GoogleNet weights trained on the Places dataset for Keras. AcharyaDebaditya KHOSHELHAMKOUROSH WINTERSTEPHAN 2020 This file (googlenet_weights.h5) contains the initial weights of GoogleNet (aka v1) trained on the Places dataset, and is used for fine-tuning task related to pose regression networks. The file was downloaded from <i><a href="https://github.com/kentsommer/keras-posenet">https://github.com/kentsommer/keras-posenet</a>.</i> Adding to the repository is only for the purpose of eliminating dependency on external URLs.<div><br></div><div>Other files contain the weights of the final trained model from our experiments of the paper Recurrent BIM-PoseNet: </div><div><br></div><div><b>SynCar</b> - Weights of model fine-tuned on Synthetic Cartoonish images.</div><div><b>SynPhoReal</b> - Weights of model fine-tuned on Synthetic photo-realistic images.</div><div><b>SynPhoRealTex </b>- Weights of model fine-tuned on Synthetic photo-realistic textured images.</div><div><b>GradmagSynCar</b> - Weights of model fine-tuned on synthetic gradmag of SynCar images.</div><div><b>EdgeRender </b>- Weights of model fine-tuned on Synthetic edge render images.</div><div><br></div><div>Other details in the name of the weight files describes the parameters, such as window length, learning rate, batch, ...., etc. <br><div><br></div><div>This is a support file for the code available at <i><a href="https://github.com/debaditya-unimelb/RecurrentBIM-PoseNet">https://github.com/debaditya-unimelb/RecurrentBIM-PoseNet</a>.</i></div></div>