diff --git a/.gitattributes b/.gitattributes index 26868d10b97f1c6961a627754ef4e39eeac528bf..aae142219378ae82f9bf1cbe423684ba540bddf6 100644 --- a/.gitattributes +++ b/.gitattributes @@ -1635,3 +1635,6 @@ parrot/lib/python3.10/site-packages/scipy/special/_ellip_harm_2.cpython-310-x86_ parrot/lib/python3.10/site-packages/pandas/_libs/lib.cpython-310-x86_64-linux-gnu.so filter=lfs diff=lfs merge=lfs -text wemm/lib/python3.10/site-packages/nvidia/cuda_nvrtc/lib/libnvrtc.so.11.2 filter=lfs diff=lfs merge=lfs -text vllm/lib/python3.10/site-packages/fastrlock/rlock.cpython-310-x86_64-linux-gnu.so filter=lfs diff=lfs merge=lfs -text +parrot/lib/python3.10/site-packages/scipy/interpolate/interpnd.cpython-310-x86_64-linux-gnu.so filter=lfs diff=lfs merge=lfs -text +vllm/lib/python3.10/site-packages/transformers/generation/__pycache__/tf_utils.cpython-310.pyc filter=lfs diff=lfs merge=lfs -text +vllm/lib/python3.10/site-packages/transformers/generation/__pycache__/logits_process.cpython-310.pyc filter=lfs diff=lfs merge=lfs -text diff --git a/parrot/lib/python3.10/site-packages/scipy/interpolate/interpnd.cpython-310-x86_64-linux-gnu.so b/parrot/lib/python3.10/site-packages/scipy/interpolate/interpnd.cpython-310-x86_64-linux-gnu.so new file mode 100644 index 0000000000000000000000000000000000000000..e1c425496245b57d4b9d2cbfa532224b3f42fc36 --- /dev/null +++ b/parrot/lib/python3.10/site-packages/scipy/interpolate/interpnd.cpython-310-x86_64-linux-gnu.so @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:858919e18e295550cd7be75e89651b5d12ad97a9e2d3c540da006b42c42ac19c +size 458776 diff --git a/parrot/lib/python3.10/site-packages/scipy/linalg/tests/data/carex_20_data.npz b/parrot/lib/python3.10/site-packages/scipy/linalg/tests/data/carex_20_data.npz new file mode 100644 index 0000000000000000000000000000000000000000..87266deb46238307347362b63a4878f2565baf56 --- /dev/null +++ b/parrot/lib/python3.10/site-packages/scipy/linalg/tests/data/carex_20_data.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:14e222d34a7118c7284a1675c6feceee77b84df951a5c6ba2a5ee9ff3054fa1d +size 31231 diff --git a/parrot/lib/python3.10/site-packages/scipy/linalg/tests/data/gendare_20170120_data.npz b/parrot/lib/python3.10/site-packages/scipy/linalg/tests/data/gendare_20170120_data.npz new file mode 100644 index 0000000000000000000000000000000000000000..ff967f2ca0d0868aacf7d7e67402599e64bab817 --- /dev/null +++ b/parrot/lib/python3.10/site-packages/scipy/linalg/tests/data/gendare_20170120_data.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a3dfab451d9d5c20243e0ed85cd8b6c9657669fb9a0f83b5be165585783d55b5 +size 2164 diff --git a/parrot/lib/python3.10/site-packages/scipy/optimize/_linprog_highs.py b/parrot/lib/python3.10/site-packages/scipy/optimize/_linprog_highs.py new file mode 100644 index 0000000000000000000000000000000000000000..eb07443bb255471e6e0ac487bd6749253bf5d133 --- /dev/null +++ b/parrot/lib/python3.10/site-packages/scipy/optimize/_linprog_highs.py @@ -0,0 +1,440 @@ +"""HiGHS Linear Optimization Methods + +Interface to HiGHS linear optimization software. +https://highs.dev/ + +.. versionadded:: 1.5.0 + +References +---------- +.. [1] Q. Huangfu and J.A.J. Hall. "Parallelizing the dual revised simplex + method." Mathematical Programming Computation, 10 (1), 119-142, + 2018. DOI: 10.1007/s12532-017-0130-5 + +""" + +import inspect +import numpy as np +from ._optimize import OptimizeWarning, OptimizeResult +from warnings import warn +from ._highs._highs_wrapper import _highs_wrapper +from ._highs._highs_constants import ( + CONST_INF, + MESSAGE_LEVEL_NONE, + HIGHS_OBJECTIVE_SENSE_MINIMIZE, + + MODEL_STATUS_NOTSET, + MODEL_STATUS_LOAD_ERROR, + MODEL_STATUS_MODEL_ERROR, + MODEL_STATUS_PRESOLVE_ERROR, + MODEL_STATUS_SOLVE_ERROR, + MODEL_STATUS_POSTSOLVE_ERROR, + MODEL_STATUS_MODEL_EMPTY, + MODEL_STATUS_OPTIMAL, + MODEL_STATUS_INFEASIBLE, + MODEL_STATUS_UNBOUNDED_OR_INFEASIBLE, + MODEL_STATUS_UNBOUNDED, + MODEL_STATUS_REACHED_DUAL_OBJECTIVE_VALUE_UPPER_BOUND + as MODEL_STATUS_RDOVUB, + MODEL_STATUS_REACHED_OBJECTIVE_TARGET, + MODEL_STATUS_REACHED_TIME_LIMIT, + MODEL_STATUS_REACHED_ITERATION_LIMIT, + + HIGHS_SIMPLEX_STRATEGY_DUAL, + + HIGHS_SIMPLEX_CRASH_STRATEGY_OFF, + + HIGHS_SIMPLEX_EDGE_WEIGHT_STRATEGY_CHOOSE, + HIGHS_SIMPLEX_EDGE_WEIGHT_STRATEGY_DANTZIG, + HIGHS_SIMPLEX_EDGE_WEIGHT_STRATEGY_DEVEX, + HIGHS_SIMPLEX_EDGE_WEIGHT_STRATEGY_STEEPEST_EDGE, +) +from scipy.sparse import csc_matrix, vstack, issparse + + +def _highs_to_scipy_status_message(highs_status, highs_message): + """Converts HiGHS status number/message to SciPy status number/message""" + + scipy_statuses_messages = { + None: (4, "HiGHS did not provide a status code. "), + MODEL_STATUS_NOTSET: (4, ""), + MODEL_STATUS_LOAD_ERROR: (4, ""), + MODEL_STATUS_MODEL_ERROR: (2, ""), + MODEL_STATUS_PRESOLVE_ERROR: (4, ""), + MODEL_STATUS_SOLVE_ERROR: (4, ""), + MODEL_STATUS_POSTSOLVE_ERROR: (4, ""), + MODEL_STATUS_MODEL_EMPTY: (4, ""), + MODEL_STATUS_RDOVUB: (4, ""), + MODEL_STATUS_REACHED_OBJECTIVE_TARGET: (4, ""), + MODEL_STATUS_OPTIMAL: (0, "Optimization terminated successfully. "), + MODEL_STATUS_REACHED_TIME_LIMIT: (1, "Time limit reached. "), + MODEL_STATUS_REACHED_ITERATION_LIMIT: (1, "Iteration limit reached. "), + MODEL_STATUS_INFEASIBLE: (2, "The problem is infeasible. "), + MODEL_STATUS_UNBOUNDED: (3, "The problem is unbounded. "), + MODEL_STATUS_UNBOUNDED_OR_INFEASIBLE: (4, "The problem is unbounded " + "or infeasible. ")} + unrecognized = (4, "The HiGHS status code was not recognized. ") + scipy_status, scipy_message = ( + scipy_statuses_messages.get(highs_status, unrecognized)) + scipy_message = (f"{scipy_message}" + f"(HiGHS Status {highs_status}: {highs_message})") + return scipy_status, scipy_message + + +def _replace_inf(x): + # Replace `np.inf` with CONST_INF + infs = np.isinf(x) + with np.errstate(invalid="ignore"): + x[infs] = np.sign(x[infs])*CONST_INF + return x + + +def _convert_to_highs_enum(option, option_str, choices): + # If option is in the choices we can look it up, if not use + # the default value taken from function signature and warn: + try: + return choices[option.lower()] + except AttributeError: + return choices[option] + except KeyError: + sig = inspect.signature(_linprog_highs) + default_str = sig.parameters[option_str].default + warn(f"Option {option_str} is {option}, but only values in " + f"{set(choices.keys())} are allowed. Using default: " + f"{default_str}.", + OptimizeWarning, stacklevel=3) + return choices[default_str] + + +def _linprog_highs(lp, solver, time_limit=None, presolve=True, + disp=False, maxiter=None, + dual_feasibility_tolerance=None, + primal_feasibility_tolerance=None, + ipm_optimality_tolerance=None, + simplex_dual_edge_weight_strategy=None, + mip_rel_gap=None, + mip_max_nodes=None, + **unknown_options): + r""" + Solve the following linear programming problem using one of the HiGHS + solvers: + + User-facing documentation is in _linprog_doc.py. + + Parameters + ---------- + lp : _LPProblem + A ``scipy.optimize._linprog_util._LPProblem`` ``namedtuple``. + solver : "ipm" or "simplex" or None + Which HiGHS solver to use. If ``None``, "simplex" will be used. + + Options + ------- + maxiter : int + The maximum number of iterations to perform in either phase. For + ``solver='ipm'``, this does not include the number of crossover + iterations. Default is the largest possible value for an ``int`` + on the platform. + disp : bool + Set to ``True`` if indicators of optimization status are to be printed + to the console each iteration; default ``False``. + time_limit : float + The maximum time in seconds allotted to solve the problem; default is + the largest possible value for a ``double`` on the platform. + presolve : bool + Presolve attempts to identify trivial infeasibilities, + identify trivial unboundedness, and simplify the problem before + sending it to the main solver. It is generally recommended + to keep the default setting ``True``; set to ``False`` if presolve is + to be disabled. + dual_feasibility_tolerance : double + Dual feasibility tolerance. Default is 1e-07. + The minimum of this and ``primal_feasibility_tolerance`` + is used for the feasibility tolerance when ``solver='ipm'``. + primal_feasibility_tolerance : double + Primal feasibility tolerance. Default is 1e-07. + The minimum of this and ``dual_feasibility_tolerance`` + is used for the feasibility tolerance when ``solver='ipm'``. + ipm_optimality_tolerance : double + Optimality tolerance for ``solver='ipm'``. Default is 1e-08. + Minimum possible value is 1e-12 and must be smaller than the largest + possible value for a ``double`` on the platform. + simplex_dual_edge_weight_strategy : str (default: None) + Strategy for simplex dual edge weights. The default, ``None``, + automatically selects one of the following. + + ``'dantzig'`` uses Dantzig's original strategy of choosing the most + negative reduced cost. + + ``'devex'`` uses the strategy described in [15]_. + + ``steepest`` uses the exact steepest edge strategy as described in + [16]_. + + ``'steepest-devex'`` begins with the exact steepest edge strategy + until the computation is too costly or inexact and then switches to + the devex method. + + Currently, using ``None`` always selects ``'steepest-devex'``, but this + may change as new options become available. + + mip_max_nodes : int + The maximum number of nodes allotted to solve the problem; default is + the largest possible value for a ``HighsInt`` on the platform. + Ignored if not using the MIP solver. + unknown_options : dict + Optional arguments not used by this particular solver. If + ``unknown_options`` is non-empty, a warning is issued listing all + unused options. + + Returns + ------- + sol : dict + A dictionary consisting of the fields: + + x : 1D array + The values of the decision variables that minimizes the + objective function while satisfying the constraints. + fun : float + The optimal value of the objective function ``c @ x``. + slack : 1D array + The (nominally positive) values of the slack, + ``b_ub - A_ub @ x``. + con : 1D array + The (nominally zero) residuals of the equality constraints, + ``b_eq - A_eq @ x``. + success : bool + ``True`` when the algorithm succeeds in finding an optimal + solution. + status : int + An integer representing the exit status of the algorithm. + + ``0`` : Optimization terminated successfully. + + ``1`` : Iteration or time limit reached. + + ``2`` : Problem appears to be infeasible. + + ``3`` : Problem appears to be unbounded. + + ``4`` : The HiGHS solver ran into a problem. + + message : str + A string descriptor of the exit status of the algorithm. + nit : int + The total number of iterations performed. + For ``solver='simplex'``, this includes iterations in all + phases. For ``solver='ipm'``, this does not include + crossover iterations. + crossover_nit : int + The number of primal/dual pushes performed during the + crossover routine for ``solver='ipm'``. This is ``0`` + for ``solver='simplex'``. + ineqlin : OptimizeResult + Solution and sensitivity information corresponding to the + inequality constraints, `b_ub`. A dictionary consisting of the + fields: + + residual : np.ndnarray + The (nominally positive) values of the slack variables, + ``b_ub - A_ub @ x``. This quantity is also commonly + referred to as "slack". + + marginals : np.ndarray + The sensitivity (partial derivative) of the objective + function with respect to the right-hand side of the + inequality constraints, `b_ub`. + + eqlin : OptimizeResult + Solution and sensitivity information corresponding to the + equality constraints, `b_eq`. A dictionary consisting of the + fields: + + residual : np.ndarray + The (nominally zero) residuals of the equality constraints, + ``b_eq - A_eq @ x``. + + marginals : np.ndarray + The sensitivity (partial derivative) of the objective + function with respect to the right-hand side of the + equality constraints, `b_eq`. + + lower, upper : OptimizeResult + Solution and sensitivity information corresponding to the + lower and upper bounds on decision variables, `bounds`. + + residual : np.ndarray + The (nominally positive) values of the quantity + ``x - lb`` (lower) or ``ub - x`` (upper). + + marginals : np.ndarray + The sensitivity (partial derivative) of the objective + function with respect to the lower and upper + `bounds`. + + mip_node_count : int + The number of subproblems or "nodes" solved by the MILP + solver. Only present when `integrality` is not `None`. + + mip_dual_bound : float + The MILP solver's final estimate of the lower bound on the + optimal solution. Only present when `integrality` is not + `None`. + + mip_gap : float + The difference between the final objective function value + and the final dual bound, scaled by the final objective + function value. Only present when `integrality` is not + `None`. + + Notes + ----- + The result fields `ineqlin`, `eqlin`, `lower`, and `upper` all contain + `marginals`, or partial derivatives of the objective function with respect + to the right-hand side of each constraint. These partial derivatives are + also referred to as "Lagrange multipliers", "dual values", and + "shadow prices". The sign convention of `marginals` is opposite that + of Lagrange multipliers produced by many nonlinear solvers. + + References + ---------- + .. [15] Harris, Paula MJ. "Pivot selection methods of the Devex LP code." + Mathematical programming 5.1 (1973): 1-28. + .. [16] Goldfarb, Donald, and John Ker Reid. "A practicable steepest-edge + simplex algorithm." Mathematical Programming 12.1 (1977): 361-371. + """ + if unknown_options: + message = (f"Unrecognized