- Sort Score
- Result 10 results
- Languages All
Results 11 - 13 of 13 for depthwise_conv_2d (0.16 sec)
-
tensorflow/compiler/mlir/quantization/tensorflow/passes/lift_quantizable_spots_as_functions.cc
// Disable quantization for the DepthwiseConv since it has no benefits in // the XLA opset. if (function_name.contains("depthwise_conv2d")) { return absl::InternalError( "DepthwiseConv2D doesn't get any benefit of quantization in XLA."); } else if (function_name.contains("conv2d")) { // For Conv2D, the channel dimension must be static to calculate the
Registered: Sun Jun 16 05:45:23 UTC 2024 - Last Modified: Fri May 10 04:07:09 UTC 2024 - 16.4K bytes - Viewed (0) -
tensorflow/compiler/mlir/lite/transforms/prepare_patterns.td
(TFL_DequantizeOp (TFL_QuantizeOp (TF_ReshapeOp $input, $shape), (UpdateShapeWithAxis<3> $qtype, $old_value))), [(UsedBy<"DepthwiseConv2D"> $old_value), (CanUpdateShapeWithAxis<3> $qtype, $old_value)], [], (addBenefit 10)>; // The axis is set to 3, because this transpose is from the legalization of
Registered: Sun Jun 16 05:45:23 UTC 2024 - Last Modified: Tue Apr 30 00:40:15 UTC 2024 - 10.5K bytes - Viewed (0) -
tensorflow/compiler/mlir/quantization/tensorflow/passes/preprocess_op.cc
// than function name. if (!function_name.starts_with("composite_")) { return failure(); } if (function_name.contains("depthwise_conv2d")) { // Uniform Quantized op requires weights of tf.DepthwiseConv2dNative to // be transformed from [H,W,C,M] to [H,W,1,CxM] where // H=height,W=width,C=channel,M=multiplier. Therefore, a reshape op is
Registered: Sun Jun 16 05:45:23 UTC 2024 - Last Modified: Thu Apr 25 16:01:03 UTC 2024 - 11.4K bytes - Viewed (0)