profile
viewpoint
If you are wondering where the data of this site comes from, please visit https://api.github.com/users/sebimarkgraf/events. GitMemory does not store any data, but only uses NGINX to cache data for a period of time. The idea behind GitMemory is simply to give users a better reading experience.
Sebastian Markgraf sebimarkgraf Karlsruhe, Germany Computer science student at KIT. Working on web and machine learning projects.

sebimarkgraf/synopses-anki 13

German Anki Decks based on KIT lectures

Novare/synopses 5

Möglichst kurz gehaltene Zusammenfassungen von KIT Vorlesungen auf Deutsch. ACHTUNG: Wir sind mittlerweile auf Anki Decks umgestiegen: https://github.com/sebimarkgraf/synopses-anki

3d-printers-distribution/3d-printers-distribution-backend 3

State of the art backend integrated with firebase and deployed on google cloud

sebimarkgraf/synopses 2

Möglichst kurz gehaltene Zusammenfassungen von KIT Vorlesungen auf Deutsch.

sebimarkgraf/kgym 1

OpenAI Gym Implementation for the kaggle environments

sebimarkgraf/resumen 1

Altklausur orientierte Zusammenfassung für die Sicherheits Vorlesung 2018

adrianleh/propa-summary 0

Summary for programming paradigms lecture

Coronicle/coronicle-gcp-functions 0

Testing Google Cloud Functions instead of a full fledged backend

starteddavemlz/eemont

started time in 5 days

issue commentdavemlz/eemont

ee.ImageCollection.getTimeSeriesByRegion breaks when specifying bands=None

Oh wow, that was quick. Thank you very much!

sebimarkgraf

comment created time in a month

issue openeddavemlz/eemont

ee.ImageCollection.getTimeSeriesByRegion breaks when specifying bands=None

Describe the bug When choosing the bands as None as specified in the default and the documentation to get all images, the following error is thrown:

TypeError: object of type 'NoneType' has no len()

To Reproduce Do not specify the bands parameter when using `´getTimeSeriesByRegion``.

Setup (please complete the following information):

  • OS: Linux
  • python version: 3.8
  • eemont version: 0.2.5
  • earthengine-api version: 0.1.269

Additional context Problem seems to be due to setNA using the bands parameter without checking for None. Potential fixes could be changing setNA or settings the bands parameter to all bands when checking for None.

created time in a month

startedonnx/onnxmltools

started time in a month

starteddavemlz/awesome-ee-spectral-indices

started time in a month

issue commentonnx/onnxmltools

ValueError: No proper operator name found for '<class 'pyspark.ml.pipeline.Pipeline'>'

You are trying to export a unfitted Pipeline. If you fit the Pipeline it becomes a PipelineModel and can be exported using the onnxmltools.

This is not an issue with the onnxmltools but could be included as helpful warning.

prashanthharshangi

comment created time in a month

startedlutzroeder/netron

started time in a month

startediterative/dvc

started time in a month

issue closedtensorflow/tensorflow

Change batching in timeseries_dataset_from_array to be optional

System information

  • TensorFlow version (you are using): 2.5.0
  • Are you willing to contribute it (Yes/No): Yes

Describe the feature and the current behavior/state. The keras.preprocessing.timeseries_dataset_from_array function works well for creating timeseries datasets, but when creating the dataset from multiple independent timeseries (e.g. different locations) one would need to merge the datasets. As the current implementation requires batching to be used and to my knowledge batched datasets cannot be concatenated, developers cannot use the generating function in that case. A simple fix would be to allow the developer to provide None for the batch size and do not perform batching in that case. This would be in line with the other parameters.

Will this change the current api? How? Small change, the documentation of batch_size would need to be updated.

Who will benefit with this feature? Users wanting to train more complex timeseries problems.

Any Other info.

closed time in 3 months

sebimarkgraf

issue commenttensorflow/tensorflow

Change batching in timeseries_dataset_from_array to be optional

@Sadulf2019 thank you very much! Definitely going to do this. Then there's no reason to keep this here open.

sebimarkgraf

comment created time in 3 months

fork sebimarkgraf/tensorflow

An Open Source Machine Learning Framework for Everyone

https://tensorflow.org

fork in 3 months

issue openedtensorflow/tensorflow

Change batching in timeseries_dataset_from_array to be optional

System information

  • TensorFlow version (you are using): 2.5.0
  • Are you willing to contribute it (Yes/No): Yes

Describe the feature and the current behavior/state. The keras.preprocessing.timeseries_dataset_from_array function works well for creating timeseries datasets, but when creating the dataset from multiple independent timeseries (e.g. different locations) one would need to merge the datasets. As the current implementation requires batching to be used and to my knowledge batched datasets cannot be concatenated, developers cannot use the generating function in that case. A simple fix would be to allow the developer to provide None for the batch size and do not perform batching in that case. This would be in line with the other parameters.

Will this change the current api? How? Small change, the documentation of batch_size would need to be updated.

Who will benefit with this feature? Users wanting to train more complex timeseries problems.

Any Other info.

created time in 3 months

fork sebimarkgraf/ismn

Readers for the data from the International Soil Moisture Network

https://ismn.earth/en/

fork in 3 months