequence_ids = [im['seq_id'] for im in data['images']]\nunique_seqs = set(im_sequence_ids)\n\n# we know total number of sequences from Kaggle \"Submit Predictions\" page\nassert len(unique_seqs) == 11028, \"expected 11028 sequence ids\"\n\nsequences = sorted(list(unique_seqs))\n\n# use Pandas to create dataframe and output as csv\ndata_dict = {\"Id\":sequences, \"Predicted\": [0] * len(sequences)}\n\ndf = pd.DataFrame(data_dict)\nprint(df.head())\n\ndf.to_csv(\"sample.submission.csv\", header=True, index=False)\nThis is really helpful, Thank you!
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "1746278": "Hi All,\nI was just trying to understand what exactly the model needs to predict and therefore searching for the prediction data format. I could not find any csv/json file (ex: sample_submission.csv) in this competition. I would really appreciate any suggestions/thoughts on this. Thanks in advance!", "1746363": "Hi @nanditab35!\n\nThe submission format is explained in the Evaluation page: https://www.kaggle.com/competitions/iwildcam2022-fgvc9/overview/evaluation Let me know if that's not clear enough.\n\nȘtefan", "1746380": "I got it. Thank you for your reply.", "1746388": "Hi @nanditab35.\n\nHere's how I created my sample_submission.csv file. Let me know if anything isn't clear.\n```\nimport json\nimport pandas as pd\nfrom pathlib import Path\n\nDATA_ROOT = Path(\"your path\")\nJSON_FILENAME = DATA_ROOT / \"iwildcam2022_test_information.json\"\n\nwith open(JSON_FILENAME) as f:\n data = json.load(f)\n\n# extract unique sequence ids from list of test images\nim_sequence_ids = [im['seq_id'] for im in data['images']]\nunique_seqs = set(im_sequence_ids)\n\n# we know total number of sequences from Kaggle \"Submit Predictions\" page\nassert len(unique_seqs) == 11028, \"expected 11028 sequence ids\"\n\nsequences = sorted(list(unique_seqs))\n\n# use Pandas to create dataframe and output as csv\ndata_dict = {\"Id\":sequences, \"Predicted\": [0] * len(sequences)}\n\ndf = pd.DataFrame(data_dict)\nprint(df.head())\n\ndf.to_csv(\"sample.submission.csv\", header=True, index=False)\n```", "1746403": "This is really helpful, Thank you!" }, "source": "meta" }