Hello everyone! First of all here is the link to my Notebook: https://www.kaggle.com/code/victoraugustocavalli/neurips-2025
\nWhen testing my code runs fine and generates the correct submission file in every way I tested…
\nI simply have no idea why my code gives this problem during submition…
\nAny help will be greatly appreciated!!!!
", "messages": [ { "id": 3269965, "postDate": "2025-08-15T12:23:05.357Z", "content": "You can try to debug it by setting your predictions to 0 and placing the code to do so at some suspicious places in your code. My bet is your model has difficulties with those degenerate cases where ingress/egress is incomplete.
", "rawMarkdown": "You can try to debug it by setting your predictions to 0 and placing the code to do so at some suspicious places in your code. My bet is your model has difficulties with those degenerate cases where ingress/egress is incomplete.", "votes": 1, "replies": [ { "id": 3269980, "postDate": "2025-08-15T13:08:30.713Z", "content": "Thank you for your reply, I think i'm missing something, because I did a new notebook with only this code:
\n numpy np # linear algebra\n pandas pd # data processing, CSV file I/O (e.g. pd.read_csv)\n glob\n os\n# This works submitting, however submitting it does \n# names_test = glob.glob(\"/kaggle/input/ariel-data-challenge-2025/train/*\")\nnames_test = glob.glob(\"/kaggle/input/ariel-data-challenge-2025/test/*\")\ndf = pd.read_csv(\"/kaggle/input/ariel-data-challenge-2025/sample_submission.csv\")\ndg = df.()\ndf.(, inplace=)\ntyped = df[\"planet_id\"].dtype\ndf[\"planet_id\"] = [(item[:]) item names_test]\ndf[\"planet_id\"] = df[\"planet_id\"].astype(typed)\ndf.loc[:,df.[:].tolist()] = np.repeat(dg.iloc[:,:]., repeats=len(names_test), axis=)\ndf.to_csv(,=)\nand I'm still getting a Notebook Threw Exception error, I will try copying someones public notebook… but I have no idea if/what I'm doing wrong here…
", "rawMarkdown": "Thank you for your reply, I think i'm missing something, because I did a new notebook with only this code:\n\n```\nimport numpy as np # linear algebra\nimport pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\nimport glob\nimport os\n# This works when not submitting, however when submitting it does not work\n# names_test = glob.glob(\"/kaggle/input/ariel-data-challenge-2025/train/*\")\nnames_test = glob.glob(\"/kaggle/input/ariel-data-challenge-2025/test/*\")\ndf = pd.read_csv(\"/kaggle/input/ariel-data-challenge-2025/sample_submission.csv\")\ndg = df.copy()\ndf.drop(0, inplace=True)\ntyped = df[\"planet_id\"].dtype\ndf[\"planet_id\"] = [int(item[46:]) for item in names_test]\ndf[\"planet_id\"] = df[\"planet_id\"].astype(typed)\ndf.loc[:,df.columns[1:].tolist()] = np.repeat(dg.iloc[:,1:].values, repeats=len(names_test), axis=0)\ndf.to_csv('submission.csv',index=False)\n```\n\nand I'm still getting a Notebook Threw Exception error, I will try copying someones public notebook... but I have no idea if/what I'm doing wrong here...", "replies": [ { "id": 3269987, "postDate": "2025-08-15T13:24:37.280Z", "content": "I can't say for sure, but this code looks super fragile:
\n[int(item[46:]) for item in names_test]
\nI would use something like
\n[os.path.split(item)[1] for item in names_test]
I just tried this change, with no luck, any other Ideas? thanks again!! I`m trying to access the submission logs, but it seems like it is not possible anymore…
", "rawMarkdown": "I just tried this change, with no luck, any other Ideas? thanks again!! I`m trying to access the submission logs, but it seems like it is not possible anymore..." }, { "id": 3270267, "postDate": "2025-08-16T05:58:43.147Z", "content": "Since you have so few lines now, you can just remove them one by one until the error stops…
", "rawMarkdown": "Since you have so few lines now, you can just remove them one by one until the error stops...", "votes": 2 }, { "id": 3270319, "postDate": "2025-08-16T08:51:13.740Z", "content": "Why do you do this:
\n\n", "rawMarkdown": "Why do you do this:\n\n> df.drop(0, inplace=True)", "votes": 1 }, { "id": 3270429, "postDate": "2025-08-16T13:59:26.363Z", "content": "df.drop(0, inplace=True)
\n
I understand that the planet in the sample submission does not necessarily exists in the test, that is why I remove it… Do you think the problem lies there? Thank for replying! One thing I did not mention is that my code runs for around 2 hours before getting the error, which is the time my code takes to preprocess 1100 planets of data, so that leads me to believe the error is at the end of the code, that is why I isolated this part.
