Hi,
I had two questions and was wondering if I could get some help.
(1) We submitted a model that passed the Synthetic Data test but was REJECTED on UW data. The submission id is: 9709250. Could we get a log of what happened, please?
(2) We are trying to still get things working (as a part of class project where we have selected this problem and its due this week). Unfortunately, in an attempt to familiarize ourselves with the data and system, it appears that we have exceeded the number of submissions allowed per week. I was wondering if it could be reset or provided a few extra builds as a one time case? Let me know if I've misunderstood submission quota rules. I tried searching the forum but couldn't find a definitive answer.
Thanks in advance!
Created by raivivek Hi @raivivek,
The error log during the training stage for both 9709347 and 9709350 was:
```
info: reading observation table..
---------------------------------------------------------------------------
ValueError Traceback (most recent call last)
/app/notebooks/COVID19.ipynb in
12 imp_mean = SimpleImputer(missing_values=np.nan, strategy='median')
13 observ["tobacco"] = observ["tobacco"].apply(lambda x: 1 if x == "Yes" else 0)
---> 14 observ["oxygen"] = imp_mean.fit_transform(observ[["oxygen"]])
/opt/conda/lib/python3.8/site-packages/pandas/core/frame.py in __setitem__(self, key, value)
3038 else:
3039 # set column
-> 3040 self._set_item(key, value)
3041
3042 def _setitem_slice(self, key: slice, value):
/opt/conda/lib/python3.8/site-packages/pandas/core/frame.py in _set_item(self, key, value)
3115 self._ensure_valid_index(value)
3116 value = self._sanitize_column(key, value)
-> 3117 NDFrame._set_item(self, key, value)
3118
3119 # check if we are modifying a copy
/opt/conda/lib/python3.8/site-packages/pandas/core/generic.py in _set_item(self, key, value)
3578 return
3579
-> 3580 NDFrame._iset_item(self, loc, value)
3581
3582 def _set_is_copy(self, ref, copy: bool_t = True) -> None:
/opt/conda/lib/python3.8/site-packages/pandas/core/generic.py in _iset_item(self, loc, value)
3567
3568 def _iset_item(self, loc: int, value) -> None:
-> 3569 self._mgr.iset(loc, value)
3570 self._clear_item_cache()
3571
/opt/conda/lib/python3.8/site-packages/pandas/core/internals/managers.py in iset(self, loc, value)
1094 blk_locs = blklocs[val_locs.indexer]
1095 if blk.should_store(value):
-> 1096 blk.set(blk_locs, value_getitem(val_locs))
1097 else:
1098 unfit_mgr_locs.append(blk.mgr_locs.as_array[blk_locs])
/opt/conda/lib/python3.8/site-packages/pandas/core/internals/blocks.py in set(self, locs, values)
329 create a new array and always creates a new Block.
330 """
--> 331 self.values[locs] = values
332
333 def delete(self, loc) -> None:
ValueError: shape mismatch: value array of shape (0,71792) could not be broadcast to indexing result of shape (1,71792)
```
Thanks,
@trberg Hi @trberg
I was wondering if you could provide a log of why this job failed: 9709347 and 9709349. Thanks! Okay, thank you! I will keep you updated and reach out again should we need help. I went ahead and changed your test model that passed validation to INVALID. You should be able to submit again. If you're able to get your model working quickly but need to get a few more submissions in, let us know. We can make something happen.
Thanks,
@trberg @trberg Thank you for the log!
We aim to present our models/early-results on Friday. I'm not sure how many tries it would take to just have the model run on the full internal UW data since we don't have that working yet (for our model). So to be honest, we could as many submissions as possible. We can revert to a quota-friendly pace after the due date. Sorry if this was not specific and unhelpful. Please use your discretion and we'll respect that. Thanks! Hi @raivivek,
Submission 9709251 failed from exceeding the RAM limit.
```
info: model type is train
/app/train.sh: line 5: 8 Killed MODEL_TYPE=${MODEL_TYPE} ipython notebooks/COVID19.ipynb
```
The limit of successful submissions should be one per day. What day is this due? And how many submissions do you need by when?
Thanks,
@trberg