borgwardtlab / set_functions_for_time_series Goto Github PK

Can you please elaborate how to run models on mimic3 dataset? I have access for the dataset but I am not sure in which directory I should put the data.

Hi, I follow the instructions and when comes the moment of running the model I get the following error

File "<string>", line 1, in <module> File "/usr/local/lib/python3.7/importlib/__init__.py", line 127, in import_module return _bootstrap._gcd_import(name[level:], package, level) File "<frozen importlib._bootstrap>", line 1006, in _gcd_import File "<frozen importlib._bootstrap>", line 983, in _find_and_load File "<frozen importlib._bootstrap>", line 967, in _find_and_load_unlocked File "<frozen importlib._bootstrap>", line 677, in _load_unlocked File "<frozen importlib._bootstrap_external>", line 728, in exec_module File "<frozen importlib._bootstrap>", line 219, in _call_with_frames_removed File "/opt/seft-hypoxia/seft/cli/fit_model.py", line 13, in <module> from seft.training_routine import TrainingLoop File "/opt/seft-hypoxia/seft/training_routine.py", line 12, in <module> from .callbacks import ( File "/opt/seft-hypoxia/seft/callbacks.py", line 8, in <module> import tensorflow_datasets as tfds File "/opt/virtualenvs/seft-NyYu6qKP-py3.7/lib/python3.7/site-packages/tensorflow_datasets/__init__.py", line 51, in <module> from tensorflow_datasets import __init__py3 as api File "/opt/virtualenvs/seft-NyYu6qKP-py3.7/lib/python3.7/site-packages/tensorflow_datasets/__init__py3.py", line 43, in <module> from tensorflow_datasets.core import tf_compat File "/opt/virtualenvs/seft-NyYu6qKP-py3.7/lib/python3.7/site-packages/tensorflow_datasets/core/__init__.py", line 21, in <module> tf_compat.ensure_tf_install() File "/opt/virtualenvs/seft-NyYu6qKP-py3.7/lib/python3.7/site-packages/tensorflow_datasets/core/tf_compat.py", line 63, in ensure_tf_install "This version of TensorFlow Datasets requires TensorFlow " ImportError: This version of TensorFlow Datasets requires TensorFlow version >= 2.1.0; Detected an installation of version 1.15.3. Please upgrade TensorFlow to proceed.
My environment is a singularity container based on ubuntu docker image with python 3.7.4

I installed poetry using pip3 install, as it was the only way I found to make poetry persistent in the container.

Hi,

If I install Seft using poetry, the environment contains tensorflow 1.15 and tensorflow_datasets 4.2, thus when I try to generate my own data set using the generated environment, I run into a dependency error (as tfds 4.2 requires tf 2.1).

Maybe the solution is to add tfds == 2.0 explicitly in SEFT pyproject.toml file?

Thank you and best regards.

Dear BorgwardtLab,

I am ran the Transformer model with the code in this repo. I assumed that this code is the Transformer model without masking. However, I am curious about the way to run this model with masking. It would be my appreciation if I can get your help! Thank you!

Hi there, I have a question about calculating dp_mask for x_t and m_dp_mask for m_t in your GRU-D implementation (file gru_d.py).

First, the dp_mask is generated from GRUCell built-in function get_dropout_mask_for_cell: code
Then, the dropout mask m_dp_mask for masking vector m_t is generated by calling _generate_dropout_mask: code
By doing so, the dp_mask and m_dp_mask zero out different elements in two inputs x_t and m_t. I can reproduce your result, however, I think that the dropout masks should be the same for x_t and m_t. Can you please clarify this for me? Did I misunderstand something in the core TensorFlow implementation/your implementation?

Thanks for the great work!

Hi!
I am trying to run quick start following your installation guide
but having problem with installing keras-transformer with poetry
It looks like there are some version problem.
Any help to solve this problem would be so much appreciated!

Thanks for amazing job

Installing keras-transformer (0.1.dev27+g347cb21 /workspace/inkyung/Seminar/Set_Functions_for_Time_Series/repos/keras-transformer): Failed
EnvCommandError