options detected: {unknown_options}. " + "These will be passed to HiGHS verbatim.") + warn(message, OptimizeWarning, stacklevel=3) + + # Map options to HiGHS enum values + simplex_dual_edge_weight_strategy_enum = _convert_to_highs_enum( + simplex_dual_edge_weight_strategy, + 'simplex_dual_edge_weight_strategy', + choices={'dantzig': HIGHS_SIMPLEX_EDGE_WEIGHT_STRATEGY_DANTZIG, + 'devex': HIGHS_SIMPLEX_EDGE_WEIGHT_STRATEGY_DEVEX, + 'steepest-devex': HIGHS_SIMPLEX_EDGE_WEIGHT_STRATEGY_CHOOSE, + 'steepest': + HIGHS_SIMPLEX_EDGE_WEIGHT_STRATEGY_STEEPEST_EDGE, + None: None}) + + c, A_ub, b_ub, A_eq, b_eq, bounds, x0, integrality = lp + + lb, ub = bounds.T.copy() # separate bounds, copy->C-cntgs + # highs_wrapper solves LHS <= A*x <= RHS, not equality constraints + with np.errstate(invalid="ignore"): + lhs_ub = -np.ones_like(b_ub)*np.inf # LHS of UB constraints is -inf + rhs_ub = b_ub # RHS of UB constraints is b_ub + lhs_eq = b_eq # Equality constraint is inequality + rhs_eq = b_eq # constraint with LHS=RHS + lhs = np.concatenate((lhs_ub, lhs_eq)) + rhs = np.concatenate((rhs_ub, rhs_eq)) + + if issparse(A_ub) or issparse(A_eq): + A = vstack((A_ub, A_eq)) + else: + A = np.vstack((A_ub, A_eq)) + A = csc_matrix(A) + + options = { + 'presolve': presolve, + 'sense': HIGHS_OBJECTIVE_SENSE_MINIMIZE, + 'solver': solver, + 'time_limit': time_limit, + 'highs_debug_level': MESSAGE_LEVEL_NONE, + 'dual_feasibility_tolerance': dual_feasibility_tolerance, + 'ipm_optimality_tolerance': ipm_optimality_tolerance, + 'log_to_console': disp, + 'mip_max_nodes': mip_max_nodes, + 'output_flag': disp, + 'primal_feasibility_tolerance': primal_feasibility_tolerance, + 'simplex_dual_edge_weight_strategy': + simplex_dual_edge_weight_strategy_enum, + 'simplex_strategy': HIGHS_SIMPLEX_STRATEGY_DUAL, + 'simplex_crash_strategy': HIGHS_SIMPLEX_CRASH_STRATEGY_OFF, + 'ipm_iteration_limit': maxiter, + 'simplex_iteration_limit': maxiter, + 'mip_rel_gap': mip_rel_gap, + } + options.update(unknown_options) + + # np.inf doesn't work; use very large constant + rhs = _replace_inf(rhs) + lhs = _replace_inf(lhs) + lb = _replace_inf(lb) + ub = _replace_inf(ub) + + if integrality is None or np.sum(integrality) == 0: + integrality = np.empty(0) + else: + integrality = np.array(integrality) + + res = _highs_wrapper(c, A.indptr, A.indices, A.data, lhs, rhs, + lb, ub, integrality.astype(np.uint8), options) + + # HiGHS represents constraints as lhs/rhs, so + # Ax + s = b => Ax = b - s + # and we need to split up s by A_ub and A_eq + if 'slack' in res: + slack = res['slack'] + con = np.array(slack[len(b_ub):]) + slack = np.array(slack[:len(b_ub)]) + else: + slack, con = None, None + + # lagrange multipliers for equalities/inequalities and upper/lower bounds + if 'lambda' in res: + lamda = res['lambda'] + marg_ineqlin = np.array(lamda[:len(b_ub)]) + marg_eqlin = np.array(lamda[len(b_ub):]) + marg_upper = np.array(res['marg_bnds'][1, :]) + marg_lower = np.array(res['marg_bnds'][0, :]) + else: + marg_ineqlin, marg_eqlin = None, None + marg_upper, marg_lower = None, None + + # this needs to be updated if we start choosing the solver intelligently + + # Convert to scipy-style status and message + highs_status = res.get('status', None) + highs_message = res.get('message', None) + status, message = _highs_to_scipy_status_message(highs_status, + highs_message) + + x = np.array(res['x']) if 'x' in res else None + sol = {'x': x, + 'slack': slack, + 'con': con, + 'ineqlin': OptimizeResult({ + 'residual': slack, + 'marginals': marg_ineqlin, + }), + 'eqlin': OptimizeResult({ + 'residual': con, + 'marginals': marg_eqlin, + }), + 'lower': OptimizeResult({ + 'residual': None if x is None else x - lb, + 'marginals': marg_lower, + }), + 'upper': OptimizeResult({ + 'residual': None if x is None else ub - x, + 'marginals': marg_upper + }), + 'fun': res.get('fun'), + 'status': status, + 'success': res['status'] == MODEL_STATUS_OPTIMAL, + 'message': message, + 'nit': res.get('simplex_nit', 0) or res.get('ipm_nit', 0), + 'crossover_nit': res.get('crossover_nit'), + } + + if np.any(x) and integrality is not None: + sol.update({ + 'mip_node_count': res.get('mip_node_count', 0), + 'mip_dual_bound': res.get('mip_dual_bound', 0.0), + 'mip_gap': res.get('mip_gap', 0.0), + }) + + return sol diff --git a/parrot/lib/python3.10/site-packages/scipy/optimize/_lsq/__pycache__/bvls.cpython-310.pyc b/parrot/lib/python3.10/site-packages/scipy/optimize/_lsq/__pycache__/bvls.cpython-310.pyc new file mode 100644 index 0000000000000000000000000000000000000000..5d532ffe22c4a7c3715d27160a442ba326f889a2 Binary files /dev/null and b/parrot/lib/python3.10/site-packages/scipy/optimize/_lsq/__pycache__/bvls.cpython-310.pyc differ diff --git a/parrot/lib/python3.10/site-packages/scipy/optimize/_lsq/__pycache__/trf_linear.cpython-310.pyc b/parrot/lib/python3.10/site-packages/scipy/optimize/_lsq/__pycache__/trf_linear.cpython-310.pyc new file mode 100644 index 0000000000000000000000000000000000000000..bd0731110a24c2744817db531684880d419199eb Binary files /dev/null and b/parrot/lib/python3.10/site-packages/scipy/optimize/_lsq/__pycache__/trf_linear.cpython-310.pyc differ diff --git a/parrot/lib/python3.10/site-packages/scipy/optimize/_lsq/bvls.py b/parrot/lib/python3.10/site-packages/scipy/optimize/_lsq/bvls.py new file mode 100644 index 0000000000000000000000000000000000000000..8f34ead4a1fc4edbb3c2ab50a204aa9a3cc21cff --- /dev/null +++ b/parrot/lib/python3.10/site-packages/scipy/optimize/_lsq/bvls.py @@ -0,0 +1,183 @@ +"""Bounded-variable least-squares algorithm.""" +import numpy as np +from numpy.linalg import norm, lstsq +from scipy.optimize import OptimizeResult + +from .common import print_header_linear, print_iteration_linear + + +def compute_kkt_optimality(g, on_bound): + """Compute the maximum violation of KKT conditions.""" + g_kkt = g * on_bound + free_set = on_bound == 0 + g_kkt[free_set] = np.abs(g[free_set]) + return np.max(g_kkt) + + +def bvls(A, b, x_lsq, lb, ub, tol, max_iter, verbose, rcond=None): + m, n = A.shape + + x = x_lsq.copy() + on_bound = np.zeros(n) + + mask = x <= lb + x[mask] = lb[mask] + on_bound[mask] = -1 + + mask = x >= ub + x[mask] = ub[mask] + on_bound[mask] = 1 + + free_set = on_bound == 0 + active_set = ~free_set + free_set, = np.nonzero(free_set) + + r = A.dot(x) - b + cost = 0.5 * np.dot(r, r) + initial_cost = cost + g = A.T.dot(r) + + cost_change = None + step_norm = None + iteration = 0 + + if verbose == 2: + print_header_linear() + + # This is the initialization loop. The requirement is that the + # least-squares solution on free variables is feasible before BVLS starts. + # One possible initialization is to set all variables to lower or upper + # bounds, but many iterations may be required from this state later on. + # The implemented ad-hoc procedure which intuitively should give a better + # initial state: find the least-squares solution on current free variables, + # if its feasible then stop, otherwise, set violating variables to + # corresponding bounds and continue on the reduced set of free variables. + + while free_set.size > 0: + if verbose == 2: + optimality = compute_kkt_optimality(g, on_bound) + print_iteration_linear(iteration, cost, cost_change, step_norm, + optimality) + + iteration += 1 + x_free_old = x[free_set].copy() + + A_free = A[:, free_set] + b_free = b - A.dot(x * active_set) + z = lstsq(A_free, b_free, rcond=rcond)[0] + + lbv = z < lb[free_set] + ubv = z > ub[free_set] + v = lbv | ubv + + if np.any(lbv): + ind = free_set[lbv] + x[ind] = lb[ind] + active_set[ind] = True + on_bound[ind] = -1 + + if np.any(ubv): + ind = free_set[ubv] + x[ind] = ub[ind] + active_set[ind] = True + on_bound[ind] = 1 + + ind = free_set[~v] + x[ind] = z[~v] + + r = A.dot(x) - b + cost_new = 0.5 * np.dot(r, r) + cost_change = cost - cost_new + cost = cost_new + g = A.T.dot(r) + step_norm = norm(x[free_set] - x_free_old) + + if np.any(v): + free_set = free_set[~v] + else: + break + + if max_iter is None: + max_iter = n + max_iter += iteration + + termination_status = None + + # Main BVLS loop. + + optimality = compute_kkt_optimality(g, on_bound) + for iteration in range(iteration, max_iter): # BVLS Loop A + if verbose == 2: + print_iteration_linear(iteration, cost, cost_change, + step_norm, optimality) + + if optimality < tol: + termination_status = 1 + + if termination_status is not None: + break + + move_to_free = np.argmax(g * on_bound) + on_bound[move_to_free] = 0 + + while True: # BVLS Loop B + + free_set = on_bound == 0 + active_set = ~free_set + free_set, = np.nonzero(free_set) + + x_free = x[free_set] + x_free_old = x_free.copy() + lb_free = lb[free_set] + ub_free = ub[free_set] + + A_free = A[:, free_set] + b_free = b - A.dot(x * active_set) + z = lstsq(A_free, b_free, rcond=rcond)[0] + + lbv, = np.nonzero(z < lb_free) + ubv, = np.nonzero(z > ub_free) + v = np.hstack((lbv, ubv)) + + if v.size > 0: + alphas = np.hstack(( + lb_free[lbv] - x_free[lbv], + ub_free[ubv] - x_free[ubv])) / (z[v] - x_free[v]) + + i = np.argmin(alphas) + i_free = v[i] + alpha = alphas[i] + + x_free *= 1 - alpha + x_free += alpha * z + x[free_set] = x_free + + if i < lbv.size: + on_bound[free_set[i_free]] = -1 + else: + on_bound[free_set[i_free]] = 1 + else: + x_free = z + x[free_set] = x_free + break + + step_norm = norm(x_free - x_free_old) + + r = A.dot(x) - b + cost_new = 0.5 * np.dot(r, r) + cost_change = cost - cost_new + + if cost_change < tol * cost: + termination_status = 2 + cost = cost_new + + g = A.T.dot(r) + optimality = compute_kkt_optimality(g, on_bound) + + if termination_status is None: + termination_status = 0 + + return OptimizeResult( + x=x, fun=r, cost=cost, optimality=optimality, active_mask=on_bound, + nit=iteration + 1, status=termination_status, + initial_cost=initial_cost) diff --git a/parrot/lib/python3.10/site-packages/scipy/optimize/_lsq/common.py b/parrot/lib/python3.10/site-packages/scipy/optimize/_lsq/common.py new file mode 100644 index 0000000000000000000000000000000000000000..995c3b64ea64670463083ae41ba038f61338cdef --- /dev/null +++ b/parrot/lib/python3.10/site-packages/scipy/optimize/_lsq/common.py @@ -0,0 +1,733 @@ +"""Functions used by least-squares algorithms.""" +from math import copysign + +import numpy as np +from numpy.linalg import norm + +from scipy.linalg import cho_factor, cho_solve, LinAlgError +from scipy.sparse import issparse +from scipy.sparse.linalg import LinearOperator, aslinearoperator + + +EPS = np.finfo(float).eps + + +# Functions related to a trust-region problem. + + +def intersect_trust_region(x, s, Delta): + """Find the intersection of a line with the boundary of a trust region. + + This function solves the quadratic equation with respect to t + ||(x + s*t)||**2 = Delta**2. + + Returns + ------- + t_neg, t_pos : tuple of float + Negative and positive roots. + + Raises + ------ + ValueError + If `s` is zero or `x` is not within the trust region. + """ + a = np.dot(s, s) + if a == 0: + raise ValueError("`s` is zero.") + + b = np.dot(x, s) + + c = np.dot(x, x) - Delta**2 + if c > 0: + raise ValueError("`x` is not within the trust region.") + + d = np.sqrt(b*b - a*c) # Root from one fourth of the discriminant. + + # Computations below avoid loss of significance, see "Numerical Recipes". + q = -(b + copysign(d, b)) + t1 = q / a + t2 = c / q + + if t1 < t2: + return t1, t2 + else: + return t2, t1 + + +def solve_lsq_trust_region(n, m, uf, s, V, Delta, initial_alpha=None, + rtol=0.01, max_iter=10): + """Solve a trust-region problem arising in least-squares minimization. + + This function implements a method described by J. J. More [1]_ and used + in MINPACK, but it relies on a single SVD of Jacobian instead of series + of Cholesky decompositions. Before running this function, compute: + ``U, s, VT = svd(J, full_matrices=False)``. + + Parameters + ---------- + n : int + Number of variables. + m : int + Number of residuals. + uf : ndarray + Computed as U.T.dot(f). + s : ndarray + Singular values of J. + V : ndarray + Transpose of VT. + Delta : float + Radius of a trust region. + initial_alpha : float, optional + Initial guess for alpha, which might be available from a previous + iteration. If None, determined automatically. + rtol : float, optional + Stopping tolerance for the root-finding procedure. Namely, the + solution ``p`` will satisfy ``abs(norm(p) - Delta) < rtol * Delta``. + max_iter : int, optional + Maximum allowed number of iterations for the root-finding procedure. + + Returns + ------- + p : ndarray, shape (n,) + Found solution of a trust-region problem. + alpha : float + Positive value such that (J.T*J + alpha*I)*p = -J.T*f. + Sometimes called Levenberg-Marquardt parameter. + n_iter : int + Number of iterations made by root-finding procedure. Zero means + that Gauss-Newton step was selected as the solution. + + References + ---------- + .. [1] More, J. J., "The Levenberg-Marquardt Algorithm: Implementation + and Theory," Numerical Analysis, ed. G. A. Watson, Lecture Notes + in Mathematics 630, Springer Verlag, pp. 105-116, 1977. + """ + def phi_and_derivative(alpha, suf, s, Delta): + """Function of which to find zero. + + It is defined as "norm of regularized (by alpha) least-squares + solution minus `Delta`". Refer to [1]_. + """ + denom = s**2 + alpha + p_norm = norm(suf / denom) + phi = p_norm - Delta + phi_prime = -np.sum(suf ** 2 / denom**3) / p_norm + return phi, phi_prime + + suf = s * uf + + # Check if J has full rank and try Gauss-Newton step. + if m >= n: + threshold = EPS * m * s[0] + full_rank = s[-1] > threshold + else: + full_rank = False + + if full_rank: + p = -V.dot(uf / s) + if norm(p) <= Delta: + return p, 0.0, 