", "rawMarkdown": "I understand that the planet in the sample submission does not necessarily exists in the test, that is why I remove it... Do you think the problem lies there? Thank for replying! One thing I did not mention is that my code runs for around 2 hours before getting the error, which is the time my code takes to preprocess 1100 planets of data, so that leads me to believe the error is at the end of the code, that is why I isolated this part." }, { "id": 3270589, "postDate": "2025-08-16T21:49:44.670Z", "content": "Try without it.
", "rawMarkdown": "Try without it.", "votes": 1 }, { "id": 3271270, "postDate": "2025-08-18T12:59:20.657Z", "content": "Hey, I have been trying many things the last few days, I tried many times removing the df.drop line, tried changing the submission code entirely, then, looking at my 20+ failed submissions I notice one of the first ones got me a Scoring error, instead of a Threw Exception error, so I'm trying to make that code work, it is basically this now:
\nnames_test = os()\ndf = pd()\ntyped = df\n ((names_test)):\n new_row = ]+df()\n df=new_row\ndf(, inplace=True)\ndf(drop=True, inplace=True)\ndf = df(typed)\ndf(, index=False)\nAny help will be greatly appreciated, I really want to participate in this competition.
\nHere is my csv submission generated with train data:
Hm you definitely have the planet IDs sorted in the wrong order, though I'm not sure that explains your issue. You want to be using the ordering in test_star_info.csv, not the ordering returned by os.listdir.
", "rawMarkdown": "Hm you definitely have the planet IDs sorted in the wrong order, though I'm not sure that explains your issue. You want to be using the ordering in test_star_info.csv, not the ordering returned by os.listdir.", "votes": 1 }, { "id": 3271508, "postDate": "2025-08-19T03:54:54.060Z", "content": "Hey, I tried to use sort_values which created a dataframe in the same order as train_star_info, using the following code:
\nnames_test = os()\ndf = pd()\ntyped = df\n ((names_test)):\n planet_id = (os(names_test)(, ))\n new_row = + df()\n df = new_row\ndf(, inplace=True)\ndf = df(typed)\ndf(by=, inplace=True)\ndf(drop=True, inplace=True)\ndf(, index=False)\nbasically added a new way to get the names, and used sort_values, visual inspection showed promising results, the planet_id column is equal in every way, but I still get a notebook scoring error… is it supposed to be this difficult?
", "rawMarkdown": "Hey, I tried to use sort_values which created a dataframe in the same order as train_star_info, using the following code:\n\n```\nnames_test = os.listdir(\"/kaggle/input/ariel-data-challenge-2025/test\")\ndf = pd.read_csv(\"/kaggle/input/ariel-data-challenge-2025/sample_submission.csv\")\ntyped = df.planet_id.dtype\nfor i in range(len(names_test)):\n planet_id = int(os.path.splitext(names_test[i])[0].replace(\"planet\", \"\"))\n new_row = [planet_id] + df.iloc[0,1:].to_list()\n df.loc[len(df)] = new_row\ndf.drop(0, inplace=True)\ndf.planet_id = df.planet_id.astype(typed)\ndf.sort_values(by=\"planet_id\", inplace=True)\ndf.reset_index(drop=True, inplace=True)\ndf.to_csv(\"submission.csv\", index=False)\n```\nbasically added a new way to get the names, and used sort_values, visual inspection showed promising results, the planet_id column is equal in every way, but I still get a notebook scoring error... is it supposed to be this difficult?\n" }, { "id": 3272069, "postDate": "2025-08-20T08:03:55.367Z", "content": "During submission, sample_submission.csv has the correct number of rows, so df.drop(0, inplace=True) doesn't delete all of the initial values.
