Abstract: Devices, systems, and techniques for experiment-based training of machine learning models (MLMs) using early stopping. The techniques include starting training tracks (TTs) that train candidate MLMs using the same training data and respective sets of training settings, performing a first evaluation of a first candidate MLM prior to completion of a corresponding first TT, and responsive to the first evaluation, placing the first TT on an inactive status, inactive status indicating that further training of the first candidate MLM is to be ceased. The techniques further include continuing at least a second TT using the training data, and responsive to conclusion of the TTs, selecting, as one or more final MLMs, the first candidate MLM or a second candidate MLM.

Abstract: Devices, systems, and techniques for provisioning of cloud-based machine learning training, optimization, and deployment services. The techniques include providing, to a remote client device, a list of available machine learning models (MLMs), receiving from the remote client device an indication of selected MLM(s) from the provided list, identifying training settings for selected MLM(s), identifying a training data for the selected MLM(s), configuring, using the identified training settings, execution of one or more processes to train the selected MLM(s) using the identified training data, and providing to the remote client device a representation of completed training of at least one MLM.

Abstract: Apparatuses, systems, and techniques to provisioning of pipelines for efficient training, retraining, configuring, deploying, and using machine learning models for inference in user-specific platforms.

Abstract: Apparatuses, systems, and techniques to use low precision quantization to train a neural network. In at least one embodiment, one or more weights of a trained model are represented by low bit integer numbers instead of using full floating point precision. Changing precision of the one or more weights is performed by first quantizing all weights and activations of a neural network except for layers that require finer granularity in representation than an 8 bit quantization can provide to generate a first trained model. Subsequently, precision of the one or more weights of the first trained model is changed again to generate a second trained model. For the second trained model, the precision of one or more weights of at least one additional layer is changed in addition to the layers that previously had precision values changed while training the neural network to generate the first trained model.

Abstract: Transfer learning can be used to enable a user to obtain a machine learning model that is fully trained for an intended inferencing task without having to train the model from scratch. A pre-trained model can be obtained that is relevant for that inferencing task. Additional training data, as may correspond to at least one additional class of data, can be used to further train this model. This model can then be pruned and retrained in order to obtain a smaller model that retains high accuracy for the intended inferencing task.

Abstract: Input layers of an element-wise operation in a neural network can be pruned such that the shape (e.g., the height, the width, and the depth) of the pruned layers matches. A pruning engine identifies all of the input layers into the element-wise operation. For each set of corresponding neurons in the input layers, the pruning engine equalizes the metrics associated with the neurons to generate an equalized metric associated with the set. The pruning engine prunes the input layers based on the equalized metrics generated for each unique set of corresponding neurons.