0 + + alpha_upper = norm(suf) / Delta + + if full_rank: + phi, phi_prime = phi_and_derivative(0.0, suf, s, Delta) + alpha_lower = -phi / phi_prime + else: + alpha_lower = 0.0 + + if initial_alpha is None or not full_rank and initial_alpha == 0: + alpha = max(0.001 * alpha_upper, (alpha_lower * alpha_upper)**0.5) + else: + alpha = initial_alpha + + for it in range(max_iter): + if alpha < alpha_lower or alpha > alpha_upper: + alpha = max(0.001 * alpha_upper, (alpha_lower * alpha_upper)**0.5) + + phi, phi_prime = phi_and_derivative(alpha, suf, s, Delta) + + if phi < 0: + alpha_upper = alpha + + ratio = phi / phi_prime + alpha_lower = max(alpha_lower, alpha - ratio) + alpha -= (phi + Delta) * ratio / Delta + + if np.abs(phi) < rtol * Delta: + break + + p = -V.dot(suf / (s**2 + alpha)) + + # Make the norm of p equal to Delta, p is changed only slightly during + # this. It is done to prevent p lie outside the trust region (which can + # cause problems later). + p *= Delta / norm(p) + + return p, alpha, it + 1 + + +def solve_trust_region_2d(B, g, Delta): + """Solve a general trust-region problem in 2 dimensions. + + The problem is reformulated as a 4th order algebraic equation, + the solution of which is found by numpy.roots. + + Parameters + ---------- + B : ndarray, shape (2, 2) + Symmetric matrix, defines a quadratic term of the function. + g : ndarray, shape (2,) + Defines a linear term of the function. + Delta : float + Radius of a trust region. + + Returns + ------- + p : ndarray, shape (2,) + Found solution. + newton_step : bool + Whether the returned solution is the Newton step which lies within + the trust region. + """ + try: + R, lower = cho_factor(B) + p = -cho_solve((R, lower), g) + if np.dot(p, p) <= Delta**2: + return p, True + except LinAlgError: + pass + + a = B[0, 0] * Delta**2 + b = B[0, 1] * Delta**2 + c = B[1, 1] * Delta**2 + + d = g[0] * Delta + f = g[1] * Delta + + coeffs = np.array( + [-b + d, 2 * (a - c + f), 6 * b, 2 * (-a + c + f), -b - d]) + t = np.roots(coeffs) # Can handle leading zeros. + t = np.real(t[np.isreal(t)]) + + p = Delta * np.vstack((2 * t / (1 + t**2), (1 - t**2) / (1 + t**2))) + value = 0.5 * np.sum(p * B.dot(p), axis=0) + np.dot(g, p) + i = np.argmin(value) + p = p[:, i] + + return p, False + + +def update_tr_radius(Delta, actual_reduction, predicted_reduction, + step_norm, bound_hit): + """Update the radius of a trust region based on the cost reduction. + + Returns + ------- + Delta : float + New radius. + ratio : float + Ratio between actual and predicted reductions. + """ + if predicted_reduction > 0: + ratio = actual_reduction / predicted_reduction + elif predicted_reduction == actual_reduction == 0: + ratio = 1 + else: + ratio = 0 + + if ratio < 0.25: + Delta = 0.25 * step_norm + elif ratio > 0.75 and bound_hit: + Delta *= 2.0 + + return Delta, ratio + + +# Construction and minimization of quadratic functions. + + +def build_quadratic_1d(J, g, s, diag=None, s0=None): + """Parameterize a multivariate quadratic function along a line. + + The resulting univariate quadratic function is given as follows:: + + f(t) = 0.5 * (s0 + s*t).T * (J.T*J + diag) * (s0 + s*t) + + g.T * (s0 + s*t) + + Parameters + ---------- + J : ndarray, sparse matrix or LinearOperator shape (m, n) + Jacobian matrix, affects the quadratic term. + g : ndarray, shape (n,) + Gradient, defines the linear term. + s : ndarray, shape (n,) + Direction vector of a line. + diag : None or ndarray with shape (n,), optional + Addition diagonal part, affects the quadratic term. + If None, assumed to be 0. + s0 : None or ndarray with shape (n,), optional + Initial point. If None, assumed to be 0. + + Returns + ------- + a : float + Coefficient for t**2. + b : float + Coefficient for t. + c : float + Free term. Returned only if `s0` is provided. + """ + v = J.dot(s) + a = np.dot(v, v) + if diag is not None: + a += np.dot(s * diag, s) + a *= 0.5 + + b = np.dot(g, s) + + if s0 is not None: + u = J.dot(s0) + b += np.dot(u, v) + c = 0.5 * np.dot(u, u) + np.dot(g, s0) + if diag is not None: + b += np.dot(s0 * diag, s) + c += 0.5 * np.dot(s0 * diag, s0) + return a, b, c + else: + return a, b + + +def minimize_quadratic_1d(a, b, lb, ub, c=0): + """Minimize a 1-D quadratic function subject to bounds. + + The free term `c` is 0 by default. Bounds must be finite. + + Returns + ------- + t : float + Minimum point. + y : float + Minimum value. + """ + t = [lb, ub] + if a != 0: + extremum = -0.5 * b / a + if lb < extremum < ub: + t.append(extremum) + t = np.asarray(t) + y = t * (a * t + b) + c + min_index = np.argmin(y) + return t[min_index], y[min_index] + + +def evaluate_quadratic(J, g, s, diag=None): + """Compute values of a quadratic function arising in least squares. + + The function is 0.5 * s.T * (J.T * J + diag) * s + g.T * s. + + Parameters + ---------- + J : ndarray, sparse matrix or LinearOperator, shape (m, n) + Jacobian matrix, affects the quadratic term. + g : ndarray, shape (n,) + Gradient, defines the linear term. + s : ndarray, shape (k, n) or (n,) + Array containing steps as rows. + diag : ndarray, shape (n,), optional + Addition diagonal part, affects the quadratic term. + If None, assumed to be 0. + + Returns + ------- + values : ndarray with shape (k,) or float + Values of the function. If `s` was 2-D, then ndarray is + returned, otherwise, float is returned. + """ + if s.ndim == 1: + Js = J.dot(s) + q = np.dot(Js, Js) + if diag is not None: + q += np.dot(s * diag, s) + else: + Js = J.dot(s.T) + q = np.sum(Js**2, axis=0) + if diag is not None: + q += np.sum(diag * s**2, axis=1) + + l = np.dot(s, g) + + return 0.5 * q + l + + +# Utility functions to work with bound constraints. + + +def in_bounds(x, lb, ub): + """Check if a point lies within bounds.""" + return np.all((x >= lb) & (x <= ub)) + + +def step_size_to_bound(x, s, lb, ub): + """Compute a min_step size required to reach a bound. + + The function computes a positive scalar t, such that x + s * t is on + the bound. + + Returns + ------- + step : float + Computed step. Non-negative value. + hits : ndarray of int with shape of x + Each element indicates whether a corresponding variable reaches the + bound: + + * 0 - the bound was not hit. + * -1 - the lower bound was hit. + * 1 - the upper bound was hit. + """ + non_zero = np.nonzero(s) + s_non_zero = s[non_zero] + steps = np.empty_like(x) + steps.fill(np.inf) + with np.errstate(over='ignore'): + steps[non_zero] = np.maximum((lb - x)[non_zero] / s_non_zero, + (ub - x)[non_zero] / s_non_zero) + min_step = np.min(steps) + return min_step, np.equal(steps, min_step) * np.sign(s).astype(int) + + +def find_active_constraints(x, lb, ub, rtol=1e-10): + """Determine which constraints are active in a given point. + + The threshold is computed using `rtol` and the absolute value of the + closest bound. + + Returns + ------- + active : ndarray of int with shape of x + Each component shows whether the corresponding constraint is active: + + * 0 - a constraint is not active. + * -1 - a lower bound is active. + * 1 - a upper bound is active. + """ + active = np.zeros_like(x, dtype=int) + + if rtol == 0: + active[x <= lb] = -1 + active[x >= ub] = 1 + return active + + lower_dist = x - lb + upper_dist = ub - x + + lower_threshold = rtol * np.maximum(1, np.abs(lb)) + upper_threshold = rtol * np.maximum(1, np.abs(ub)) + + lower_active = (np.isfinite(lb) & + (lower_dist <= np.minimum(upper_dist, lower_threshold))) + active[lower_active] = -1 + + upper_active = (np.isfinite(ub) & + (upper_dist <= np.minimum(lower_dist, upper_threshold))) + active[upper_active] = 1 + + return active + + +def make_strictly_feasible(x, lb, ub, rstep=1e-10): + """Shift a point to the interior of a feasible region. + + Each element of the returned vector is at least at a relative distance + `rstep` from the closest bound. If ``rstep=0`` then `np.nextafter` is used. + """ + x_new = x.copy() + + active = find_active_constraints(x, lb, ub, rstep) + lower_mask = np.equal(active, -1) + upper_mask = np.equal(active, 1) + + if rstep == 0: + x_new[lower_mask] = np.nextafter(lb[lower_mask], ub[lower_mask]) + x_new[upper_mask] = np.nextafter(ub[upper_mask], lb[upper_mask]) + else: + x_new[lower_mask] = (lb[lower_mask] + + rstep * np.maximum(1, np.abs(lb[lower_mask]))) + x_new[upper_mask] = (ub[upper_mask] - + rstep * np.maximum(1, np.abs(ub[upper_mask]))) + + tight_bounds = (x_new < lb) | (x_new > ub) + x_new[tight_bounds] = 0.5 * (lb[tight_bounds] + ub[tight_bounds]) + + return x_new + + +def CL_scaling_vector(x, g, lb, ub): + """Compute Coleman-Li scaling vector and its derivatives. + + Components of a vector v are defined as follows:: + + | ub[i] - x[i], if g[i] < 0 and ub[i] < np.inf + v[i] = | x[i] - lb[i], if g[i] > 0 and lb[i] > -np.inf + | 1, otherwise + + According to this definition v[i] >= 0 for all i. It differs from the + definition in paper [1]_ (eq. (2.2)), where the absolute value of v is + used. Both definitions are equivalent down the line. + Derivatives of v with respect to x take value 1, -1 or 0 depending on a + case. + + Returns + ------- + v : ndarray with shape of x + Scaling vector. + dv : ndarray with shape of x + Derivatives of v[i] with respect to x[i], diagonal elements of v's + Jacobian. + + References + ---------- + .. [1] M.A. Branch, T.F. Coleman, and Y. Li, "A Subspace, Interior, + and Conjugate Gradient Method for Large-Scale Bound-Constrained + Minimization Problems," SIAM Journal on Scientific Computing, + Vol. 21, Number 1, pp 1-23, 1999. + """ + v = np.ones_like(x) + dv = np.zeros_like(x) + + mask = (g < 0) & np.isfinite(ub) + v[mask] = ub[mask] - x[mask] + dv[mask] = -1 + + mask = (g > 0) & np.isfinite(lb) + v[mask] = x[mask] - lb[mask] + dv[mask] = 1 + + return v, dv + + +def reflective_transformation(y, lb, ub): + """Compute reflective transformation and its gradient.""" + if in_bounds(y, lb, ub): + return y, np.ones_like(y) + + lb_finite = np.isfinite(lb) + ub_finite = np.isfinite(ub) + + x = y.copy() + g_negative = np.zeros_like(y, dtype=bool) + + mask = lb_finite & ~ub_finite + x[mask] = np.maximum(y[mask], 2 * lb[mask] - y[mask]) + g_negative[mask] = y[mask] < lb[mask] + + mask = ~lb_finite & ub_finite + x[mask] = np.minimum(y[mask], 2 * ub[mask] - y[mask]) + g_negative[mask] = y[mask] > ub[mask] + + mask = lb_finite & ub_finite + d = ub - lb + t = np.remainder(y[mask] - lb[mask], 2 * d[mask]) + x[mask] = lb[mask] + np.minimum(t, 2 * d[mask] - t) + g_negative[mask] = t > d[mask] + + g = np.ones_like(y) + g[g_negative] = -1 + + return x, g + + +# Functions to display algorithm's progress. + + +def print_header_nonlinear(): + print("{:^15}{:^15}{:^15}{:^15}{:^15}{:^15}" + .format("Iteration", "Total nfev", "Cost", "Cost reduction", + "Step norm", "Optimality")) + + +def print_iteration_nonlinear(iteration, nfev, cost, cost_reduction, + step_norm, optimality): + if cost_reduction is None: + cost_reduction = " " * 15 + else: + cost_reduction = f"{cost_reduction:^15.2e}" + + if step_norm is None: + step_norm = " " * 15 + else: + step_norm = f"{step_norm:^15.2e}" + + print("{:^15}{:^15}{:^15.4e}{}{}{:^15.2e}" + .format(iteration, nfev, cost, cost_reduction, + step_norm, optimality)) + + +def print_header_linear(): + print("{:^15}{:^15}{:^15}{:^15}{:^15}" + .format("Iteration", "Cost", "Cost reduction", "Step norm", + "Optimality")) + + +def print_iteration_linear(iteration, cost, cost_reduction, step_norm, + optimality): + if cost_reduction is None: + cost_reduction = " " * 15 + else: + cost_reduction = f"{cost_reduction:^15.2e}" + + if step_norm is None: + step_norm = " " * 15 + else: + step_norm = f"{step_norm:^15.2e}" + + print(f"{iteration:^15}{cost:^15.4e}{cost_reduction}{step_norm}{optimality:^15.2e}") + + +# Simple helper functions. + + +def compute_grad(J, f): + """Compute gradient of the least-squares cost function.""" + if isinstance(J, LinearOperator): + return J.rmatvec(f) + else: + return J.T.dot(f) + + +def compute_jac_scale(J, scale_inv_old=None): + """Compute variables scale based on the Jacobian matrix.""" + if issparse(J): + scale_inv = np.asarray(J.power(2).sum(axis=0)).ravel()**0.5 + else: + scale_inv = np.sum(J**2, axis=0)**0.5 + + if scale_inv_old is None: + scale_inv[scale_inv == 0] = 1 + else: + scale_inv = np.maximum(scale_inv, scale_inv_old) + + return 1 / scale_inv, scale_inv + + +def left_multiplied_operator(J, d): + """Return diag(d) J as LinearOperator.""" + J = aslinearoperator(J) + + def matvec(x): + return d * J.matvec(x) + + def matmat(X): + return d[:, np.newaxis] * J.matmat(X) + + def rmatvec(x): + return J.rmatvec(x.ravel() * d) + + return LinearOperator(J.shape, matvec=matvec, matmat=matmat, + rmatvec=rmatvec) + + +def right_multiplied_operator(J, d): + """Return J diag(d) as LinearOperator.""" + J = aslinearoperator(J) + + def matvec(x): + return J.matvec(np.ravel(x) * d) + + def matmat(X): + return J.matmat(X * d[:, np.newaxis]) + + def rmatvec(x): + return d * J.rmatvec(x) + + return LinearOperator(J.shape, matvec=matvec, matmat=matmat, + rmatvec=rmatvec) + + +def regularized_lsq_operator(J, diag): + """Return a matrix arising in regularized least squares as LinearOperator. + + The matrix is + [ J ] + [ D ] + where D is diagonal matrix with elements from `diag`. + """ + J = aslinearoperator(J) + m, n = J.shape + + def matvec(x): + return np.hstack((J.matvec(x), diag * x)) + + def rmatvec(x): + x1 = x[:m] + x2 = x[m:] + return J.rmatvec(x1) + diag * x2 + + return LinearOperator((m + n, n), matvec=matvec, rmatvec=rmatvec) + + +def right_multiply(J, d, copy=True): + """Compute J diag(d). + + If `copy` is False, `J` is modified in place (unless being LinearOperator). + """ + if