Skip to content

Latest commit

 

History

History
886 lines (444 loc) · 27 KB

CHANGELOG.md

File metadata and controls

886 lines (444 loc) · 27 KB

Change Log

Unreleased

Added

  • Add support for parameter expression in qiskit translation

0.12.0

Added

  • Add translation of r gate from qiskit

  • Add det method at backends

  • Add fermion Gaussian state simulator in fgs.py

  • Add partial_transpose and entanglement_negativity method in quantum.py

  • Add reduced_wavefunction method in quantum.py to get reduced pure state

Changed

  • move ensemble module to applications/ai (breaking changes)

  • tc2qiskit now record qiskit measure with incremental clbit from 0

Fixed

  • Support degenerate eigenvalue for jax backend eigh method when using AD

  • Fixed cu gate translation from qiskit to avoid qiskit bug

  • Fixed jax refactoring (0.4.24) where SVD and QR return a namedtuple instead of a tuple

  • Fix qiskit<1.0 and tf<2.16

0.11.0

Added

  • Add multiple GPU VQE examples using jax pmap

  • Add with_prob option to general_kraus so that the probability of each option can be returned together

  • Add benchmark example showcasing new way of implementing matrix product using vmap

  • Add keras3 example showcasing integration with tc

  • Add circuit copy method that avoid shallow copy issue Circuit.copy()

  • Add end to end infrastructures and methods for classical shadow in shadows.py

  • Add classical shadow tutorial

  • Add NN-VQE tutorial

Fixed

  • improve the adaptive_vmap to support internal jit and pytree output

  • fix pauli_gates dtype unchange issue when set new dtype (not recommend to use this attr anymore)

  • fix rem apply_correction bug when non-numpy backend is set

  • fix tf warning for cast with higher version of tf

Changed

  • The static method BaseCircuit.copy is renamed as BaseCircuit.copy_nodes (breaking changes)

0.10.0

Added

  • c.measure_instruction(*qubits) now supports multiple ints specified at the same time

  • c.expectation_ps() now also supports ps argument directly (pauli structures)

  • Add tc version print in tc.about() method

  • tc now supports fancy batch indexing for gates, e.g. c.rxx([0, 1, 2], [1, 2, 3], theta=K.ones([3]))

  • Task management via group tag (when submit_task and list_tasks)

  • batch_expectation_ps now supports local device without topology and thus unify the interface for numerical exact simulation, numerical simulation with measurement shots and QPU experiments

  • introduce two stage compiling for batch_expectation_ps to save some compiling overhead

  • Add experimental support for ODE backend pulse level control simulation/analog quantum computing

  • make the pulse level control support differentiating the end time

  • Add new qem module with qem methods: zne, dd and rc

Fixed

  • tc.results.counts.plot_histogram now can dispatch kws to corresponding qiskit method

  • New implementation for c.inverse() to partially avoid unrecognized gate name issue

  • Fixed bug for batch_expectation_ps for jax backend

  • Partially fix the SVD numerical stability bug on tf backend when using MPSCircuit

  • List syntax for gate now supports range

0.9.1

Added

  • Add tc.TorchHardwarLayer for shortcut layer construction of quantum hardware experiments

  • Add cotengra contractor setup shortcut

  • Add simplecompiler module to assite qiskit compile for better performance when targeting rz native basis

Changed

  • Add compiler and cloud namespace to the global tensorcircuit namespace

  • Refactor composed compiler pipeline interface to include simple_compiler, using DefaultCompiler for now (breaking)

  • Refactor batch_submit_template wrapper to make it a standard abstraction layer between tc cloud infras and batch_expectation_ps abstraction, providing another way to adpot other cloud providers with only batch_submit_template implemented

Fixed

  • submit_task return (list of dict vs dict) follows the data type of provided circuit instead of the number of circuits

  • Fix qubit mapping related bug when using batch_expectation_ps or simple_compile

0.9.0

Added

  • Cloud module for Tencent QCloud is now merged into the master branch and ready to release

  • Add tc.about() to print related software versions and configs

  • Torch support is upgraded to 2.0, and now support native vmap and native functional grad, and thus vvag. Still jit support is conflict with these functional transformations and be turned off by default

