MLUGS

Protokoll

Vorstellungsrunde

• Andreas, Software-Entwickler, AX-Semantics
• Uwe, Raumfahrt, ST2C
• Johanna, Software-Entwickler, AX-Semantics
• Tin, Wirtschafts-Ing, Daimler
• Benedikt, Data-Scientist, Daimler
• Matthias, VWL, Data-Berater
• Thilo, Data-Engineer, Freelancer
• Thomas, Software-Entwickler, Vialytics
• Michael, Software-Entwickler, AX
• Janis, , P3-Systems
• Dominik, Mathematiker, Sprint1
• Thomas, Physiker/Data-Analyst
• Lukas, Datascientist, Sprint1
• Alex, Security/KI, Bosch
• David, Data-Analyst-Consultant, Diconium

Andreas: AllenNLP Intro

slides: https://github.com/mfa/talks/tree/master/allennlp-intro

Thilo: Improving Standard ResNet-50

• Convolutions, ReLu, BatchNorm
• weitere: LayerNorm, InstanceNorm, GroupNorm
• https://pytorch.org/docs/stable/nn.html#torch.nn.GroupNorm
• ResNeXt
• SENet
• PolyNet
• Deformable Convolution
• NasNet
• https://github.com/tensorflow/cleverhans
• https://github.com/bethgelab/foolbox

• http://cocodataset.org/#detection-leaderboard

• https://github.com/NVIDIA/apex (auch für AllenNLP)

misc

links:

• https://nvlabs.github.io/SPADE/
• https://blogs.nvidia.com/blog/2019/03/18/gaugan-photorealistic-landscapes-nvidia-research/
• https://rapids.ai
• https://github.com/rapidsai/cuml
• https://github.com/understand-ai/anonymizer
• https://github.com/ageitgey/face_recognition

next

• date: 2019-04-16
• vorträge:
• Tin: Facerecognition