Typo

docs/performance.rst

@@ -14,7 +14,7 @@ Introduction
 
 There are a growing number of quantum simulators available for research and industry use. Many of them perform quite well for smaller number of qubits, and are suitable for non-rigorous experimental explorations. Fewer projects are suitable as "high performance" candidates in the >32 qubit range. Many rely on the common approach often described as the "Schrödinger method,"  doubling RAM usage by a factor of 2 per fully interoperable qubit, or else Feynman path integrals, which can become intractible at arbitrary circuit depth. Inspired by IBM's `Pareto-Efficient Quantum Circuit Simulation Using Tensor Contraction Deferral` [Pednault2017]_ paper, instead with a Dirac "ket" centered approach to Schmidt decomposition, and with more recent attention to potential improvements inspired by Gottesman-Knill stabilizer simulators, Qrack can execute surprisingly general circuits past 32 qubits in width on modest single nodes.
 
-Qrack is an open-source quantum computer simulator option, implemented in C++, directly wrapped for Python via "PyQrack,: supporting integration into other popular compilers and interfaces, suitable for utilization in a wide variety of projects. As such, it is an ideal test-bed for establishing a set of benchmarks useful for comparing performance between various quantum simulators.
+Qrack is an open-source quantum computer simulator option, implemented in C++, directly wrapped for Python via "PyQrack," supporting integration into other popular compilers and interfaces, suitable for utilization in a wide variety of projects. As such, it is an ideal test-bed for establishing a set of benchmarks useful for comparing performance between various quantum simulators.
 
 Qrack provides a "QEngineCPU" and a "QEngineOCL" that represent non-OpenCL and OpenCL base implementations for Schrödinger method simulation. "QHybrid" switches off between these two types internally for best performance at low qubit widths. "QStabilizerHybrid" switches off internally between Gottesman-Knill "stabilizer" simulation and Schrödinger method. For general use cases, the "QUnit" layer provides explicit Schmidt decomposition on top of another engine type (per [Pednault2017]_). "QPager" segments a Schrödinger method simulation into equally sized "pages" that can be run on multiple OpenCL devices or multiple maximum allocation segments of a single device, increasing greatest maximally entangled width. A "QEngine" type is always the base layer, and QUnit, QStabilizerHybrid, and QPager types may be layered over these, and over each other.