  • Add torch_interfaces_kws that support static keyword arguments when wrapping with the interface

  • Add gpu_memory_share function and enable it by default

  • Add scan methods for backends

  • Add example demontrating how jax compiling time can be accelerated by jax.lax.scan

Fixed

  • Add tests and fixed some missing methods for cupy backend, cupy backend is now ready to use (though still not guaranteed)

  • Fix adjoint gate numpy conversion for fixed gate case

  • Sometime, tf just return IndexedSlice instead of tensor from gradient API, partially fix this in tc backend methods

Changed

  • Upgraded black and mypy==1.2.0 (breaking change for developers)

0.8.0

Added

  • Add initial_mapping circuit method to return a new circuit with given logical_physical_mapping

  • Add get_positional_logical_mapping circuit method to return the mapping when only part of the qubits are measured

  • results.rem.ReadoutMit class now support three layers of abstraction on qubits: positional, logical, and physical

  • Add an example script demonstrating how tc can use external contraction path finder wirtten in Julia

  • Add cals_from_api method for ReadoutMit class which can acquire the readout error information from the api

  • Add experimental compiler module

  • Make the compiler infra more ready for a pipeline compling

  • When translating to qiskit, multicontrol gate is manipulated specifically instead of a general unitary

  • Add qft blocks in template module

  • Add Tensorcircuit MacOS (univerisal) installation guide

  • Add KerasLayer without jit (quantum hardware compatible)

  • Add regularizer support for KerasLayer

  • Add methods in quantum module for translating ps list and xyz argument dict

  • Add templates.ensemble.bagging module for bagging ensemble method

  • The speed of Pauli string sum Hamiltonian generation is improved by a divide-and-conquer sum

Fixed

  • Circuit nosify in noise model now support all circuit attributes apart from qubit number

  • Some string warnings are fixed by using r-string

  • Fix bug in tc.quantum.quimb2qop when mps is the input

  • Fix bug in translation.py when qiskit is not installed

  • Rem results after apply_correction is now sorted

  • Fix KerasLayer so that it supports null weights

  • Fix tf optimizer bug and optimizer compatibility issue with tf2.11

0.7.0

Added

  • Add c.probability() method to return probability amplitude

  • Add results module including funtionalities on count dict manipulation and readout error mitigation (local/global calibriation, scalable counts and expectation mitigation from research papers)

  • Add _extra_qir to store information on hardware level measurement and reset

  • Add enable_instruction option in to_qiskit method that enables measurements in qiskit export

  • Add circuit method measure_instruction, barrier_instruction and reset_instruction for hardware level instruction flags

  • Auto unroll composite qiskit instructions when translating to tc circuit

  • Add binding_parameters argument for translating parameterized qiskit circuit to tc circuit

  • Add keep_measure_order bool option to from_openqasm methods so that the measure instruction order is kept by qiskit

  • Add Chinese translation for doc Sharpbit

  • Add circuit_constructor argument for qiskit2tc method, so that we can support more circuit class than circuit and dmcircuit

Fixed

  • Fix adjoint possible bug with agnostic backend

  • Fix sigmoid bug on pytorch backend

  • Fix relu bug on pytorch backend

  • Ignore ComplexWarning for cast method on numpy and jax backend

  • Fix vjp method bug on tensorflow backend, where none is replaced with zeros

0.6.0

Added

  • Add native support for rxx, ryy and rzz gates for translation from qiskit

  • Add from_openqasm and from_openqasm_file methods for Circuit

  • Add circuit_params argument for translation from qiskit to make the interface universal and consistent with other from_ methods

  • Add shifts tuple parameter for experimental.parameter_shift_grad API so that we can also customize finite difference gradient from this method

  • Add std method for backends

  • Add NoiseModel class to programmably configure the global error model when simulating the quantum circuit

  • Add tc.channels.composedkraus to compose the different Kraus operators as a new one

  • Add direct support for noise model via sample_expectation_ps and expectation methods, both Monte Carlo trajectory and density matrix evolution approaches are supported

Changed

  • Improve the efficiency of sample_expectation_ps method by using cached state

Fixed

  • Fixed unitary_kraus of Circuit class support for multi-qubit kraus channels, previous implementation fails to reshape the multi-qubit kraus tensor as matrix

  • Fixed kraus_to_super_gate bug when multi-qubit kraus channels are presented on tensorflow backend

0.5.0

Added

  • Finished quantum noise modeling and simulation development stage 1. Add more quantum channels and the differentiable transformation between different channel forms. Add readout error support for sample and sample_expectation_ps methods.

