Abstract: A training device may comprise a ramp block, which may include a short side facet; an incline facet; a right face, and a left face disposed on opposing sides of the training device; a main base ridge and a plurality of alternate base ridges disposed along a bottom border of the right and left face; and a ramp block base disposed along the incline facet. The right and left face may be integrally formed with the short side and incline facet. A right and left plurality of holes may be disposed on the right and left face. The training device may further include a height adjustment member comprising a height adjustment member base configured to rest on a flat surface; a securing slot; and a frame configured to provide height adjustment via the plurality of holes and securing slot. The securing slot may align with the right/left plurality of holes.

Abstract: A training device may include a frame. The frame may include a main shaft having proximal and distal ends; a left mount and a right mount each disposed perpendicular to the main shaft; a weight guard extending radially from the main shaft; a sheath protruding from the main shaft; and a fastener well disposed on the sheath. Further, the training device may include a t-bar. The t-bar may include a support member comprising a plurality of holes. The sheath may be sized to accept the support member. A left rail and a right rail may each be disposed orthogonal to the support member.

Abstract: A training device may include an outer layer; a first end and a second end located at opposing ends of the training device; an adjustable handle including a first piece, a second piece, and a fastener loop; and a contents disposed within a void defined by the outer layer and first and second ends. Each end may include a cuff. The fastener loop may be configured to accept the first piece and the second piece.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for performing optimization using an optimizer neural network. One of the methods includes for each optimizer network parameter, randomly sampling a perturbation value; generating a plurality of sets of candidate values for the optimizer network parameters, for each set of candidate values of the optimizer network parameters: determining a respective loss value representing a performance of the optimizer neural network in updating one or more sets of inner parameters in accordance with the set of candidate of values of the optimizer network parameters; and updating the current values of the optimizer network parameters based on the loss values for the plurality of sets of candidate values of the optimizer network parameters.