copy and not isinstance(J, LinearOperator): + J = J.copy() + + if issparse(J): + J.data *= d.take(J.indices, mode='clip') # scikit-learn recipe. + elif isinstance(J, LinearOperator): + J = right_multiplied_operator(J, d) + else: + J *= d + + return J + + +def left_multiply(J, d, copy=True): + """Compute diag(d) J. + + If `copy` is False, `J` is modified in place (unless being LinearOperator). + """ + if copy and not isinstance(J, LinearOperator): + J = J.copy() + + if issparse(J): + J.data *= np.repeat(d, np.diff(J.indptr)) # scikit-learn recipe. + elif isinstance(J, LinearOperator): + J = left_multiplied_operator(J, d) + else: + J *= d[:, np.newaxis] + + return J + + +def check_termination(dF, F, dx_norm, x_norm, ratio, ftol, xtol): + """Check termination condition for nonlinear least squares.""" + ftol_satisfied = dF < ftol * F and ratio > 0.25 + xtol_satisfied = dx_norm < xtol * (xtol + x_norm) + + if ftol_satisfied and xtol_satisfied: + return 4 + elif ftol_satisfied: + return 2 + elif xtol_satisfied: + return 3 + else: + return None + + +def scale_for_robust_loss_function(J, f, rho): + """Scale Jacobian and residuals for a robust loss function. + + Arrays are modified in place. + """ + J_scale = rho[1] + 2 * rho[2] * f**2 + J_scale[J_scale < EPS] = EPS + J_scale **= 0.5 + + f *= rho[1] / J_scale + + return left_multiply(J, J_scale, copy=False), f diff --git a/parrot/lib/python3.10/site-packages/scipy/optimize/_lsq/dogbox.py b/parrot/lib/python3.10/site-packages/scipy/optimize/_lsq/dogbox.py new file mode 100644 index 0000000000000000000000000000000000000000..6bb5abbe79028afed7b110603a0d5dfd6affae7f --- /dev/null +++ b/parrot/lib/python3.10/site-packages/scipy/optimize/_lsq/dogbox.py @@ -0,0 +1,331 @@ +""" +Dogleg algorithm with rectangular trust regions for least-squares minimization. + +The description of the algorithm can be found in [Voglis]_. The algorithm does +trust-region iterations, but the shape of trust regions is rectangular as +opposed to conventional elliptical. The intersection of a trust region and +an initial feasible region is again some rectangle. Thus, on each iteration a +bound-constrained quadratic optimization problem is solved. + +A quadratic problem is solved by well-known dogleg approach, where the +function is minimized along piecewise-linear "dogleg" path [NumOpt]_, +Chapter 4. If Jacobian is not rank-deficient then the function is decreasing +along this path, and optimization amounts to simply following along this +path as long as a point stays within the bounds. A constrained Cauchy step +(along the anti-gradient) is considered for safety in rank deficient cases, +in this situations the convergence might be slow. + +If during iterations some variable hit the initial bound and the component +of anti-gradient points outside the feasible region, then a next dogleg step +won't make any progress. At this state such variables satisfy first-order +optimality conditions and they are excluded before computing a next dogleg +step. + +Gauss-Newton step can be computed exactly by `numpy.linalg.lstsq` (for dense +Jacobian matrices) or by iterative procedure `scipy.sparse.linalg.lsmr` (for +dense and sparse matrices, or Jacobian being LinearOperator). The second +option allows to solve very large problems (up to couple of millions of +residuals on a regular PC), provided the Jacobian matrix is sufficiently +sparse. But note that dogbox is not very good for solving problems with +large number of constraints, because of variables exclusion-inclusion on each +iteration (a required number of function evaluations might be high or accuracy +of a solution will be poor), thus its large-scale usage is probably limited +to unconstrained problems. + +References +---------- +.. [Voglis] C. Voglis and I. E. Lagaris, "A Rectangular Trust Region Dogleg + Approach for Unconstrained and Bound Constrained Nonlinear + Optimization", WSEAS International Conference on Applied + Mathematics, Corfu, Greece, 2004. +.. [NumOpt] J. Nocedal and S. J. Wright, "Numerical optimization, 2nd edition". +""" +import numpy as np +from numpy.linalg import lstsq, norm + +from scipy.sparse.linalg import LinearOperator, aslinearoperator, lsmr +from scipy.optimize import OptimizeResult + +from .common import ( + step_size_to_bound, in_bounds, update_tr_radius, evaluate_quadratic, + build_quadratic_1d, minimize_quadratic_1d, compute_grad, + compute_jac_scale, check_termination, scale_for_robust_loss_function, + print_header_nonlinear, print_iteration_nonlinear) + + +def lsmr_operator(Jop, d, active_set): + """Compute LinearOperator to use in LSMR by dogbox algorithm. + + `active_set` mask is used to excluded active variables from computations + of matrix-vector products. + """ + m, n = Jop.shape + + def matvec(x): + x_free = x.ravel().copy() + x_free[active_set] = 0 + return Jop.matvec(x * d) + + def rmatvec(x): + r = d * Jop.rmatvec(x) + r[active_set] = 0 + return r + + return LinearOperator((m, n), matvec=matvec, rmatvec=rmatvec, dtype=float) + + +def find_intersection(x, tr_bounds, lb, ub): + """Find intersection of trust-region bounds and initial bounds. + + Returns + ------- + lb_total, ub_total : ndarray with shape of x + Lower and upper bounds of the intersection region. + orig_l, orig_u : ndarray of bool with shape of x + True means that an original bound is taken as a corresponding bound + in the intersection region. + tr_l, tr_u : ndarray of bool with shape of x + True means that a trust-region bound is taken as a corresponding bound + in the intersection region. + """ + lb_centered = lb - x + ub_centered = ub - x + + lb_total = np.maximum(lb_centered, -tr_bounds) + ub_total = np.minimum(ub_centered, tr_bounds) + + orig_l = np.equal(lb_total, lb_centered) + orig_u = np.equal(ub_total, ub_centered) + + tr_l = np.equal(lb_total, -tr_bounds) + tr_u = np.equal(ub_total, tr_bounds) + + return lb_total, ub_total, orig_l, orig_u, tr_l, tr_u + + +def dogleg_step(x, newton_step, g, a, b, tr_bounds, lb, ub): + """Find dogleg step in a rectangular region. + + Returns + ------- + step : ndarray, shape (n,) + Computed dogleg step. + bound_hits : ndarray of int, shape (n,) + Each component shows whether a corresponding variable hits the + initial bound after the step is taken: + * 0 - a variable doesn't hit the bound. + * -1 - lower bound is hit. + * 1 - upper bound is hit. + tr_hit : bool + Whether the step hit the boundary of the trust-region. + """ + lb_total, ub_total, orig_l, orig_u, tr_l, tr_u = find_intersection( + x, tr_bounds, lb, ub + ) + bound_hits = np.zeros_like(x, dtype=int) + + if in_bounds(newton_step, lb_total, ub_total): + return newton_step, bound_hits, False + + to_bounds, _ = step_size_to_bound(np.zeros_like(x), -g, lb_total, ub_total) + + # The classical dogleg algorithm would check if Cauchy step fits into + # the bounds, and just return it constrained version if not. But in a + # rectangular trust region it makes sense to try to improve constrained + # Cauchy step too. Thus, we don't distinguish these two cases. + + cauchy_step = -minimize_quadratic_1d(a, b, 0, to_bounds)[0] * g + + step_diff = newton_step - cauchy_step + step_size, hits = step_size_to_bound(cauchy_step, step_diff, + lb_total, ub_total) + bound_hits[(hits < 0) & orig_l] = -1 + bound_hits[(hits > 0) & orig_u] = 1 + tr_hit = np.any((hits < 0) & tr_l | (hits > 0) & tr_u) + + return cauchy_step + step_size * step_diff, bound_hits, tr_hit + + +def dogbox(fun, jac, x0, f0, J0, lb, ub, ftol, xtol, gtol, max_nfev, x_scale, + loss_function, tr_solver, tr_options, verbose): + f = f0 + f_true = f.copy() + nfev = 1 + + J = J0 + njev = 1 + + if loss_function is not None: + rho = loss_function(f) + cost = 0.5 * np.sum(rho[0]) + J, f = scale_for_robust_loss_function(J, f, rho) + else: + cost = 0.5 * np.dot(f, f) + + g = compute_grad(J, f) + + jac_scale = isinstance(x_scale, str) and x_scale == 'jac' + if jac_scale: + scale, scale_inv = compute_jac_scale(J) + else: + scale, scale_inv = x_scale, 1 / x_scale + + Delta = norm(x0 * scale_inv, ord=np.inf) + if Delta == 0: + Delta = 1.0 + + on_bound = np.zeros_like(x0, dtype=int) + on_bound[np.equal(x0, lb)] = -1 + on_bound[np.equal(x0, ub)] = 1 + + x = x0 + step = np.empty_like(x0) + + if max_nfev is None: + max_nfev = x0.size * 100 + + termination_status = None + iteration = 0 + step_norm = None + actual_reduction = None + + if verbose == 2: + print_header_nonlinear() + + while True: + active_set = on_bound * g < 0 + free_set = ~active_set + + g_free = g[free_set] + g_full = g.copy() + g[active_set] = 0 + + g_norm = norm(g, ord=np.inf) + if g_norm < gtol: + termination_status = 1 + + if verbose == 2: + print_iteration_nonlinear(iteration, nfev, cost, actual_reduction, + step_norm, g_norm) + + if termination_status is not None or nfev == max_nfev: + break + + x_free = x[free_set] + lb_free = lb[free_set] + ub_free = ub[free_set] + scale_free = scale[free_set] + + # Compute (Gauss-)Newton and build quadratic model for Cauchy step. + if tr_solver == 'exact': + J_free = J[:, free_set] + newton_step = lstsq(J_free, -f, rcond=-1)[0] + + # Coefficients for the quadratic model along the anti-gradient. + a, b = build_quadratic_1d(J_free, g_free, -g_free) + elif tr_solver == 'lsmr': + Jop = aslinearoperator(J) + + # We compute lsmr step in scaled variables and then + # transform back to normal variables, if lsmr would give exact lsq + # solution, this would be equivalent to not doing any + # transformations, but from experience it's better this way. + + # We pass active_set to make computations as if we selected + # the free subset of J columns, but without actually doing any + # slicing, which is expensive for sparse matrices and impossible + # for LinearOperator. + + lsmr_op = lsmr_operator(Jop, scale, active_set) + newton_step = -lsmr(lsmr_op, f, **tr_options)[0][free_set] + newton_step *= scale_free + + # Components of g for active variables were zeroed, so this call + # is correct and equivalent to using J_free and g_free. + a, b = build_quadratic_1d(Jop, g, -g) + + actual_reduction = -1.0 + while actual_reduction <= 0 and nfev < max_nfev: + tr_bounds = Delta * scale_free + + step_free, on_bound_free, tr_hit = dogleg_step( + x_free, newton_step, g_free, a, b, tr_bounds, lb_free, ub_free) + + step.fill(0.0) + step[free_set] = step_free + + if tr_solver == 'exact': + predicted_reduction = -evaluate_quadratic(J_free, g_free, + step_free) + elif tr_solver == 'lsmr': + predicted_reduction = -evaluate_quadratic(Jop, g, step) + + # gh11403 ensure that solution is fully within bounds. + x_new = np.clip(x + step, lb, ub) + + f_new = fun(x_new) + nfev += 1 + + step_h_norm = norm(step * scale_inv, ord=np.inf) + + if not np.all(np.isfinite(f_new)): + Delta = 0.25 * step_h_norm + continue + + # Usual trust-region step quality estimation. + if loss_function is not None: + cost_new = loss_function(f_new, cost_only=True) + else: + cost_new = 0.5 * np.dot(f_new, f_new) + actual_reduction = cost - cost_new + + Delta, ratio = update_tr_radius( + Delta, actual_reduction, predicted_reduction, + step_h_norm, tr_hit + ) + + step_norm = norm(step) + termination_status = check_termination( + actual_reduction, cost, step_norm, norm(x), ratio, ftol, xtol) + + if termination_status is not None: + break + + if actual_reduction > 0: + on_bound[free_set] = on_bound_free + + x = x_new + # Set variables exactly at the boundary. + mask = on_bound == -1 + x[mask] = lb[mask] + mask = on_bound == 1 + x[mask] = ub[mask] + + f = f_new + f_true = f.copy() + + cost = cost_new + + J = jac(x, f) + njev += 1 + + if loss_function is not None: + rho = loss_function(f) + J, f = scale_for_robust_loss_function(J, f, rho) + + g = compute_grad(J, f) + + if jac_scale: + scale, scale_inv = compute_jac_scale(J, scale_inv) + else: + step_norm = 0 + actual_reduction = 0 + + iteration += 1 + + if termination_status is None: + termination_status = 0 + + return OptimizeResult( + x=x, cost=cost, fun=f_true, jac=J, grad=g_full, optimality=g_norm, + active_mask=on_bound, nfev=nfev, njev=njev, status=termination_status) diff --git a/parrot/lib/python3.10/site-packages/scipy/optimize/_lsq/least_squares.py b/parrot/lib/python3.10/site-packages/scipy/optimize/_lsq/least_squares.py new file mode 100644 index 0000000000000000000000000000000000000000..db8bb31c7b1530fd48ac7ae58cf501e2b0081a91 --- /dev/null +++ b/parrot/lib/python3.10/site-packages/scipy/optimize/_lsq/least_squares.py @@ -0,0 +1,967 @@ +"""Generic interface for least-squares minimization.""" +from warnings import warn + +import numpy as np +from numpy.linalg import norm + +from scipy.sparse import issparse +from scipy.sparse.linalg import LinearOperator +from scipy.optimize import _minpack, OptimizeResult +from scipy.optimize._numdiff import approx_derivative, group_columns +from scipy.optimize._minimize import Bounds + +from .trf import trf +from .dogbox import dogbox +from .common import EPS, in_bounds, make_strictly_feasible + + +TERMINATION_MESSAGES = { + -1: "Improper input parameters status returned from `leastsq`", + 0: "The maximum number of function evaluations is exceeded.", + 1: "`gtol` termination condition is satisfied.", + 2: "`ftol` termination condition is satisfied.", + 3: "`xtol` termination condition is satisfied.", + 4: "Both `ftol` and `xtol` termination conditions are satisfied." +} + + +FROM_MINPACK_TO_COMMON = { + 0: -1, # Improper input parameters from MINPACK. + 1: 2, + 2: 3, + 3: 4, + 4: 1, + 5: 0 + # There are 6, 7, 8 for too small tolerance parameters, + # but we guard against it by checking ftol, xtol, gtol beforehand. +} + + +def call_minpack(fun, x0, jac, ftol, xtol, gtol, max_nfev, x_scale, diff_step): + n = x0.size + + if diff_step is None: + epsfcn = EPS + else: + epsfcn = diff_step**2 + + # Compute MINPACK's `diag`, which is inverse of our `x_scale` and + # ``x_scale='jac'`` corresponds to ``diag=None``. + if isinstance(x_scale, str) and x_scale == 'jac': + diag = None + else: + diag = 1 / x_scale + + full_output = True + col_deriv = False + factor = 100.0 + + if jac is None: + if max_nfev is None: + # n squared to account for Jacobian evaluations. + max_nfev = 100 * n * (n + 1) + x, info, status = _minpack._lmdif( + fun, x0, (), full_output, ftol, xtol, gtol, + max_nfev, epsfcn, factor, diag) + else: + if max_nfev is None: + max_nfev = 100 * n + x, info, status = _minpack._lmder( + fun, jac, x0, (), full_output, col_deriv, + ftol, xtol, gtol, max_nfev, factor, diag) + + f = info['fvec'] + + if callable(jac): + J = jac(x) + else: + J = np.atleast_2d(approx_derivative(fun, x)) + + cost = 0.5 * np.dot(f, f) + g = J.T.dot(f) + g_norm = norm(g, ord=np.inf) + + nfev = info['nfev'] + njev = info.get('njev', None) + + status = FROM_MINPACK_TO_COMMON[status] + active_mask = np.zeros_like(x0, dtype=int) + + return OptimizeResult( + x=x, cost=cost, fun=f, jac=J, grad=g, optimality=g_norm, + active_mask=active_mask, nfev=nfev, njev=njev, status=status) + + +def prepare_bounds(bounds, n): + lb, ub = (np.asarray(b, dtype=float) for b in bounds) + if lb.ndim == 0: + lb = np.resize(lb, n) + + if ub.ndim == 0: + ub = np.resize(ub, n) + + return lb, ub + + +def check_tolerance(ftol, xtol, gtol, method): + def check(tol, name): + if tol is None: + tol = 0 + elif tol < EPS: + warn(f"Setting `{name}` below the machine epsilon ({EPS:.2e}) effectively " + f"disables the corresponding termination condition.", + stacklevel=3) + return tol + + ftol = check(ftol, "ftol") + xtol = check(xtol, "xtol") + gtol = check(gtol, "gtol") + + if method == "lm" and (ftol < EPS or xtol < EPS or gtol < EPS): + raise ValueError("All tolerances must be higher than machine epsilon " + f"({EPS:.2e}) for method 'lm'.") + elif ftol < EPS and xtol < EPS and gtol < EPS: + raise ValueError("At least one of the tolerances must be higher than " + f"machine epsilon ({EPS:.2e}).") + + return ftol, xtol, gtol + + +def check_x_scale(x_scale, x0): + if isinstance(x_scale, str) and x_scale == 'jac': + return x_scale + + try: + x_scale = np.asarray(x_scale, dtype=float) + valid = np.all(np.isfinite(x_scale)) and np.all(x_scale > 0) + except (ValueError, TypeError): + valid = False + + if not valid: + raise ValueError("`x_scale` must be 'jac' or array_like with " + "positive numbers.") + + if x_scale.ndim == 0: + x_scale = np.resize(x_scale, x0.shape) + + if x_scale.shape != x0.shape: + raise ValueError("Inconsistent shapes between `x_scale` and `x0`.") + + return x_scale + + +def check_jac_sparsity(jac_sparsity, m, n): + if jac_sparsity is None: + return None + + if not issparse(jac_sparsity): + jac_sparsity = np.atleast_2d(jac_sparsity) + + if jac_sparsity.shape != (m, n): + raise ValueError("`jac_sparsity` has wrong shape.") + + return jac_sparsity, group_columns(jac_sparsity) + + +# Loss functions. + + +def huber(z, rho, cost_only): + mask = z <= 1 + rho[0, mask] = z[mask] + rho[0, ~mask] = 2 * z[~mask]**0.5 - 1 + if cost_only: + return + rho[1, mask] = 1 + rho[1, ~mask] = z[~mask]**-0.5 + rho[2, mask] = 0 + rho[2, ~mask] = -0.5 * z[~mask]**-1.5 + + +def soft_l1(z, rho, cost_only): + t = 1 + z + rho[0] = 2 * (t**0.5 - 1) + if cost_only: + return + rho[1] = t**-0.5 + rho[2] = -0.5 * t**-1.5 + + +def cauchy(z, rho, cost_only): + rho[0] = np.log1p(z) + if cost_only: + return + t = 1 + z + rho[1] = 1 / t + rho[2] = -1 / t**2 + + +def arctan(z, rho, cost_only): + rho[0] = np.arctan(z) + if cost_only: + return + t = 1 + z**2 + rho[1] = 1 / t + rho[2] = -2 * z / t**2 + + +IMPLEMENTED_LOSSES = dict(linear=None, huber=huber, soft_l1=soft_l1, + cauchy=cauchy, arctan=arctan) + + +def construct_loss_function(m, loss, f_scale): + if loss == 'linear': + return None + + if not callable(loss): + loss = IMPLEMENTED_LOSSES[loss] + rho = np.empty((3, m)) + + def loss_function(f, cost_only=False): + z = (f / f_scale) ** 2 + loss(z, rho, cost_only=cost_only) + if cost_only: + return 0.5 * f_scale ** 2 * np.sum(rho[0]) + rho[0] *= f_scale ** 2 + rho[2] /= f_scale ** 2 + return rho + else: + def loss_function(f, cost_only=False): + z = (f / f_scale) ** 2 + rho = loss(z) + if cost_only: + return 0.5 * f_scale ** 2 * np.sum(rho[0]) + rho[0] *= f_scale ** 2 + rho[2] /= f_scale ** 2 + return rho + + return loss_function + + +def least_squares( + fun, x0, jac='2-point', bounds=(-np.inf, np.inf), method='trf', + ftol=1e-8, xtol=1e-8, gtol=1e-8, x_scale=1.0, loss='linear', + f_scale=1.0, diff_step=None, tr_solver=None, tr_options={}, + jac_sparsity=None, max_nfev=None, verbose=0, args=(), kwargs={}): + """Solve a nonlinear least-squares problem with bounds on the variables. + + Given the residuals f(x) (an m-D real function of n real + variables) and the loss function rho(s) (a scalar function), `least_squares` + finds a local minimum of the cost function F(x):: + + minimize F(x) = 0.5 * sum(rho(f_i(x)**2), i = 0, ..., m - 1) + subject to lb <= x <= ub + + The purpose of the loss function rho(s) is to reduce the influence of + outliers on the solution. + + Parameters + ---------- + fun : callable + Function which computes the vector of residuals, with the signature + ``fun(x, *args, **kwargs)``, i.e., the minimization proceeds with + respect to its first argument. The argument ``x`` passed to this + function is an ndarray of shape (n,) (never a scalar, even for n=1). + It must allocate and return a 1-D array_like of shape (m,) or a scalar. + If the argument ``x`` is complex or the function ``fun`` returns + complex residuals, it must be wrapped in a real function of real + arguments, as shown at the end of the Examples section. + x0 : array_like with shape (n,) or float + Initial guess on independent variables. If float, it will be treated + as a 1-D array with one element. When `method` is 'trf', the initial + guess might be slightly adjusted to lie sufficiently within the given + `bounds`. + jac : {'2-point', '3-point', 'cs', callable}, optional + Method of computing the Jacobian matrix (an m-by-n matrix, where + element (i, j) is the partial derivative of f[i] with respect to + x[j]). The keywords select a finite difference scheme for numerical + estimation. The scheme '3-point' is more accurate, but requires + twice as many operations as '2-point' (default). The scheme 'cs' + uses complex steps, and while potentially the most accurate, it is + applicable only when `fun` correctly handles complex inputs and + can be analytically continued to the complex plane. Method 'lm' + always uses the '2-point' scheme. If callable, it is used as + ``jac(x, *args, **kwargs)`` and should return a good approximation + (or the exact value) for the Jacobian as an array_like (np.atleast_2d + is applied), a sparse matrix (csr_matrix preferred for performance) or + a `scipy.sparse.linalg.LinearOperator`. + bounds : 2-tuple of array_like or `Bounds`, optional + There are two ways to specify bounds: + + 1. Instance of `Bounds` class + 2. Lower and upper bounds on independent variables. Defaults to no + bounds. Each array must match the size of `x0` or be a scalar, + in the latter case a bound will be the same for all variables. + Use ``np.inf`` with an appropriate sign to disable bounds on all + or some variables. + method : {'trf', 'dogbox', 'lm'}, optional + Algorithm to perform minimization. + + * 'trf' : Trust Region Reflective algorithm, particularly suitable + for large sparse problems with bounds. Generally robust method. + * 'dogbox' : dogleg algorithm with rectangular trust regions, + typical use case is small problems with bounds. Not recommended + for problems with rank-deficient Jacobian. + * 'lm' : Levenberg-Marquardt algorithm as implemented in MINPACK. + Doesn't handle bounds and sparse Jacobians. Usually the most + efficient method for small unconstrained problems. + + Default is 'trf'. See Notes for more information. + ftol : float or None, optional + Tolerance for termination by the change of the cost function. Default + is 1e-8. The optimization process is stopped when ``dF < ftol * F``, + and there was an adequate agreement between a local quadratic model and + the true model in the last step. + + If None and 'method' is not 'lm', the termination by this condition is + disabled. If 'method' is 'lm', this tolerance must be higher than + machine epsilon. + xtol : float or None, optional + Tolerance for termination by the change of the independent variables. + Default is 1e-8. The exact condition depends on the `method` used: + + * For 'trf' and 'dogbox' : ``norm(dx) < xtol * (xtol + norm(x))``. + * For 'lm' : ``Delta < xtol * norm(xs)``, where ``Delta`` is + a trust-region radius and ``xs`` is the value of ``x`` + scaled according to `x_scale` parameter (see below). + + If None and 'method' is not 'lm', the termination by this condition is + disabled. If 'method' is 'lm', this tolerance must be higher than + machine epsilon. + gtol : float or None, optional + Tolerance for termination by the norm of the gradient. Default is 1e-8. + The exact condition depends on a `method` used: + + * For 'trf' : ``norm(g_scaled, ord=np.inf) < gtol``, where + ``g_scaled`` is the value of the gradient scaled to account for + the presence of the bounds [STIR]_. + * For 'dogbox' : ``norm(g_free, ord=np.inf) < gtol``, where + ``g_free`` is the gradient with respect to the variables which + are not in the optimal state on the boundary. + * For 'lm' : the maximum absolute value of the cosine of angles + between columns of the Jacobian and the residual vector is less + than `gtol`, or the residual vector is zero. + + If None and 'method' is not 'lm', the termination by this condition is + disabled. If 'method' is 'lm', this tolerance must be higher than + machine epsilon. + x_scale : array_like or 'jac', optional + Characteristic scale of each variable. Setting `x_scale` is equivalent + to reformulating the problem in scaled variables ``xs = x / x_scale``. + An alternative view is that the size of a trust region along jth + dimension is proportional to ``x_scale[j]``. Improved convergence may + be achieved by setting `x_scale` such that a step of a given size + along any of the scaled variables has a similar effect on the cost + function. If set to 'jac', the scale is iteratively updated using the + inverse norms of the columns of the Jacobian matrix (as described in + [JJMore]_). + loss : str or callable, optional + Determines the loss function. The following keyword values are allowed: + + * 'linear' (default) : ``rho(z) = z``. Gives a standard + least-squares problem. + * 'soft_l1' : ``rho(z) = 2 * ((1 + z)**0.5 - 1)``. The smooth + approximation of l1 (absolute value) loss. Usually a good + choice for robust least squares. + * 'huber' : ``rho(z) = z if z <= 1 else 2*z**0.5 - 1``. Works + similarly to 'soft_l1'. + * 'cauchy' : ``rho(z) = ln(1 + z)``. Severely weakens outliers + influence, but may cause difficulties in optimization process. + * 'arctan' : ``rho(z) = arctan(z)``. Limits a maximum loss on + a single residual, has properties similar to 'cauchy'. + + If callable, it must take a 1-D ndarray ``z=f**2`` and return an + array_like with shape (3, m) where row 0 contains function values, + row 1 contains first derivatives and row 2 contains second + derivatives. Method 'lm' supports only 'linear' loss. + f_scale : float, optional + Value of soft margin between inlier and outlier residuals, default + is 1.0. The loss function is evaluated as follows + ``rho_(f**2) = C**2 * rho(f**2 / C**2)``, where ``C`` is `f_scale`, + and ``rho`` is determined by `loss` parameter. This parameter has + no effect with ``loss='linear'``, but for other `loss` values it is + of crucial importance. + max_nfev : None or int, optional + Maximum number of function evaluations before the termination. + If None (default), the value is chosen automatically: + + * For 'trf' and 'dogbox' : 100 * n. + * For 'lm' : 100 * n if `jac` is callable and 100 * n * (n + 1) + otherwise (because 'lm' counts function calls in Jacobian + estimation). + + diff_step : None or array_like, optional + Determines the relative step size for the finite difference + approximation of the Jacobian. The actual step is computed as + ``x * diff_step``. If None (default), then `diff_step` is taken to be + a conventional "optimal" power of machine epsilon for the finite + difference scheme used [NR]_. + tr_solver : {None, 'exact', 'lsmr'}, optional + Method for solving trust-region subproblems, relevant only for 'trf' + and 'dogbox' methods. + + * 'exact' is suitable for not very large problems with dense + Jacobian matrices. The computational complexity per iteration is + comparable to a singular value decomposition of the Jacobian + matrix. + * 'lsmr' is suitable for problems with sparse and large Jacobian + matrices. It uses the iterative procedure + `scipy.sparse.linalg.lsmr` for finding a solution of a linear + least-squares problem and only requires matrix-vector product + evaluations. + + If None (default), the solver is chosen based on the type of Jacobian + returned on the first iteration. + tr_options : dict, optional + Keyword options passed to trust-region solver. + + * ``tr_solver='exact'``: `tr_options` are ignored. + * ``tr_solver='lsmr'``: options for `scipy.sparse.linalg.lsmr`. + Additionally, ``method='trf'`` supports 'regularize' option + (bool, default is True), which adds a regularization term to the + normal equation, which