My progress in 3 days of Submission Scoring Error has changed to Submission CSV Not Found. Questions: when sending Notepad for verification, make a prediction on test.csv or train.csv or sample_submission.csv. which DCM folder to use ../input/osic-pulmonary-fibrosis-progression/test or ../input/osic-pulmonary-fibrosis-progression/train.
\nSuch moments as -12 133 I took into account, save SUBMISSINO3_pred 2.to_csv(\"./submission.csv\", index=False), internet off
You either have a syntax error that comes up when running on the private test set or you are running into memory issues during the run on the private test set. If you are trying to use the dicom images for predictions remember that there are roughly 200 patients on the hidden test set.
\nYour notebook is exiting before it outputs predictions and therefore no submission.csv file to provide a leader board score.
", "rawMarkdown": "You either have a syntax error that comes up when running on the private test set or you are running into memory issues during the run on the private test set. If you are trying to use the dicom images for predictions remember that there are roughly 200 patients on the hidden test set.\n\nYour notebook is exiting before it outputs predictions and therefore no submission.csv file to provide a leader board score.", "replies": [ { "id": 985973, "postDate": "2020-08-26T06:11:09.713Z", "content": "Hi, my algorithm does the following: loads train.csv and loads test. csv defines directories for train dicom and test. dicom, as you can see everything is loaded dynamically.
\nTRAIN = pd.read_csv(\"../input/osic-pulmonary-fibrosis-progression/train.csv\")
\nTEST = pd.read_csv(\"../input/osic-pulmonary-fibrosis-progression/test.csv\")
\nSUB = pd.read_csv(\"../input/osic-pulmonary-fibrosis-progression/sample_submission.csv\")
\nTRAIN_DIR = \"../input/osic-pulmonary-fibrosis-progression/train/\"
\nTEST_DIR = \"../input/osic-pulmonary-fibrosis-progression/test/\"
\nTEST[\"hack\"] = \"../input/fibrose-zefir/linze5.png\"
\nTRAIN[\"hack\"] = \"../input/fibrose-zefir/linze5.png\"
\nTRAIN = TRAIN.loc[~TRAIN['Patient'].isin([\"ID00011637202177653955184\",\"ID00052637202186188008618\"])]
\nTRAIN.reset_index(drop=True, inplace=True)
\nTEST.reset_index(drop=True, inplace=True)
\nNext, I train EfficientNetB0 installed not over the Internet!!! on table data and images 128x128 batch=32
\n200 patients, are you sure? Up to this point, it seemed to me that their number corresponds to train.csv, then the logic of the Patient column makes sense
note
\nI have predictions for the entire train. csv, for every patient from -12 to 133. When I send sample_submission everything works. When I do SUB2 = pd. merge (TEST, MY_SUB, on='Patient_Week', has='left') or SUB2 = pd. merge(sample_submission, MY_SUB, on= 'Patient_Week', how= 'left') I get an error.
\nConclusion I have only one there are restrictions on FVC and Confidence, probably it is |FVC_true-FVC_pred|>????
@zefirchik Your answers are slightly confusing, but from the look of the notebook link you provided above it seems to me that you are hard-coding everything. You are not actually predicting on the hidden test set.
\nYou have just merged your csv files which is not going to work and your submission.csv format is not correct. Your patient ids are all over the place since you have sorted according to the week number. You are supposed to keep the FVC and Confidence values for a single patient from weeks -12 to 133 together (You may ask the competition hosts if this is an issue, because I'm just assuming).
