The SDSS-IV dataset contains information about various astronomical bodies such as Galaxies, Stars, and Quasars captured by observatories. Inspired by our work on deep multimodal learning, which utilized transfer learning to classify the SDSS-IV dataset, we further extended our research in the fine tuning of these architectures to study the effect in the classification scenario. Architectures such as Resnet-50, DenseNet-121 VGG-16, Xception, EfficientNetB2, MobileNetV2 and NasnetMobile have been built using layer wise fine tuning at different levels. Our findings suggest that freezing all layers with Imagenet weights and adding a final trainable layer may not be the optimal solution. Further, baseline models and models that have higher number of trainable layers performed similarly in certain architectures. Model need to be fine tuned at different levels and a specific training ratio is required for a model to be termed ideal. Different architectures had different responses to the change in the number of trainable layers w.r.t accuracies. While models such as DenseNet-121, Xception, EfficientNetB2 achieved peak accuracies that were relatively consistent with near perfect training curves, models such as Resnet-50,VGG-16, MobileNetV2 and NasnetMobile had lower, delayed peak accuracies with poorly fitting training curves. It was also found that though mobile neural networks have lesser parameters and model size, they may not always be ideal for deployment on a low computational device as they had consistently lower validation accuracies. Customized evaluation metrics such as Tuning Parameter Ratio and Tuning Layer Ratio are used for model evaluation.
翻译:SDSS-IV 数据集包含关于各种天体的信息,如星座、星星和由天文台捕获的Quasars等。我们深多式联运学习工作,利用转移学习对 SDSS-IV 数据集进行分类,我们进一步扩展了对这些架构的微调研究,以研究分类设想方案的影响。Resnet-50、DenseNet-121 VGG-16、Xception、高效NetB2、MovelNetV2和NasnetMobile等架构是在不同级别进行层级精细微调的构建的。我们的研究结果表明,将带有图像网重量的所有层冻结并添加最后可训练层也许不是最佳的解决方案。此外,基线模型和模型中具有与某些结构相似的可训练层的更高数量。模型需要在不同级别上进行微调和具体的培训比率才能被称作理想模式。不同的结构对可训练层的低比率值(w.r.t)和NasnetMostairmoverial)的变化有不同的反应模式,而像DencyNet 121、Xcrecial deal ruval ruvalal stration2 listrational strational-rucural 这样的模型则被使用,它们被找到的更接近于一个更精确的精确的精确的模型, rucurvial-deal-deal-deal-deal-deal-deal-deal-deal-deal-deal-deal-deal-ral-ral-deal-ral-deal-deal-rvaldaltradaltradal-cturdal-rvaldaldaldal-deal-deal-deal-deal-rvald-deal-deal-rvaldal-deal-deal-deal-deal-deal-deal-deal-deal-deal-deal-deal-deal-deal-deal-ral-deal-deal-deal-deal-deal-deal-deal-mod-ral-deal-mod-mod-modal-mod-modal-ral-mod-mod-mod-deal-ral-