\nI strongly recommend you to not construct the dataframe by yourself, but to update the values in sample_submission.csv like
df = pd.read_csv()\n planet_id list_planet_id:\n # Update df[df.planet_id] == planet_id]\ndf.to_csv(, =)\nOh really? That's good to know.
", "rawMarkdown": "Oh really? That's good to know.", "votes": 1 }, { "id": 3272428, "postDate": "2025-08-20T23:36:33.057Z", "content": "OK, I tried out what you told me and it FREAKING WORKED, hahaha
\nHere is the code for the dummy data submission, gonna make it work with my actual notebook.
\ndf = pd.read_csv(, =)\ndg = pd.DataFrame(=df.index, =df.columns)\n i df.index:\n dg.loc[i] = np.repeat(0.5, len(df.columns))\ndf.reset_index(=)\ndg.reset_index(=)\n col dg.columns:\n dg[col] = dg[col].astype(df[col].dtype)\ndg.to_csv(, =)\nThank you so much for everyone of you who took some time to reply to my query. ❤️
\nNow after a week of trying to make it work I finally can progress a little.
\nI can tell you that the correct results are not 0.5 hahaha.
A good check would be to run it with all the test data replaced by train data (i.e. replace lines like \"test_planet_list = glob.glob(\"/kaggle/input/ariel-data-challenge-2025/test/\")\" by \"test_planet_list = glob.glob(\"/kaggle/input/ariel-data-challenge-2025/train/\")\". Perhaps your code fails in some way if there is more than one test planet. You'd run this offline of course, not as a submission.
\nOtherwise, you can submit only parts of your script (writing a dummy submission at the end), so that you can zoom into where the error occurs.
\nGood luck, these can take a lot of perserverance…
", "rawMarkdown": "A good check would be to run it with all the test data replaced by train data (i.e. replace lines like \"test_planet_list = glob.glob(\"/kaggle/input/ariel-data-challenge-2025/test/*\")\" by \"test_planet_list = glob.glob(\"/kaggle/input/ariel-data-challenge-2025/train/*\")\". Perhaps your code fails in some way if there is more than one test planet. You'd run this offline of course, not as a submission.\n\nOtherwise, you can submit only parts of your script (writing a dummy submission at the end), so that you can zoom into where the error occurs.\n\nGood luck, these can take a lot of perserverance...", "votes": 2, "replies": [ { "id": 3269702, "postDate": "2025-08-14T21:48:58.603Z", "content": "Hey thank you for your reply, actually I have done that already (replacing the line in glob with train) and it works as is, only when submitting the notebook i get an error.
\nI really don't know what to do, maybe inference with 1200 hidden test planets fails? however I have tried that with all training data and it takes less than 1 minute to run…
\nI really don't know. :(
", "rawMarkdown": "Hey thank you for your reply, actually I have done that already (replacing the line in glob with train) and it works as is, only when submitting the notebook i get an error.\n\nI really don't know what to do, maybe inference with 1200 hidden test planets fails? however I have tried that with all training data and it takes less than 1 minute to run...\n\nI really don't know. :(" } ] }, { "id": 3269683, "postDate": "2025-08-14T20:38:01.547Z", "content": "Hello everyone! First of all here is the link to my Notebook: https://www.kaggle.com/code/victoraugustocavalli/neurips-2025
\nWhen testing my code runs fine and generates the correct submission file in every way I tested…
\nI simply have no idea why my code gives this problem during submition…
\nAny help will be greatly appreciated!!!!
", "rawMarkdown": "Hello everyone! First of all here is the link to my Notebook: https://www.kaggle.com/code/victoraugustocavalli/neurips-2025\n\nWhen testing my code runs fine and generates the correct submission file in every way I tested...\n\nI simply have no idea why my code gives this problem during submition...\n\nAny help will be greatly appreciated!!!!" } ], "comments": [ { "id": 3269965, "author_name": "DennisSakva", "author_url": "", "post_date": "2025-08-15T12:23:05.357000", "content": "You can try to debug it by setting your predictions to 0 and placing the code to do so at some suspicious places in your code. My bet is your model has difficulties with those degenerate cases where ingress/egress is incomplete.