  • Add new parameter shift gradient API that supports finite measurement shots and the corresponding example scripts

  • Add openqasm format transformation method c.to_openqasm()

  • Add native support for phase and cphase gates when transforming to qiskit

  • Add native support for rxx, ryy, rzz and u, cu gates when transforming to qiskit

  • Add native support for u gate when transforming from qiskit

  • Add circuit from_qsim_file method to load Google random circuit structure

  • Add searchsorted method for backend

  • Add probability_sample method for backend as an alternative for random_choice since it supports status as external randomness format

  • Add status support for sample and sample_expection_ps methods

Changed

  • The inner mechanism for sample_expectation_ps is changed to sample representation from count representation for a fast speed

Fixed

  • Fixed the breaking change introduced in jax 0.3.18, jax._src is no longer imported into the from the public jax namespace.

  • tc.quantum.correlation_from_samples now fix the sign error with odd number of spins

  • Updated to the latest version of mypy and get rid of lots of type: ignored

  • Fix the dtype bug when float is pass to u gate or phase gate

  • Fix to qiskit bug when parameterized gate has default nonset parameters

  • Fix iswap gate translation to qiskit with support for parameters

0.4.1

Added

  • Add support for weighted graph QAOA in tc.templates.blocks.QAOA_block

  • Add AbstractCircuit to and from json capability (experimental support, subject to change)

  • Add alias for sd and td gate

  • Add phase gate and cphase gate for the circuit

  • Add U gate and CU gate following OpenQASM 3.0 convention

  • Add tc.gates.get_u_parameter to solve the three Euler angle for the U gate given the matrix

  • Add GateVF.ided() method to kron quantum gate with identity gate

  • Add batched_parameters_structures.py example to demonstrate nested vmap and architecture search possibility

  • Add gate_count method on AbstractCircuit to count gates of given type in the circuit

  • Add gate_summary method on AbstractCircuit to count gate by type as a dict

  • Add ccx as another alias for toffoli gate

Changed

  • Seperate channel auto register for circuit class with unitary and general case

  • The old standalone depolarizing implementation can now be called via c.depolarizing_reference

Fixed

  • Move iswap gate to vgates list

  • Fix possible bug when vmap is nested in different order (only affect tensorflow backend)

  • Fix bug when multi input function accept the same variable in different args and gradient or jvp/vjp is required (only affect tensorflow backend)

  • Fix the use of rev over rev jacobian for hessian method and back to the efficient solution of fwd over rev order due to the solution of nested vmap issue on tf backend

  • Identify potential bug in unitary_kraus2 implementation, change to unitary_kraus instead

0.4.0

Added

  • Add sample_expectation_ps method for BaseCircuit, which measure the Pauli string expectation considering measurement shots

  • Add alias expps for expectation_ps and sexpps for sampled_expectation_ps

  • Add counts_d2s and counts_s2d in quantum module to transform different representation of measurement shots results

  • Add vmap enhanced parameter_shift_grad in experimental module (API subjects to change)

  • Add parameter_shift.py script in examples showcasing how to use parameter shift grad wrapper

  • Add vmap_randomness.py script in examples showcasing how to vmap external random generators

  • Add noise_sampling_jit.py script showcasing how real device simulation with sample method is efficiently implemented with Monte Carlo and jit

  • Add jit support and external random management for tc.quantum.measurement_counts

  • Add MPO gate support for multiple qubit gates in MPSCircuit simulator

  • Add the expectation_ps method to MPSCircuit (moving to AbstractCircuit)

  • Add six format of measurement results support and their transformation in quantum module

  • Add format option in Circuit.sample while maintain the backward compatibility

  • Add tc.utils.arg_alias which is a decorator that adds alias argument for function with the doc fixed accordingly