Command ['/root/.cache/pypoetry/virtualenvs/seft-ih-2Jpha-py3.7/bin/pip', 'install', '--no-deps', '-U', '/workspace/inkyu
ng/Seminar/Set_Functions_for_Time_Series/repos/keras-transformer'] errored with the following return code 1, and output:
Processing ./repos/keras-transformer
ERROR: Command errored out with exit status 1:
command: /root/.cache/pypoetry/virtualenvs/seft-ih-2Jpha-py3.7/bin/python -c 'import sys, setuptools, tokenize; sys.
argv[0] = '"'"'/tmp/pip-req-build-h6aratez/setup.py'"'"'; file='"'"'/tmp/pip-req-build-h6aratez/setup.py'"'"';f=getattr
(tokenize, '"'"'open'"'"', open)(file);code=f.read().replace('"'"'\r\n'"'"', '"'"'\n'"'"');f.close();exec(compile(code, file, '"'"'exec'"'"'))' egg_info --egg-base /tmp/pip-pip-egg-info-dk1gxjae
cwd: /tmp/pip-req-build-h6aratez/
Complete output (28 lines):
WARNING: The wheel package is not available.
Traceback (most recent call last):
File "", line 1, in
File "/tmp/pip-req-build-h6aratez/setup.py", line 35, in
include_package_data=True,
File "/root/.cache/pypoetry/virtualenvs/seft-ih-2Jpha-py3.7/lib/python3.7/site-packages/setuptools/init.py", line 163, in setup
return distutils.core.setup(**attrs)
File "/usr/lib/python3.7/distutils/core.py", line 108, in setup
_setup_distribution = dist = klass(attrs)
File "/root/.cache/pypoetry/virtualenvs/seft-ih-2Jpha-py3.7/lib/python3.7/site-packages/setuptools/dist.py", line 430, in init
k: v for k, v in attrs.items()
File "/usr/lib/python3.7/distutils/dist.py", line 292, in init
self.finalize_options()
File "/root/.cache/pypoetry/virtualenvs/seft-ih-2Jpha-py3.7/lib/python3.7/site-packages/setuptools/dist.py", line 721, in finalize_options
ep(self)
File "/root/.cache/pypoetry/virtualenvs/seft-ih-2Jpha-py3.7/lib/python3.7/site-packages/setuptools/dist.py", line 728, in _finalize_setup_keywords
ep.load()(self, ep.name, value)
File "/tmp/pip-req-build-h6aratez/.eggs/setuptools_scm-5.0.1-py3.7.egg/setuptools_scm/integration.py", line 26, inversion_keyword
dist.metadata.version = _get_version(config)
File "/tmp/pip-req-build-h6aratez/.eggs/setuptools_scm-5.0.1-py3.7.egg/setuptools_scm/init.py", line 173, in _get_version
parsed_version = _do_parse(config)
File "/tmp/pip-req-build-h6aratez/.eggs/setuptools_scm-5.0.1-py3.7.egg/setuptools_scm/init.py", line 142, in _do_parse
"use git+https://github.com/user/proj.git#egg=proj" % config.absolute_root
LookupError: setuptools-scm was unable to detect version for '/tmp/pip-req-build-h6aratez'.

  Make sure you're either building from a fully intact git repository or PyPI tarballs. Most other sources (such as GitHub's tarballs, a git checkout without the .git folder) don't contain the necessary metadata and will not work.

  For example, if you're using pip, instead of https://github.com/user/proj/archive/master.zip use git+https://github.com/user/proj.git#egg=proj
  ----------------------------------------

ERROR: Command errored out with exit status 1: python setup.py egg_info Check the logs for full command output.
WARNING: You are using pip version 20.2.2; however, version 20.3.3 is available.
You should consider upgrading via the '/root/.cache/pypoetry/virtualenvs/seft-ih-2Jpha-py3.7/bin/python -m pip install --upgrade pip' command.

at ~/.poetry/lib/poetry/utils/env.py:1074 in run
1070│ output = subprocess.check_output(
1071│ cmd, stderr=subprocess.STDOUT, **kwargs
1072│ )
1073│ except CalledProcessError as e:
→ 1074│ raise EnvCommandError(e, input=input)
1075│
1076│ return decode(output)
1077│
1078│ def execute(self, bin, *args, **kwargs):
`

The changes that @ExpectationMax made in August are not reflected in the commit id of the submodule. Is this safe to update?

When I run the TransformerModel, the AttributeError appears, and the codes are:

Hello, I am interested in running your algorithm and your code for my data (irregularly sampled time series), do you have documentation or advice on how can I do that?

Thank you and best regards.

Hi there,

Thank you for both the medical_ts_dataset repo and this repo. Very reproducible work, which is super exciting for healthcare research.

I was wondering if you can help me reconcile your work and this paper: https://arxiv.org/abs/2101.10318

Both has a MIMIC3 table using AUROC as the point of comparison. However, the numbers and ordering of methods are quite different. I understand that this work followed mimic3-benchmarks train-test-val splits and IP-net/mTAND followed a different data split. However, do you suppose that would contribute to such a difference , or could it also be some sort of data-processing/normalization process that are different between the two works?

Mainly, I'm also interested in the authors take on the current research progress for irregular time-series classification. There have been many papers demonstrating superior performance over GRU-D yet in this work, GRU-D is quite strong if not the best for MIMIC3, in addition to being very strong for Physionet 2012 and 2019. (I also followed up with the issue between different masking for x_t and m_t, and setting them to be the same doesn't downgrade performance that much, which implies that GRU-D is still strong). Is it still SOTA?

When trying to run the following cmd:

$ poetry run seft_fit_model --dataset physionet2019 --balance --log_dir test_transformer TransformerModel

I get the following error on my M3 Macbook Pro:

zsh: illegal hardware instruction poetry run seft_fit_model --dataset physionet2019 --balance --log_dir

Hi Max, in the "SF for TS" paper you called your novel model "SeFT-ATTN". I was expecting to see the exact same name in this repo to see how you implemented it, but can't find it anywhere. Does the "deep_set_attention.py" file contain the SeFT-ATTN implementation?