improves convergence if the Jacobian is + rank-deficient [Byrd]_ (eq. 3.4). + + jac_sparsity : {None, array_like, sparse matrix}, optional + Defines the sparsity structure of the Jacobian matrix for finite + difference estimation, its shape must be (m, n). If the Jacobian has + only few non-zero elements in *each* row, providing the sparsity + structure will greatly speed up the computations [Curtis]_. A zero + entry means that a corresponding element in the Jacobian is identically + zero. If provided, forces the use of 'lsmr' trust-region solver. + If None (default), then dense differencing will be used. Has no effect + for 'lm' method. + verbose : {0, 1, 2}, optional + Level of algorithm's verbosity: + + * 0 (default) : work silently. + * 1 : display a termination report. + * 2 : display progress during iterations (not supported by 'lm' + method). + + args, kwargs : tuple and dict, optional + Additional arguments passed to `fun` and `jac`. Both empty by default. + The calling signature is ``fun(x, *args, **kwargs)`` and the same for + `jac`. + + Returns + ------- + result : OptimizeResult + `OptimizeResult` with the following fields defined: + + x : ndarray, shape (n,) + Solution found. + cost : float + Value of the cost function at the solution. + fun : ndarray, shape (m,) + Vector of residuals at the solution. + jac : ndarray, sparse matrix or LinearOperator, shape (m, n) + Modified Jacobian matrix at the solution, in the sense that J^T J + is a Gauss-Newton approximation of the Hessian of the cost function. + The type is the same as the one used by the algorithm. + grad : ndarray, shape (m,) + Gradient of the cost function at the solution. + optimality : float + First-order optimality measure. In unconstrained problems, it is + always the uniform norm of the gradient. In constrained problems, + it is the quantity which was compared with `gtol` during iterations. + active_mask : ndarray of int, shape (n,) + Each component shows whether a corresponding constraint is active + (that is, whether a variable is at the bound): + + * 0 : a constraint is not active. + * -1 : a lower bound is active. + * 1 : an upper bound is active. + + Might be somewhat arbitrary for 'trf' method as it generates a + sequence of strictly feasible iterates and `active_mask` is + determined within a tolerance threshold. + nfev : int + Number of function evaluations done. Methods 'trf' and 'dogbox' do + not count function calls for numerical Jacobian approximation, as + opposed to 'lm' method. + njev : int or None + Number of Jacobian evaluations done. If numerical Jacobian + approximation is used in 'lm' method, it is set to None. + status : int + The reason for algorithm termination: + + * -1 : improper input parameters status returned from MINPACK. + * 0 : the maximum number of function evaluations is exceeded. + * 1 : `gtol` termination condition is satisfied. + * 2 : `ftol` termination condition is satisfied. + * 3 : `xtol` termination condition is satisfied. + * 4 : Both `ftol` and `xtol` termination conditions are satisfied. + + message : str + Verbal description of the termination reason. + success : bool + True if one of the convergence criteria is satisfied (`status` > 0). + + See Also + -------- + leastsq : A legacy wrapper for the MINPACK implementation of the + Levenberg-Marquadt algorithm. + curve_fit : Least-squares minimization applied to a curve-fitting problem. + + Notes + ----- + Method 'lm' (Levenberg-Marquardt) calls a wrapper over least-squares + algorithms implemented in MINPACK (lmder, lmdif). It runs the + Levenberg-Marquardt algorithm formulated as a trust-region type algorithm. + The implementation is based on paper [JJMore]_, it is very robust and + efficient with a lot of smart tricks. It should be your first choice + for unconstrained problems. Note that it doesn't support bounds. Also, + it doesn't work when m < n. + + Method 'trf' (Trust Region Reflective) is motivated by the process of + solving a system of equations, which constitute the first-order optimality + condition for a bound-constrained minimization problem as formulated in + [STIR]_. The algorithm iteratively solves trust-region subproblems + augmented by a special diagonal quadratic term and with trust-region shape + determined by the distance from the bounds and the direction of the + gradient. This enhancements help to avoid making steps directly into bounds + and efficiently explore the whole space of variables. To further improve + convergence, the algorithm considers search directions reflected from the + bounds. To obey theoretical requirements, the algorithm keeps iterates + strictly feasible. With dense Jacobians trust-region subproblems are + solved by an exact method very similar to the one described in [JJMore]_ + (and implemented in MINPACK). The difference from the MINPACK + implementation is that a singular value decomposition of a Jacobian + matrix is done once per iteration, instead of a QR decomposition and series + of Givens rotation eliminations. For large sparse Jacobians a 2-D subspace + approach of solving trust-region subproblems is used [STIR]_, [Byrd]_. + The subspace is spanned by a scaled gradient and an approximate + Gauss-Newton solution delivered by `scipy.sparse.linalg.lsmr`. When no + constraints are imposed the algorithm is very similar to MINPACK and has + generally comparable performance. The algorithm works quite robust in + unbounded and bounded problems, thus it is chosen as a default algorithm. + + Method 'dogbox' operates in a trust-region framework, but considers + rectangular trust regions as opposed to conventional ellipsoids [Voglis]_. + The intersection of a current trust region and initial bounds is again + rectangular, so on each iteration a quadratic minimization problem subject + to bound constraints is solved approximately by Powell's dogleg method + [NumOpt]_. The required Gauss-Newton step can be computed exactly for + dense Jacobians or approximately by `scipy.sparse.linalg.lsmr` for large + sparse Jacobians. The algorithm is likely to exhibit slow convergence when + the rank of Jacobian is less than the number of variables. The algorithm + often outperforms 'trf' in bounded problems with a small number of + variables. + + Robust loss functions are implemented as described in [BA]_. The idea + is to modify a residual vector and a Jacobian matrix on each iteration + such that computed gradient and Gauss-Newton Hessian approximation match + the true gradient and Hessian approximation of the cost function. Then + the algorithm proceeds in a normal way, i.e., robust loss functions are + implemented as a simple wrapper over standard least-squares algorithms. + + .. versionadded:: 0.17.0 + + References + ---------- + .. [STIR] M. A. Branch, T. F. Coleman, and Y. Li, "A Subspace, Interior, + and Conjugate Gradient Method for Large-Scale Bound-Constrained + Minimization Problems," SIAM Journal on Scientific Computing, + Vol. 21, Number 1, pp 1-23, 1999. + .. [NR] William H. Press et. al., "Numerical Recipes. The Art of Scientific + Computing. 3rd edition", Sec. 5.7. + .. [Byrd] R. H. Byrd, R. B. Schnabel and G. A. Shultz, "Approximate + solution of the trust region problem by minimization over + two-dimensional subspaces", Math. Programming, 40, pp. 247-263, + 1988. + .. [Curtis] A. Curtis, M. J. D. Powell, and J. Reid, "On the estimation of + sparse Jacobian matrices", Journal of the Institute of + Mathematics and its Applications, 13, pp. 117-120, 1974. + .. [JJMore] J. J. More, "The Levenberg-Marquardt Algorithm: Implementation + and Theory," Numerical Analysis, ed. G. A. Watson, Lecture + Notes in Mathematics 630, Springer Verlag, pp. 105-116, 1977. + .. [Voglis] C. Voglis and I. E. Lagaris, "A Rectangular Trust Region + Dogleg Approach for Unconstrained and Bound Constrained + Nonlinear Optimization", WSEAS International Conference on + Applied Mathematics, Corfu, Greece, 2004. + .. [NumOpt] J. Nocedal and S. J. Wright, "Numerical optimization, + 2nd edition", Chapter 4. + .. [BA] B. Triggs et. al., "Bundle Adjustment - A Modern Synthesis", + Proceedings of the International Workshop on Vision Algorithms: + Theory and Practice, pp. 298-372, 1999. + + Examples + -------- + In this example we find a minimum of the Rosenbrock function without bounds + on independent variables. + + >>> import numpy as np + >>> def fun_rosenbrock(x): + ... return np.array([10 * (x[1] - x[0]**2), (1 - x[0])]) + + Notice that we only provide the vector of the residuals. The algorithm + constructs the cost function as a sum of squares of the residuals, which + gives the Rosenbrock function. The exact minimum is at ``x = [1.0, 1.0]``. + + >>> from scipy.optimize import least_squares + >>> x0_rosenbrock = np.array([2, 2]) + >>> res_1 = least_squares(fun_rosenbrock, x0_rosenbrock) + >>> res_1.x + array([ 1., 1.]) + >>> res_1.cost + 9.8669242910846867e-30 + >>> res_1.optimality + 8.8928864934219529e-14 + + We now constrain the variables, in such a way that the previous solution + becomes infeasible. Specifically, we require that ``x[1] >= 1.5``, and + ``x[0]`` left unconstrained. To this end, we specify the `bounds` parameter + to `least_squares` in the form ``bounds=([-np.inf, 1.5], np.inf)``. + + We also provide the analytic Jacobian: + + >>> def jac_rosenbrock(x): + ... return np.array([ + ... [-20 * x[0], 10], + ... [-1, 0]]) + + Putting this all together, we see that the new solution lies on the bound: + + >>> res_2 = least_squares(fun_rosenbrock, x0_rosenbrock, jac_rosenbrock, + ... bounds=([-np.inf, 1.5], np.inf)) + >>> res_2.x + array([ 1.22437075, 1.5 ]) + >>> res_2.cost + 0.025213093946805685 + >>> res_2.optimality + 1.5885401433157753e-07 + + Now we solve a system of equations (i.e., the cost function should be zero + at a minimum) for a Broyden tridiagonal vector-valued function of 100000 + variables: + + >>> def fun_broyden(x): + ... f = (3 - x) * x + 1 + ... f[1:] -= x[:-1] + ... f[:-1] -= 2 * x[1:] + ... return f + + The corresponding Jacobian matrix is sparse. We tell the algorithm to + estimate it by finite differences and provide the sparsity structure of + Jacobian to significantly speed up this process. + + >>> from scipy.sparse import lil_matrix + >>> def sparsity_broyden(n): + ... sparsity = lil_matrix((n, n), dtype=int) + ... i = np.arange(n) + ... sparsity[i, i] = 1 + ... i = np.arange(1, n) + ... sparsity[i, i - 1] = 1 + ... i = np.arange(n - 1) + ... sparsity[i, i + 1] = 1 + ... return sparsity + ... + >>> n = 100000 + >>> x0_broyden = -np.ones(n) + ... + >>> res_3 = least_squares(fun_broyden, x0_broyden, + ... jac_sparsity=sparsity_broyden(n)) + >>> res_3.cost + 4.5687069299604613e-23 + >>> res_3.optimality + 1.1650454296851518e-11 + + Let's also solve a curve fitting problem using robust loss function to + take care of outliers in the data. Define the model function as + ``y = a + b * exp(c * t)``, where t is a predictor variable, y is an + observation and a, b, c are parameters to estimate. + + First, define the function which generates the data with noise and + outliers, define the model parameters, and generate data: + + >>> from numpy.random import default_rng + >>> rng = default_rng() + >>> def gen_data(t, a, b, c, noise=0., n_outliers=0, seed=None): + ... rng = default_rng(seed) + ... + ... y = a + b * np.exp(t * c) + ... + ... error = noise * rng.standard_normal(t.size) + ... outliers = rng.integers(0, t.size, n_outliers) + ... error[outliers] *= 10 + ... + ... return y + error + ... + >>> a = 0.5 + >>> b = 2.0 + >>> c = -1 + >>> t_min = 0 + >>> t_max = 10 + >>> n_points = 15 + ... + >>> t_train = np.linspace(t_min, t_max, n_points) + >>> y_train = gen_data(t_train, a, b, c, noise=0.1, n_outliers=3) + + Define function for computing residuals and initial estimate of + parameters. + + >>> def fun(x, t, y): + ... return x[0] + x[1] * np.exp(x[2] * t) - y + ... + >>> x0 = np.array([1.0, 1.0, 0.0]) + + Compute a standard least-squares solution: + + >>> res_lsq = least_squares(fun, x0, args=(t_train, y_train)) + + Now compute two solutions with two different robust loss functions. The + parameter `f_scale` is set to 0.1, meaning that inlier residuals should + not significantly exceed 0.1 (the noise level used). + + >>> res_soft_l1 = least_squares(fun, x0, loss='soft_l1', f_scale=0.1, + ... args=(t_train, y_train)) + >>> res_log = least_squares(fun, x0, loss='cauchy', f_scale=0.1, + ... args=(t_train, y_train)) + + And, finally, plot all the curves. We see that by selecting an appropriate + `loss` we can get estimates close to optimal even in the presence of + strong outliers. But keep in mind that generally it is recommended to try + 'soft_l1' or 'huber' losses first (if at all necessary) as the other two + options may cause difficulties in optimization process. + + >>> t_test = np.linspace(t_min, t_max, n_points * 10) + >>> y_true = gen_data(t_test, a, b, c) + >>> y_lsq = gen_data(t_test, *res_lsq.x) + >>> y_soft_l1 = gen_data(t_test, *res_soft_l1.x) + >>> y_log = gen_data(t_test, *res_log.x) + ... + >>> import matplotlib.pyplot as plt + >>> plt.plot(t_train, y_train, 'o') + >>> plt.plot(t_test, y_true, 'k', linewidth=2, label='true') + >>> plt.plot(t_test, y_lsq, label='linear loss') + >>> plt.plot(t_test, y_soft_l1, label='soft_l1 loss') + >>> plt.plot(t_test, y_log, label='cauchy loss') + >>> plt.xlabel("t") + >>> plt.ylabel("y") + >>> plt.legend() + >>> plt.show() + + In the next example, we show how complex-valued residual functions of + complex variables can be optimized with ``least_squares()``. Consider the + following function: + + >>> def f(z): + ... return z - (0.5 + 