  • Add quantum channel auto resgisteration method in Circuit class

Changed

  • rxx, ryy, rzz gates now has 1/2 factor before theta consitent with rx, ry, rz gates. (breaking change)

  • replace status arguments in sample method as random_generator (new convention: status for 0, 1 uniform randomness and random_generator for random key) (breaking change)

  • Rewrite the expectation method of MPSCircuit to make it general

  • Adjusted the initialization method for MPSCircuit (move the from_wavefunction method and allow QuVector input) (breaking change)

  • tc.quantum.measurement_counts aliased as tc.quantum.measurement_results and change the function arguments (breaking change)

  • Refactor backend to use multiple inheritance approach instead of reflection method

Fixed

  • Add jit support for sample method when allow_state=True

  • Fix the bug that 128 type is converted to 64 value

  • Fix arg_alias bug when the keyword arguments is None by design

  • Fix arg_alias when the docstring for each argument is in multiple lines

  • Noise channel apply methods in DMCircuit can also absorb status keyword (directly omitting it) for a consistent API with Circuit

0.3.1

Added

  • Add overload of plus sign for TensorFlow SparseTensor

  • Add support of dict or list of tensors as the input for tc.KerasLayer

  • Add support of multiple tensor inputs for tc.TorchLayer

  • Both gate index and expectation(_ps) now support negative qubit index, eg. c.H(-1) with the same meaning as numpy indexing

  • Add tc.templates.measurements.parameterized_local_measurements for local Pauli string evaluation

Changed

  • Change pytest xdist option in check_all.sh to -n auto

  • Further refactor the circuit abstraction as AbstractCircuit and BaseCircuit

Fixed

  • Add sgates, vgates, mpogates and gate_aliases back to BaseCircuit for backward compatibility

  • Add auto conversion of dtype for unitary gate

0.3.0

Added

  • Add device and device_move methods on backend with universal device string representation

  • Add to_qir, from_qir, to_circuit, inverse, from_qiskit, to_qiskit, sample, prepend, append, cond_measurment, select_gate method for DMCircuit class

  • Add status arguments as external randomness for perfect_sampling and measure methods

  • DMCircuit now supports mps_inputs and mpo_dminputs

  • Add decorator tc.interfaces.args_to_tensor to auto convert function inputs as tensor format

Changed

  • Refactor circuit and dmcircuit with common methods now in basecircuit.py, and merge dmcircuit.py and dmcircuit2.py, now DMCircuit supports MPO gate and qir representation

Fixed

  • Patially solve the issue with visualization on cond_measure (#50)

0.2.2

Added

  • PyTorch backend support multi pytrees version of tree_map

  • Add dtype backend method which returns the dtype string

  • Add TensorFlow interface

  • Add to_dlpack and from_dlpack method on backends

  • Add enable_dlpack option on interfaces and torchnn

  • Add inverse method for Circuit (#26)

Changed

  • Refactor interfaces code as a submodule and add pytree support for args

  • Change the way to register global setup internally, so that we can skip the list of all submodules

  • Refactor the tensortrans code to a pytree perspective

Fixed

  • Fixed numpy method bug in pytorch backend when the input tensor requires grad (#24) and when the tensor is on GPU (#25)

  • Fixed TorchLayer parameter list auto registeration

  • Pytorch interface is now device aware (#25)

0.2.1

Added

  • Add enable_lightcone option in circuit expectation method, where only gates within casual lightcone of local observable is contracted.

  • Add benchmark function into utils

Fixed

  • Fixed a vital bug on circuit expectation evaluation, a wrongly transposed operator connection is fixed.