", "votes": 1, "replies": [ { "id": 3269980, "author_name": "Victor Cavalli", "author_url": "", "post_date": "2025-08-15T13:08:30.713000", "content": "Thank you for your reply, I think i'm missing something, because I did a new notebook with only this code:
\n numpy np # linear algebra\n pandas pd # data processing, CSV file I/O (e.g. pd.read_csv)\n glob\n os\n# This works submitting, however submitting it does \n# names_test = glob.glob(\"/kaggle/input/ariel-data-challenge-2025/train/*\")\nnames_test = glob.glob(\"/kaggle/input/ariel-data-challenge-2025/test/*\")\ndf = pd.read_csv(\"/kaggle/input/ariel-data-challenge-2025/sample_submission.csv\")\ndg = df.()\ndf.(, inplace=)\ntyped = df[\"planet_id\"].dtype\ndf[\"planet_id\"] = [(item[:]) item names_test]\ndf[\"planet_id\"] = df[\"planet_id\"].astype(typed)\ndf.loc[:,df.[:].tolist()] = np.repeat(dg.iloc[:,:]., repeats=len(names_test), axis=)\ndf.to_csv(,=)\nand I'm still getting a Notebook Threw Exception error, I will try copying someones public notebook… but I have no idea if/what I'm doing wrong here…
", "votes": 0, "replies": [ { "id": 3269987, "author_name": "DennisSakva", "author_url": "", "post_date": "2025-08-15T13:24:37.280000", "content": "I can't say for sure, but this code looks super fragile:
\n[int(item[46:]) for item in names_test]
\nI would use something like
\n[os.path.split(item)[1] for item in names_test]
I just tried this change, with no luck, any other Ideas? thanks again!! I`m trying to access the submission logs, but it seems like it is not possible anymore…
", "votes": 0, "replies": [] }, { "id": 3270267, "author_name": "Jeroen Cottaar", "author_url": "", "post_date": "2025-08-16T05:58:43.147000", "content": "Since you have so few lines now, you can just remove them one by one until the error stops…
", "votes": 2, "replies": [] }, { "id": 3270319, "author_name": "CPMP", "author_url": "", "post_date": "2025-08-16T08:51:13.740000", "content": "Why do you do this:
\n\n", "votes": 1, "replies": [] }, { "id": 3270429, "author_name": "Victor Cavalli", "author_url": "", "post_date": "2025-08-16T13:59:26.363000", "content": "df.drop(0, inplace=True)
\n
I understand that the planet in the sample submission does not necessarily exists in the test, that is why I remove it… Do you think the problem lies there? Thank for replying! One thing I did not mention is that my code runs for around 2 hours before getting the error, which is the time my code takes to preprocess 1100 planets of data, so that leads me to believe the error is at the end of the code, that is why I isolated this part.
", "votes": 0, "replies": [] }, { "id": 3270589, "author_name": "CPMP", "author_url": "", "post_date": "2025-08-16T21:49:44.670000", "content": "Try without it.
", "votes": 1, "replies": [] }, { "id": 3271270, "author_name": "Victor Cavalli", "author_url": "", "post_date": "2025-08-18T12:59:20.657000", "content": "Hey, I have been trying many things the last few days, I tried many times removing the df.drop line, tried changing the submission code entirely, then, looking at my 20+ failed submissions I notice one of the first ones got me a Scoring error, instead of a Threw Exception error, so I'm trying to make that code work, it is basically this now:
\nnames_test = os()\ndf = pd()\ntyped = df\n ((names_test)):\n new_row = ]+df()\n df=new_row\ndf(, inplace=True)\ndf(drop=True, inplace=True)\ndf = df(typed)\ndf(, index=False)\nAny help will be greatly appreciated, I really want to participate in this competition.
\nHere is my csv submission generated with train data:
Hm you definitely have the planet IDs sorted in the wrong order, though I'm not sure that explains your issue. You want to be using the ordering in test_star_info.csv, not the ordering returned by os.listdir.