0.5j) + + We wrap it into a function of real variables that returns real residuals + by simply handling the real and imaginary parts as independent variables: + + >>> def f_wrap(x): + ... fx = f(x[0] + 1j*x[1]) + ... return np.array([fx.real, fx.imag]) + + Thus, instead of the original m-D complex function of n complex + variables we optimize a 2m-D real function of 2n real variables: + + >>> from scipy.optimize import least_squares + >>> res_wrapped = least_squares(f_wrap, (0.1, 0.1), bounds=([0, 0], [1, 1])) + >>> z = res_wrapped.x[0] + res_wrapped.x[1]*1j + >>> z + (0.49999999999925893+0.49999999999925893j) + + """ + if method not in ['trf', 'dogbox', 'lm']: + raise ValueError("`method` must be 'trf', 'dogbox' or 'lm'.") + + if jac not in ['2-point', '3-point', 'cs'] and not callable(jac): + raise ValueError("`jac` must be '2-point', '3-point', 'cs' or " + "callable.") + + if tr_solver not in [None, 'exact', 'lsmr']: + raise ValueError("`tr_solver` must be None, 'exact' or 'lsmr'.") + + if loss not in IMPLEMENTED_LOSSES and not callable(loss): + raise ValueError("`loss` must be one of {} or a callable." + .format(IMPLEMENTED_LOSSES.keys())) + + if method == 'lm' and loss != 'linear': + raise ValueError("method='lm' supports only 'linear' loss function.") + + if verbose not in [0, 1, 2]: + raise ValueError("`verbose` must be in [0, 1, 2].") + + if max_nfev is not None and max_nfev <= 0: + raise ValueError("`max_nfev` must be None or positive integer.") + + if np.iscomplexobj(x0): + raise ValueError("`x0` must be real.") + + x0 = np.atleast_1d(x0).astype(float) + + if x0.ndim > 1: + raise ValueError("`x0` must have at most 1 dimension.") + + if isinstance(bounds, Bounds): + lb, ub = bounds.lb, bounds.ub + bounds = (lb, ub) + else: + if len(bounds) == 2: + lb, ub = prepare_bounds(bounds, x0.shape[0]) + else: + raise ValueError("`bounds` must contain 2 elements.") + + if method == 'lm' and not np.all((lb == -np.inf) & (ub == np.inf)): + raise ValueError("Method 'lm' doesn't support bounds.") + + if lb.shape != x0.shape or ub.shape != x0.shape: + raise ValueError("Inconsistent shapes between bounds and `x0`.") + + if np.any(lb >= ub): + raise ValueError("Each lower bound must be strictly less than each " + "upper bound.") + + if not in_bounds(x0, lb, ub): + raise ValueError("`x0` is infeasible.") + + x_scale = check_x_scale(x_scale, x0) + + ftol, xtol, gtol = check_tolerance(ftol, xtol, gtol, method) + + if method == 'trf': + x0 = make_strictly_feasible(x0, lb, ub) + + def fun_wrapped(x): + return np.atleast_1d(fun(x, *args, **kwargs)) + + f0 = fun_wrapped(x0) + + if f0.ndim != 1: + raise ValueError("`fun` must return at most 1-d array_like. " + f"f0.shape: {f0.shape}") + + if not np.all(np.isfinite(f0)): + raise ValueError("Residuals are not finite in the initial point.") + + n = x0.size + m = f0.size + + if method == 'lm' and m < n: + raise ValueError("Method 'lm' doesn't work when the number of " + "residuals is less than the number of variables.") + + loss_function = construct_loss_function(m, loss, f_scale) + if callable(loss): + rho = loss_function(f0) + if rho.shape != (3, m): + raise ValueError("The return value of `loss` callable has wrong " + "shape.") + initial_cost = 0.5 * np.sum(rho[0]) + elif loss_function is not None: + initial_cost = loss_function(f0, cost_only=True) + else: + initial_cost = 0.5 * np.dot(f0, f0) + + if callable(jac): + J0 = jac(x0, *args, **kwargs) + + if issparse(J0): + J0 = J0.tocsr() + + def jac_wrapped(x, _=None): + return jac(x, *args, **kwargs).tocsr() + + elif isinstance(J0, LinearOperator): + def jac_wrapped(x, _=None): + return jac(x, *args, **kwargs) + + else: + J0 = np.atleast_2d(J0) + + def jac_wrapped(x, _=None): + return np.atleast_2d(jac(x, *args, **kwargs)) + + else: # Estimate Jacobian by finite differences. + if method == 'lm': + if jac_sparsity is not None: + raise ValueError("method='lm' does not support " + "`jac_sparsity`.") + + if jac != '2-point': + warn(f"jac='{jac}' works equivalently to '2-point' for method='lm'.", + stacklevel=2) + + J0 = jac_wrapped = None + else: + if jac_sparsity is not None and tr_solver == 'exact': + raise ValueError("tr_solver='exact' is incompatible " + "with `jac_sparsity`.") + + jac_sparsity = check_jac_sparsity(jac_sparsity, m, n) + + def jac_wrapped(x, f): + J = approx_derivative(fun, x, rel_step=diff_step, method=jac, + f0=f, bounds=bounds, args=args, + kwargs=kwargs, sparsity=jac_sparsity) + if J.ndim != 2: # J is guaranteed not sparse. + J = np.atleast_2d(J) + + return J + + J0 = jac_wrapped(x0, f0) + + if J0 is not None: + if J0.shape != (m, n): + raise ValueError( + f"The return value of `jac` has wrong shape: expected {(m, n)}, " + f"actual {J0.shape}." + ) + + if not isinstance(J0, np.ndarray): + if method == 'lm': + raise ValueError("method='lm' works only with dense " + "Jacobian matrices.") + + if tr_solver == 'exact': + raise ValueError( + "tr_solver='exact' works only with dense " + "Jacobian matrices.") + + jac_scale = isinstance(x_scale, str) and x_scale == 'jac' + if isinstance(J0, LinearOperator) and jac_scale: + raise ValueError("x_scale='jac' can't be used when `jac` " + "returns LinearOperator.") + + if tr_solver is None: + if isinstance(J0, np.ndarray): + tr_solver = 'exact' + else: + tr_solver = 'lsmr' + + if method == 'lm': + result = call_minpack(fun_wrapped, x0, jac_wrapped, ftol, xtol, gtol, + max_nfev, x_scale, diff_step) + + elif method == 'trf': + result = trf(fun_wrapped, jac_wrapped, x0, f0, J0, lb, ub, ftol, xtol, + gtol, max_nfev, x_scale, loss_function, tr_solver, + tr_options.copy(), verbose) + + elif method == 'dogbox': + if tr_solver == 'lsmr' and 'regularize' in tr_options: + warn("The keyword 'regularize' in `tr_options` is not relevant " + "for 'dogbox' method.", + stacklevel=2) + tr_options = tr_options.copy() + del tr_options['regularize'] + + result = dogbox(fun_wrapped, jac_wrapped, x0, f0, J0, lb, ub, ftol, + xtol, gtol, max_nfev, x_scale, loss_function, + tr_solver, tr_options, verbose) + + result.message = TERMINATION_MESSAGES[result.status] + result.success = result.status > 0 + + if verbose >= 1: + print(result.message) + print("Function evaluations {}, initial cost {:.4e}, final cost " + "{:.4e}, first-order optimality {:.2e}." + .format(result.nfev, initial_cost, result.cost, + result.optimality)) + + return result diff --git a/parrot/lib/python3.10/site-packages/scipy/optimize/_lsq/lsq_linear.py b/parrot/lib/python3.10/site-packages/scipy/optimize/_lsq/lsq_linear.py new file mode 100644 index 0000000000000000000000000000000000000000..fdf4d26020109d55a6aea2be3009181a388c722d --- /dev/null +++ b/parrot/lib/python3.10/site-packages/scipy/optimize/_lsq/lsq_linear.py @@ -0,0 +1,362 @@ +"""Linear least squares with bound constraints on independent variables.""" +import numpy as np +from numpy.linalg import norm +from scipy.sparse import issparse, csr_matrix +from scipy.sparse.linalg import LinearOperator, lsmr +from scipy.optimize import OptimizeResult +from scipy.optimize._minimize import Bounds + +from .common import in_bounds, compute_grad +from .trf_linear import trf_linear +from .bvls import bvls + + +def prepare_bounds(bounds, n): + if len(bounds) != 2: + raise ValueError("`bounds` must contain 2 elements.") + lb, ub = (np.asarray(b, dtype=float) for b in bounds) + + if lb.ndim == 0: + lb = np.resize(lb, n) + + if ub.ndim == 0: + ub = np.resize(ub, n) + + return lb, ub + + +TERMINATION_MESSAGES = { + -1: "The algorithm was not able to make progress on the last iteration.", + 0: "The maximum number of iterations is exceeded.", + 1: "The first-order optimality measure is less than `tol`.", + 2: "The relative change of the cost function is less than `tol`.", + 3: "The unconstrained solution is optimal." +} + + +def lsq_linear(A, b, bounds=(-np.inf, np.inf), method='trf', tol=1e-10, + lsq_solver=None, lsmr_tol=None, max_iter=None, + verbose=0, *, lsmr_maxiter=None,): + r"""Solve a linear least-squares problem with bounds on the variables. + + Given a m-by-n design matrix A and a target vector b with m elements, + `lsq_linear` solves the following optimization problem:: + + minimize 0.5 * ||A x - b||**2 + subject to lb <= x <= ub + + This optimization problem is convex, hence a found minimum (if iterations + have converged) is guaranteed to be global. + + Parameters + ---------- + A : array_like, sparse matrix of LinearOperator, shape (m, n) + Design matrix. Can be `scipy.sparse.linalg.LinearOperator`. + b : array_like, shape (m,) + Target vector. + bounds : 2-tuple of array_like or `Bounds`, optional + Lower and upper bounds on parameters. Defaults to no bounds. + There are two ways to specify the bounds: + + - Instance of `Bounds` class. + + - 2-tuple of array_like: Each element of the tuple must be either + an array with the length equal to the number of parameters, or a + scalar (in which case the bound is taken to be the same for all + parameters). Use ``np.inf`` with an appropriate sign to disable + bounds on all or some parameters. + + method : 'trf' or 'bvls', optional + Method to perform minimization. + + * 'trf' : Trust Region Reflective algorithm adapted for a linear + least-squares problem. This is an interior-point-like method + and the required number of iterations is weakly correlated with + the number of variables. + * 'bvls' : Bounded-variable least-squares algorithm. This is + an active set method, which requires the number of iterations + comparable to the number of variables. Can't be used when `A` is + sparse or LinearOperator. + + Default is 'trf'. + tol : float, optional + Tolerance parameter. The algorithm terminates if a relative change + of the cost function is less than `tol` on the last iteration. + Additionally, the first-order optimality measure is considered: + + * ``method='trf'`` terminates if the uniform norm of the gradient, + scaled to account for the presence of the bounds, is less than + `tol`. + * ``method='bvls'`` terminates if Karush-Kuhn-Tucker conditions + are satisfied within `tol` tolerance. + + lsq_solver : {None, 'exact', 'lsmr'}, optional + Method of solving unbounded least-squares problems throughout + iterations: + + * 'exact' : Use dense QR or SVD decomposition approach. Can't be + used when `A` is sparse or LinearOperator. + * 'lsmr' : Use `scipy.sparse.linalg.lsmr` iterative procedure + which requires only matrix-vector product evaluations. Can't + be used with ``method='bvls'``. + + If None (default), the solver is chosen based on type of `A`. + lsmr_tol : None, float or 'auto', optional + Tolerance parameters 'atol' and 'btol' for `scipy.sparse.linalg.lsmr` + If None (default), it is set to ``1e-2 * tol``. If 'auto', the + tolerance will be adjusted based on the optimality of the current + iterate, which can speed up the optimization process, but is not always + reliable. + max_iter : None or int, optional + Maximum number of iterations before termination. If None (default), it + is set to 100 for ``method='trf'`` or to the number of variables for + ``method='bvls'`` (not counting iterations for 'bvls' initialization). + verbose : {0, 1, 2}, optional + Level of algorithm's verbosity: + + * 0 : work silently (default). + * 1 : display a termination report. + * 2 : display progress during iterations. + lsmr_maxiter : None or int, optional + Maximum number of iterations for the lsmr least squares solver, + if it is used (by setting ``lsq_solver='lsmr'``). If None (default), it + uses lsmr's default of ``min(m, n)`` where ``m`` and ``n`` are the + number of rows and columns of `A`, respectively. Has no effect if + ``lsq_solver='exact'``. + + Returns + ------- + OptimizeResult with the following fields defined: + x : ndarray, shape (n,) + Solution found. + cost : float + Value of the cost function at the solution. + fun : ndarray, shape (m,) + Vector of residuals at the solution. + optimality : float + First-order optimality measure. The exact meaning depends on `method`, + refer to the description of `tol` parameter. + active_mask : ndarray of int, shape (n,) + Each component shows whether a corresponding constraint is active + (that is, whether a variable is at the bound): + + * 0 : a constraint is not active. + * -1 : a lower bound is active. + * 1 : an upper bound is active. + + Might be somewhat arbitrary for the `trf` method as it generates a + sequence of strictly feasible iterates and active_mask is determined + within a tolerance threshold. + unbounded_sol : tuple + Unbounded least squares solution tuple returned by the least squares + solver (set with `lsq_solver` option). If `lsq_solver` is not set or is + set to ``'exact'``, the tuple contains an ndarray of shape (n,) with + the unbounded solution, an ndarray with the sum of squared residuals, + an int with the rank of `A`, and an ndarray with the singular values + of `A` (see NumPy's ``linalg.lstsq`` for more information). If + `lsq_solver` is set to ``'lsmr'``, the tuple contains an ndarray of + shape (n,) with the unbounded solution, an int with the exit code, + an int with the number of iterations, and five floats with + various norms and the condition number of `A` (see SciPy's + ``sparse.linalg.lsmr`` for more information). This output can be + useful for determining the convergence of the least squares solver, + particularly the iterative ``'lsmr'`` solver. The unbounded least + squares problem is to minimize ``0.5 * ||A x - b||**2``. + nit : int + Number of iterations. Zero if the unconstrained solution is optimal. + status : int + Reason for algorithm termination: + + * -1 : the algorithm was not able to make progress on the last + iteration. + * 0 : the