", "rawMarkdown": "@zefirchik Your answers are slightly confusing, but from the look of the notebook link you provided above it seems to me that you are hard-coding everything. You are not actually predicting on the hidden test set. \n\nYou have just merged your csv files which is not going to work and your submission.csv format is not correct. Your patient ids are all over the place since you have sorted according to the week number. You are supposed to keep the FVC and Confidence values for a single patient from weeks -12 to 133 together (You may ask the competition hosts if this is an issue, because I'm just assuming)." } ] }, { "id": 985819, "postDate": "2020-08-26T03:17:08.310Z", "content": "the same to you , run notebook successfully and there is a submission CSV file output ,but commit showing no file [
\n", "rawMarkdown": "the same to you , run notebook successfully and there is a submission CSV file output ,but commit showing no file [\n\n](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F1722747%2F40c505835ffd71df8e6d4ac1963acc7d%2F2020-08-26%2011-09-46.png?generation=1598411741367176&alt=media)", "replies": [ { "id": 986668, "postDate": "2020-08-26T17:12:49.907Z", "content": "I think I understand. I was training Label Encoder() on a training set and it should have been like this
\nPatient = LabelEncoder()
\nSex = LabelEncoder()
\nSmokingStatus = LabelEncoder()
\ntrain_pac = TRAIN_C[\"Patient\"].unique().tolist()
\ntrain_pac.extend(TEST_C[\"Patient\"].unique().tolist())
\nall_pacient = np.unique(train_pac)
\nPatient.fit(all_pacient)
\nSex.fit(TRAIN[\"Sex\"].unique())
\nSmokingStatus.fit(TRAIN[\"SmokingStatus\"].unique())
You should commit your kernel, this way it will run automatically the entire kernel from start to finish, make sure there are no errors else it will stop.
\nI'm sorry but I don't really understand your text. After running your entire kernel, an output file should exist within this kernel. If you name it submission.csv, Kaggle will automatically detect when submitting.
\nsubmission.to_csv(\"submission.csv\", index = False)
\nwhere submission is your dataframe, or whatever you use for the predictions
", "rawMarkdown": "You should commit your kernel, this way it will run automatically the entire kernel from start to finish, make sure there are no errors else it will stop.\n\nI'm sorry but I don't really understand your text. After running your entire kernel, an output file should exist within this kernel. If you name it submission.csv, Kaggle will automatically detect when submitting.\n\nsubmission.to_csv(\"submission.csv\", index = False)\n\nwhere submission is your dataframe, or whatever you use for the predictions" }, { "id": 985329, "postDate": "2020-08-25T16:43:14.707Z", "content": "My progress in 3 days of Submission Scoring Error has changed to Submission CSV Not Found. Questions: when sending Notepad for verification, make a prediction on test.csv or train.csv or sample_submission.csv. which DCM folder to use ../input/osic-pulmonary-fibrosis-progression/test or ../input/osic-pulmonary-fibrosis-progression/train.
\nSuch moments as -12 133 I took into account, save SUBMISSINO3_pred 2.to_csv(\"./submission.csv\", index=False), internet off
You either have a syntax error that comes up when running on the private test set or you are running into memory issues during the run on the private test set. If you are trying to use the dicom images for predictions remember that there are roughly 200 patients on the hidden test set.
\nYour notebook is exiting before it outputs predictions and therefore no submission.csv file to provide a leader board score.
", "votes": 0, "replies": [ { "id": 985973, "author_name": "zefirchik", "author_url": "", "post_date": "2020-08-26T06:11:09.713000", "content": "Hi, my algorithm does the following: loads train.csv and loads test. csv defines directories for train dicom and test. dicom, as you can see everything is loaded dynamically.
\nTRAIN = pd.read_csv(\"../input/osic-pulmonary-fibrosis-progression/train.csv\")
\nTEST = pd.read_csv(\"../input/osic-pulmonary-fibrosis-progression/test.csv\")
\nSUB = pd.read_csv(\"../input/osic-pulmonary-fibrosis-progression/sample_submission.csv\")
\nTRAIN_DIR = \"../input/osic-pulmonary-fibrosis-progression/train/\"
\nTEST_DIR = \"../input/osic-pulmonary-fibrosis-progression/test/\"
\nTEST[\"hack\"] = \"../input/fibrose-zefir/linze5.png\"
\nTRAIN[\"hack\"] = \"../input/fibrose-zefir/linze5.png\"
\nTRAIN = TRAIN.loc[~TRAIN['Patient'].isin([\"ID00011637202177653955184\",\"ID00052637202186188008618\"])]
\nTRAIN.reset_index(drop=True, inplace=True)
\nTEST.reset_index(drop=True, inplace=True)
\nNext, I train EfficientNetB0 installed not over the Internet!!! on table data and images 128x128 batch=32
\n200 patients, are you sure? Up to this point, it seemed to me that their number corresponds to train.csv, then the logic of the Patient column makes sense
note
\nI have predictions for the entire train. csv, for every patient from -12 to 133. When I send sample_submission everything works. When I do SUB2 = pd. merge (TEST, MY_SUB, on='Patient_Week', has='left') or SUB2 = pd. merge(sample_submission, MY_SUB, on= 'Patient_Week', how= 'left') I get an error.