", "votes": 1, "replies": [] }, { "id": 3271508, "author_name": "Victor Cavalli", "author_url": "", "post_date": "2025-08-19T03:54:54.060000", "content": "Hey, I tried to use sort_values which created a dataframe in the same order as train_star_info, using the following code:
\nnames_test = os()\ndf = pd()\ntyped = df\n ((names_test)):\n planet_id = (os(names_test)(, ))\n new_row = + df()\n df = new_row\ndf(, inplace=True)\ndf = df(typed)\ndf(by=, inplace=True)\ndf(drop=True, inplace=True)\ndf(, index=False)\nbasically added a new way to get the names, and used sort_values, visual inspection showed promising results, the planet_id column is equal in every way, but I still get a notebook scoring error… is it supposed to be this difficult?
", "votes": 0, "replies": [] }, { "id": 3272069, "author_name": "c-number", "author_url": "", "post_date": "2025-08-20T08:03:55.367000", "content": "During submission, sample_submission.csv has the correct number of rows, so df.drop(0, inplace=True) doesn't delete all of the initial values.
\nI strongly recommend you to not construct the dataframe by yourself, but to update the values in sample_submission.csv like
df = pd.read_csv()\n planet_id list_planet_id:\n # Update df[df.planet_id] == planet_id]\ndf.to_csv(, =)\nOh really? That's good to know.
", "votes": 1, "replies": [] }, { "id": 3272428, "author_name": "Victor Cavalli", "author_url": "", "post_date": "2025-08-20T23:36:33.057000", "content": "OK, I tried out what you told me and it FREAKING WORKED, hahaha
\nHere is the code for the dummy data submission, gonna make it work with my actual notebook.
\ndf = pd.read_csv(, =)\ndg = pd.DataFrame(=df.index, =df.columns)\n i df.index:\n dg.loc[i] = np.repeat(0.5, len(df.columns))\ndf.reset_index(=)\ndg.reset_index(=)\n col dg.columns:\n dg[col] = dg[col].astype(df[col].dtype)\ndg.to_csv(, =)\nThank you so much for everyone of you who took some time to reply to my query. ❤️
\nNow after a week of trying to make it work I finally can progress a little.
\nI can tell you that the correct results are not 0.5 hahaha.
A good check would be to run it with all the test data replaced by train data (i.e. replace lines like \"test_planet_list = glob.glob(\"/kaggle/input/ariel-data-challenge-2025/test/\")\" by \"test_planet_list = glob.glob(\"/kaggle/input/ariel-data-challenge-2025/train/\")\". Perhaps your code fails in some way if there is more than one test planet. You'd run this offline of course, not as a submission.
\nOtherwise, you can submit only parts of your script (writing a dummy submission at the end), so that you can zoom into where the error occurs.
\nGood luck, these can take a lot of perserverance…
", "votes": 2, "replies": [ { "id": 3269702, "author_name": "Victor Cavalli", "author_url": "", "post_date": "2025-08-14T21:48:58.603000", "content": "Hey thank you for your reply, actually I have done that already (replacing the line in glob with train) and it works as is, only when submitting the notebook i get an error.
\nI really don't know what to do, maybe inference with 1200 hidden test planets fails? however I have tried that with all training data and it takes less than 1 minute to run…
\nI really don't know. :(
", "votes": 0, "replies": [] } ] } ], "raw_markdown_by_id": { "3269965": "You can try to debug it by setting your predictions to 0 and placing the code to do so at some suspicious places in your code. My bet is your model has difficulties with those degenerate cases where ingress/egress is incomplete.", "3269688": "A good check would be to run it with all the test data replaced by train data (i.e. replace lines like \"test_planet_list = glob.glob(\"/kaggle/input/ariel-data-challenge-2025/test/*\")\" by \"test_planet_list = glob.glob(\"/kaggle/input/ariel-data-challenge-2025/train/*\")\". Perhaps your code fails in some way if there is more than one test planet. You'd run this offline of course, not as a submission.\n\nOtherwise, you can submit only parts of your script (writing a dummy submission at the end), so that you can zoom into where the error occurs.\n\nGood luck, these can take a lot of perserverance...", "3269683": "Hello everyone! First of all here is the link to my Notebook: https://www.kaggle.com/code/victoraugustocavalli/neurips-2025\n\nWhen testing my code runs fine and generates the correct submission file in every way I tested...\n\nI simply have no idea why my code gives this problem during submition...\n\nAny help will be greatly appreciated!!!!" } }