  • Name passed in gate application now works as Node name

0.2.0

Added

  • Add PyTorch nn Module wrapper in torchnn

  • Add reverse, mod, left_shift, right_shift, arange methods on backend

  • Brand new sample API with batch support and sampling from state support

  • add more methods in global namespace, and add alias KerasLayer/TorchLayer

Fixed

  • Fixed bug in merge single gates when all gates are single-qubit ones

Changed

  • The default contractor enable preprocessing feature where single-qubit gates are merged firstly

0.1.3

Added

  • Add more type auto conversion for tc.gates.Gate as inputs

  • Add tree_flatten and tree_unflatten method on backends

  • Add torch optimizer to the backend agnostic optimizer abstraction

Changed

  • Refactor the tree utils, add native torch support for pytree utils

Fixed

  • grad in torch backend now support pytrees

  • fix float parameter issue in translation to qiskit circuit (#19)

0.1.2

Added

  • Add rxx, ryy and rzz gate

Fixed

  • Fix installation issue with tensorflow requirements on MACOS with M1 chip

  • Improve M1 macOS compatibility with unjit tensorflow ops

  • Fixed SVD backprop bug on jax backend of wide matrix

  • mps_input dtype auto correction enabled

0.1.1

Added

  • Add quoperator method to get QuOperator representation of the circuit unitary

  • Add coo_sparse_matrix_from_numpy method on backend, where the scipy coo matrix is converted to sparse tensor in corresponding backend

  • Add sparse tensor to scipy coo matrix implementation in numpy method

Changed

  • tc.quantum.PauliStringSum2COO, tc.quantum.PauliStringSum2Dense, and tc.quantum.heisenberg_hamiltonian now return the tensor in current backend format if numpy option sets to False. (Breaking change: previously, the return are fixed in TensorFlow format)

0.1.0

Added

  • DMCircuit also supports array instead of gate as the operator

Fixed

  • fix translation issue to qiskit when the input parameter is in numpy form

  • type conversion in measure API when high precision is set

  • fix bug in to_qiskit with new version qiskit

0.0.220509

Added

  • Add eigvalsh method on backend

Changed

  • post_select method return the measurement result int tensor now, consistent with cond_measure

  • Circuit.measure now point to measure_jit

0.0.220413

Added

  • Add expectation_ps method for DMCircuit

  • Add measure and sample for DMCircuit

Fixed

  • With Circuit.vis_tex, for the Circuit has customized input state, the default visualization is psi instead of all zeros now

  • general_kraus is synced with apply_general_kraus for DMCircuit

  • Fix dtype incompatible issue in kraus methods between status and prob

0.0.220402

Added

  • add utils.append to build function pipeline

  • add mean method on backends

  • add trigonometric methods on backends

  • add conditional_gate to support quantum ops based on previous measurment results

  • add expectation_ps as shortcut to get Pauli string expectation

  • add append and prepend to compose circuits

  • add matrix method to get the circuit unitary matrix

Changed

  • change the return information of unitary_kraus and general_kraus methods

  • add alias for any gate as unitary

0.0.220328

Added

  • add QuOperator convert tools which can convert MPO in the form of TensorNetwork and Quimb into MPO in the form of QuOperator

Changed

  • quantum Hamiltonian generation now support the direct return of numpy form matrix

Fixed

  • unitary_kraus and general_kraus API now supports the mix input of array and Node as kraus list

0.0.220318

Added

  • add gradient free scipy interface for optimization

  • add qiskit circuit to tensorcircuit circuit methods

  • add draw method on circuit from qiskit transform pipeline

Changed

  • futher refactor VQNHE code in applications

  • add alias sample for perfect_sampling method

  • optimize VQNHE pipeline for a more stable training loop (breaking changes in some APIs)

Fixed

  • Circuit inputs will convert to tensor first

0.0.220311

Added

  • add sigmoid method on backends

  • add MPO expectation template function for MPO evaluation on circuit

  • add operator_expectation in templates.measurements for a unified expectation interface

  • add templates.chems module for interface between tc and openfermion on quantum chemistry related tasks

  • add tc.Circuit to Qiskit QuantumCircuit transformation

Fixed

  • fix the bug in QuOperator.from_local_tensor where the dtype should always be in numpy context

  • fix MPO copy when apply MPO gate on the circuit

Changed

  • allow multi-qubit gate in multicontrol gate

0.0.220301

Added

  • new universal contraction analyse tools and pseudo contraction rehearsals for debug

  • add gather1d method on backends for 1d tensor indexing

  • add dataset module in template submodule for dataset preprocessing and embedding