\nConclusion I have only one there are restrictions on FVC and Confidence, probably it is |FVC_true-FVC_pred|>????
@zefirchik Your answers are slightly confusing, but from the look of the notebook link you provided above it seems to me that you are hard-coding everything. You are not actually predicting on the hidden test set.
\nYou have just merged your csv files which is not going to work and your submission.csv format is not correct. Your patient ids are all over the place since you have sorted according to the week number. You are supposed to keep the FVC and Confidence values for a single patient from weeks -12 to 133 together (You may ask the competition hosts if this is an issue, because I'm just assuming).
", "votes": 0, "replies": [] } ] }, { "id": 985819, "author_name": "Feng", "author_url": "", "post_date": "2020-08-26T03:17:08.310000", "content": "the same to you , run notebook successfully and there is a submission CSV file output ,but commit showing no file [
\n", "votes": 0, "replies": [ { "id": 986668, "author_name": "zefirchik", "author_url": "", "post_date": "2020-08-26T17:12:49.907000", "content": "I think I understand. I was training Label Encoder() on a training set and it should have been like this
\nPatient = LabelEncoder()
\nSex = LabelEncoder()
\nSmokingStatus = LabelEncoder()
\ntrain_pac = TRAIN_C[\"Patient\"].unique().tolist()
\ntrain_pac.extend(TEST_C[\"Patient\"].unique().tolist())
\nall_pacient = np.unique(train_pac)
\nPatient.fit(all_pacient)
\nSex.fit(TRAIN[\"Sex\"].unique())
\nSmokingStatus.fit(TRAIN[\"SmokingStatus\"].unique())
You should commit your kernel, this way it will run automatically the entire kernel from start to finish, make sure there are no errors else it will stop.
\nI'm sorry but I don't really understand your text. After running your entire kernel, an output file should exist within this kernel. If you name it submission.csv, Kaggle will automatically detect when submitting.
\nsubmission.to_csv(\"submission.csv\", index = False)
\nwhere submission is your dataframe, or whatever you use for the predictions
", "votes": 0, "replies": [] }, { "id": 985971, "author_name": "", "author_url": "", "post_date": "2020-08-26T06:10:45.617000", "content": "", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "985863": "You either have a syntax error that comes up when running on the private test set or you are running into memory issues during the run on the private test set. If you are trying to use the dicom images for predictions remember that there are roughly 200 patients on the hidden test set.\n\nYour notebook is exiting before it outputs predictions and therefore no submission.csv file to provide a leader board score.", "985819": "the same to you , run notebook successfully and there is a submission CSV file output ,but commit showing no file [\n\n](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F1722747%2F40c505835ffd71df8e6d4ac1963acc7d%2F2020-08-26%2011-09-46.png?generation=1598411741367176&alt=media)", "985667": "You should commit your kernel, this way it will run automatically the entire kernel from start to finish, make sure there are no errors else it will stop.\n\nI'm sorry but I don't really understand your text. After running your entire kernel, an output file should exist within this kernel. If you name it submission.csv, Kaggle will automatically detect when submitting.\n\nsubmission.to_csv(\"submission.csv\", index = False)\n\nwhere submission is your dataframe, or whatever you use for the predictions", "985329": "My progress in 3 days of Submission Scoring Error has changed to Submission CSV Not Found. Questions: when sending Notepad for verification, make a prediction on test.csv or train.csv or sample_submission.csv. which DCM folder to use ../input/osic-pulmonary-fibrosis-progression/test or ../input/osic-pulmonary-fibrosis-progression/train.\nSuch moments as -12 133 I took into account, save SUBMISSINO3_pred 2.to_csv(\"./submission.csv\", index=False), internet off", "985971": "" } }