ConvReLU3d

Quantization-aware fused 3-D conv + ReLU — trainable, fake-quant per forward.

A Conv3d whose forward applies ReLU before the output fake-quant, so the activation observer sees the true post-ReLU (non-negative) range and calibrates a grid that spends all of its codes on values the layer can actually emit. Fusing conv and ReLU also lets a single activation observer stand in for both, which is exactly the shape lucid.quantization.convert needs to collapse them into one kernel.

Like the base Conv3d, the weight and the (post-ReLU) output are fake-quantized every forward under a straight-through estimator (STE): rounding happens in the forward pass while gradients pass straight through to the float kernel, so it stays fully trainable and learns to compensate for the eventual int8 rounding. Built from a fused float lucid.nn.intrinsic.ConvReLU3d by lucid.quantization.prepare_qat; lucid.quantization.convert folds it into a single quantized lucid.nn.quantized.ConvReLU3d.

On each forward the kernel is fake-quantized, F.conv3d runs, ReLU clips, and the result is fake-quantized onto the observed post-ReLU grid:

where $\star$ is the 3-D cross-correlation, $s, z$ are the observer's scale / zero-point and $q_{\min}, q_{\max}$ the grid bounds. Because the activation observer sits outside the ReLU, the calibrated $(s, z)$ describe the non-negative output.