  • MPO format quantum gate is natively support now

  • add multicontrol gates in MPO format

Fixed

  • fixed real operation on some methods in templates.measurements

Changed

  • add gatef key in circuit IR dict for the gate function, while replace gate key with the gate node or MPO (breaking change)

0.0.220126

Added

  • add td and sd gates for dagger version of T gate and S gate

  • add argmax and argmin as backend methods

  • add expectation_before methods for tc.Circuit for further manipulation on the tensornetwork

Changed

  • refined repr for tc.gates.Gate

  • expectation API now supports int index besides list indexes

Fixed

  • make consistent Gate return for channels

  • fixed bug on list optimizer for contraction

  • stability for QR operator in terms of automatic differentiation

0.0.220118

Added

  • add hessian method on backends

  • add further automatic pipelines for visualization by generating pdf or images

  • add reshape2 method on backends as a short cut to reshape a tensor with all legs 2-d

  • add reshapem method on backends to reshape any tensor as a square matrix

  • add controlled and ocontrolled API to generate more gates

  • add crx, cry, crz gate on Circuit

  • add __repr__ and __str__ for backend object

  • tc.expectation now support ket arg as quvector form

Fixed

  • sizen correctly returns 1 for tensor of no shape

  • fixed convert_to_tensor bug in numpy backend in TensorNetwork

  • any_gate also support Gate format instead of matrix

  • matrix_for_gate works now for backends more than numpy

Changed

  • expectation API now also accepts plain tensor instead of tc.Gate.

  • DMCircuit and DMCircuit2 are all pointing the efficent implementations (breaking changes)

0.0.220106

Added

  • add solve method on backends to solve linear equations

  • add full quantum natural gradient examples and qng method in experimental module

  • add concat method to backends

  • add stop_gradient method to backends

  • add has_aux arg on vvag method

  • add imag method on backends

  • add Circuit.vis_tex interface that returns the quantikz circuit latex

Changed

  • contractor, dtype and backend set are default to return objects, with tc.runtime_backend("jax") as K or K = tc.set_backend("jax") could work

  • change perfect_sampling to use measure_jit behind the scene

  • anygate automatically reshape the unitary input to 2-d leg for users' good

  • quantum.renyi_entropy computation with correct prefactor

  • Circuit gate can provided other names by name attr

  • example_block support param auto reshape for users' good

Fixed

  • make four algorithms for quantum natural gradient consistent and correct

  • torch real is now a real

0.0.211223

Added

  • add quantum.heisenberg_hamiltonian for hamiltonian generation shortcut

  • add has_aux parameter in backend methods grad and value_and_grad, the semantic syntax is the same as jax

  • add optimizer class on tensorflow and jax backend, so that a minimal and unified backend agnostic optimizer interface is provided

  • add quantum.mutual_information, add support on mixed state for quantum.reduced_density_matrix

  • add jvp methods for tensorflow, jax, torch backends, and ensure pytree support in jvp and vjp interfaces for tensorflow and jax backends; also ensure complex support for jvp and vjp

  • add jacfwd and jacrev for backend methods (experimental API, may have bugs and subject to changes)

Fixed

  • fix matmul bug on tensornetwork tensorflow backend

Changed

  • delete qcode IR for Circuit, use qir instead (breaking changes)

  • basic circuit running is ok on pytorch backend with some complex support fixing

0.0.211216

Added

  • add get_random_state and random_split methods to backends

  • add qir representation of circuit, c.to_qir() and Circuit.from_qir() methods

  • fine-grained control on backend, dtype and contractor setup: tc.set_function_backend() for function level decorator and tc.runtime_backend() as with context manager

  • add state_centric decorator in tc.templates.blocks to transform circuit-to-circuit funtion to state-to-state function

  • add interfaces.scipy_optimize_interface to transform quantum function into scipy.optimize.minimize campatible form

Fixed

  • avoid error on watch non tf.Tensor in tensorflow backend grad method

  • circuit preprocessing simplification with only single qubit gates

  • avoid the bug when random from jax backend with jitted function

  • refresh the state cache in Circuit when new gate is applied

Changed

  • refactor tc.gates (breaking API on rgate -> r_gate, iswapgate -> iswap_gate)