maximum number of iterations is exceeded. + * 1 : the first-order optimality measure is less than `tol`. + * 2 : the relative change of the cost function is less than `tol`. + * 3 : the unconstrained solution is optimal. + + message : str + Verbal description of the termination reason. + success : bool + True if one of the convergence criteria is satisfied (`status` > 0). + + See Also + -------- + nnls : Linear least squares with non-negativity constraint. + least_squares : Nonlinear least squares with bounds on the variables. + + Notes + ----- + The algorithm first computes the unconstrained least-squares solution by + `numpy.linalg.lstsq` or `scipy.sparse.linalg.lsmr` depending on + `lsq_solver`. This solution is returned as optimal if it lies within the + bounds. + + Method 'trf' runs the adaptation of the algorithm described in [STIR]_ for + a linear least-squares problem. The iterations are essentially the same as + in the nonlinear least-squares algorithm, but as the quadratic function + model is always accurate, we don't need to track or modify the radius of + a trust region. The line search (backtracking) is used as a safety net + when a selected step does not decrease the cost function. Read more + detailed description of the algorithm in `scipy.optimize.least_squares`. + + Method 'bvls' runs a Python implementation of the algorithm described in + [BVLS]_. The algorithm maintains active and free sets of variables, on + each iteration chooses a new variable to move from the active set to the + free set and then solves the unconstrained least-squares problem on free + variables. This algorithm is guaranteed to give an accurate solution + eventually, but may require up to n iterations for a problem with n + variables. Additionally, an ad-hoc initialization procedure is + implemented, that determines which variables to set free or active + initially. It takes some number of iterations before actual BVLS starts, + but can significantly reduce the number of further iterations. + + References + ---------- + .. [STIR] M. A. Branch, T. F. Coleman, and Y. Li, "A Subspace, Interior, + and Conjugate Gradient Method for Large-Scale Bound-Constrained + Minimization Problems," SIAM Journal on Scientific Computing, + Vol. 21, Number 1, pp 1-23, 1999. + .. [BVLS] P. B. Start and R. L. Parker, "Bounded-Variable Least-Squares: + an Algorithm and Applications", Computational Statistics, 10, + 129-141, 1995. + + Examples + -------- + In this example, a problem with a large sparse matrix and bounds on the + variables is solved. + + >>> import numpy as np + >>> from scipy.sparse import rand + >>> from scipy.optimize import lsq_linear + >>> rng = np.random.default_rng() + ... + >>> m = 20000 + >>> n = 10000 + ... + >>> A = rand(m, n, density=1e-4, random_state=rng) + >>> b = rng.standard_normal(m) + ... + >>> lb = rng.standard_normal(n) + >>> ub = lb + 1 + ... + >>> res = lsq_linear(A, b, bounds=(lb, ub), lsmr_tol='auto', verbose=1) + # may vary + The relative change of the cost function is less than `tol`. + Number of iterations 16, initial cost 1.5039e+04, final cost 1.1112e+04, + first-order optimality 4.66e-08. + """ + if method not in ['trf', 'bvls']: + raise ValueError("`method` must be 'trf' or 'bvls'") + + if lsq_solver not in [None, 'exact', 'lsmr']: + raise ValueError("`solver` must be None, 'exact' or 'lsmr'.") + + if verbose not in [0, 1, 2]: + raise ValueError("`verbose` must be in [0, 1, 2].") + + if issparse(A): + A = csr_matrix(A) + elif not isinstance(A, LinearOperator): + A = np.atleast_2d(np.asarray(A)) + + if method == 'bvls': + if lsq_solver == 'lsmr': + raise ValueError("method='bvls' can't be used with " + "lsq_solver='lsmr'") + + if not isinstance(A, np.ndarray): + raise ValueError("method='bvls' can't be used with `A` being " + "sparse or LinearOperator.") + + if lsq_solver is None: + if isinstance(A, np.ndarray): + lsq_solver = 'exact' + else: + lsq_solver = 'lsmr' + elif lsq_solver == 'exact' and not isinstance(A, np.ndarray): + raise ValueError("`exact` solver can't be used when `A` is " + "sparse or LinearOperator.") + + if len(A.shape) != 2: # No ndim for LinearOperator. + raise ValueError("`A` must have at most 2 dimensions.") + + if max_iter is not None and max_iter <= 0: + raise ValueError("`max_iter` must be None or positive integer.") + + m, n = A.shape + + b = np.atleast_1d(b) + if b.ndim != 1: + raise ValueError("`b` must have at most 1 dimension.") + + if b.size != m: + raise ValueError("Inconsistent shapes between `A` and `b`.") + + if isinstance(bounds, Bounds): + lb = bounds.lb + ub = bounds.ub + else: + lb, ub = prepare_bounds(bounds, n) + + if lb.shape != (n,) and ub.shape != (n,): + raise ValueError("Bounds have wrong shape.") + + if np.any(lb >= ub): + raise ValueError("Each lower bound must be strictly less than each " + "upper bound.") + + if lsmr_maxiter is not None and lsmr_maxiter < 1: + raise ValueError("`lsmr_maxiter` must be None or positive integer.") + + if not ((isinstance(lsmr_tol, float) and lsmr_tol > 0) or + lsmr_tol in ('auto', None)): + raise ValueError("`lsmr_tol` must be None, 'auto', or positive float.") + + if lsq_solver == 'exact': + unbd_lsq = np.linalg.lstsq(A, b, rcond=-1) + elif lsq_solver == 'lsmr': + first_lsmr_tol = lsmr_tol # tol of first call to lsmr + if lsmr_tol is None or lsmr_tol == 'auto': + first_lsmr_tol = 1e-2 * tol # default if lsmr_tol not defined + unbd_lsq = lsmr(A, b, maxiter=lsmr_maxiter, + atol=first_lsmr_tol, btol=first_lsmr_tol) + x_lsq = unbd_lsq[0] # extract the solution from the least squares solver + + if in_bounds(x_lsq, lb, ub): + r = A @ x_lsq - b + cost = 0.5 * np.dot(r, r) + termination_status = 3 + termination_message = TERMINATION_MESSAGES[termination_status] + g = compute_grad(A, r) + g_norm = norm(g, ord=np.inf) + + if verbose > 0: + print(termination_message) + print(f"Final cost {cost:.4e}, first-order optimality {g_norm:.2e}") + + return OptimizeResult( + x=x_lsq, fun=r, cost=cost, optimality=g_norm, + active_mask=np.zeros(n), unbounded_sol=unbd_lsq, + nit=0, status=termination_status, + message=termination_message, success=True) + + if method == 'trf': + res = trf_linear(A, b, x_lsq, lb, ub, tol, lsq_solver, lsmr_tol, + max_iter, verbose, lsmr_maxiter=lsmr_maxiter) + elif method == 'bvls': + res = bvls(A, b, x_lsq, lb, ub, tol, max_iter, verbose) + + res.unbounded_sol = unbd_lsq + res.message = TERMINATION_MESSAGES[res.status] + res.success = res.status > 0 + + if verbose > 0: + print(res.message) + print( + f"Number of iterations {res.nit}, initial cost {res.initial_cost:.4e}, " + f"final cost {res.cost:.4e}, first-order optimality {res.optimality:.2e}." + ) + + del res.initial_cost + + return res diff --git a/parrot/lib/python3.10/site-packages/scipy/optimize/_lsq/trf.py b/parrot/lib/python3.10/site-packages/scipy/optimize/_lsq/trf.py new file mode 100644 index 0000000000000000000000000000000000000000..9154bdba5b2cc41883811ba1820dfc251e515d6c --- /dev/null +++ b/parrot/lib/python3.10/site-packages/scipy/optimize/_lsq/trf.py @@ -0,0 +1,560 @@ +"""Trust Region Reflective algorithm for least-squares optimization. + +The algorithm is based on ideas from paper [STIR]_. The main idea is to +account for the presence of the bounds by appropriate scaling of the variables (or, +equivalently, changing a trust-region shape). Let's introduce a vector v: + + | ub[i] - x[i], if g[i] < 0 and ub[i] < np.inf + v[i] = | x[i] - lb[i], if g[i] > 0 and lb[i] > -np.inf + | 1, otherwise + +where g is the gradient of a cost function and lb, ub are the bounds. Its +components are distances to the bounds at which the anti-gradient points (if +this distance is finite). Define a scaling matrix D = diag(v**0.5). +First-order optimality conditions can be stated as + + D^2 g(x) = 0. + +Meaning that components of the gradient should be zero for strictly interior +variables, and components must point inside the feasible region for variables +on the bound. + +Now consider this system of equations as a new optimization problem. If the +point x is strictly interior (not on the bound), then the left-hand side is +differentiable and the Newton step for it satisfies + + (D^2 H + diag(g) Jv) p = -D^2 g + +where H is the Hessian matrix (or its J^T J approximation in least squares), +Jv is the Jacobian matrix of v with components -1, 1 or 0, such that all +elements of matrix C = diag(g) Jv are non-negative. Introduce the change +of the variables x = D x_h (_h would be "hat" in LaTeX). In the new variables, +we have a Newton step satisfying + + B_h p_h = -g_h, + +where B_h = D H D + C, g_h = D g. In least squares B_h = J_h^T J_h, where +J_h = J D. Note that J_h and g_h are proper Jacobian and gradient with respect +to "hat" variables. To guarantee global convergence we formulate a +trust-region problem based on the Newton step in the new variables: + + 0.5 * p_h^T B_h p + g_h^T p_h -> min, ||p_h|| <= Delta + +In the original space B = H + D^{-1} C D^{-1}, and the equivalent trust-region +problem is + + 0.5 * p^T B p + g^T p -> min, ||D^{-1} p|| <= Delta + +Here, the meaning of the matrix D becomes more clear: it alters the shape +of a trust-region, such that large steps towards the bounds are not allowed. +In the implementation, the trust-region problem is solved in "hat" space, +but handling of the bounds is done in the original space (see below and read +the code). + +The introduction of the matrix D doesn't allow to ignore bounds, the algorithm +must keep iterates strictly feasible (to satisfy aforementioned +differentiability), the parameter theta controls step back from the boundary +(see the code for details). + +The algorithm does another important trick. If the trust-region solution +doesn't fit into the bounds, then a reflected (from a firstly encountered +bound) search direction is considered. For motivation and analysis refer to +[STIR]_ paper (and other papers of the authors). In practice, it doesn't need +a lot of justifications, the algorithm simply chooses the best step among +three: a constrained trust-region step, a reflected step and a constrained +Cauchy step (a minimizer along -g_h in "hat" space, or -D^2 g in the original +space). + +Another feature is that a trust-region radius control strategy is modified to +account for appearance of the diagonal C matrix (called diag_h in the code). + +Note that all described peculiarities are completely gone as we consider +problems without bounds (the algorithm becomes a standard trust-region type +algorithm very similar to ones implemented in MINPACK). + +The implementation supports two methods of solving the trust-region problem. +The first, called 'exact', applies SVD on Jacobian and then solves the problem +very accurately using the algorithm described in [JJMore]_. It is not +applicable to large problem. The second, called 'lsmr', uses the 2-D subspace +approach (sometimes called "indefinite dogleg"), where the problem is solved +in a subspace spanned by the gradient and the approximate Gauss-Newton step +found by ``scipy.sparse.linalg.lsmr``. A 2-D trust-region problem is +reformulated as a 4th order algebraic equation and solved very accurately by +``numpy.roots``. The subspace approach allows to solve very large problems +(up to couple of millions of residuals on a regular PC), provided the Jacobian +matrix is sufficiently sparse. + +References +---------- +.. [STIR] Branch, M.A., T.F. Coleman, and Y. Li, "A Subspace, Interior, + and Conjugate Gradient Method for Large-Scale Bound-Constrained + Minimization Problems," SIAM Journal on Scientific Computing, + Vol. 21, Number 1, pp 1-23, 1999. +.. [JJMore] More, J. J., "The Levenberg-Marquardt Algorithm: Implementation + and Theory," Numerical Analysis, ed. G. A. Watson, Lecture +""" +import numpy as np +from numpy.linalg import norm +from scipy.linalg import svd, qr +from scipy.sparse.linalg import lsmr +from scipy.optimize import OptimizeResult + +from .common import ( + step_size_to_bound, find_active_constraints, in_bounds, + make_strictly_feasible, intersect_trust_region, solve_lsq_trust_region, + solve_trust_region_2d, minimize_quadratic_1d, build_quadratic_1d, + evaluate_quadratic, right_multiplied_operator, regularized_lsq_operator, + CL_scaling_vector, compute_grad, compute_jac_scale, check_termination, + update_tr_radius, scale_for_robust_loss_function, print_header_nonlinear, + print_iteration_nonlinear) + + +def trf(fun, jac, x0, f0, J0, lb, ub, ftol, xtol, gtol, max_nfev, x_scale, + loss_function, tr_solver, tr_options, verbose): + # For efficiency, it makes sense to run the simplified version of the + # algorithm when no bounds are imposed. We decided to write the two + # separate functions. It violates the DRY principle, but the individual + # functions are kept the most readable. + if np.all(lb == -np.inf) and np.all(ub == np.inf): + return trf_no_bounds( + fun, jac, x0, f0, J0, ftol, xtol, gtol, max_nfev, x_scale, + loss_function, tr_solver, tr_options, verbose) + else: + return trf_bounds( + fun, jac, x0, f0, J0, lb, ub, ftol, xtol, gtol, max_nfev, x_scale, + loss_function, tr_solver, tr_options, verbose) + + +def select_step(x, J_h, diag_h, g_h, p, p_h, d, Delta, lb, ub, theta): + """Select the best step according to Trust Region Reflective algorithm.""" + if in_bounds(x + p, lb, ub): + p_value = evaluate_quadratic(J_h, g_h, p_h, diag=diag_h) + return p, p_h, -p_value + + p_stride, hits = step_size_to_bound(x, p, lb, ub) + + # Compute the reflected direction. + r_h = np.copy(p_h) + r_h[hits.astype(bool)] *= -1 + r = d * r_h + + # Restrict trust-region step, such that it hits the bound. + p *= p_stride + p_h *= p_stride + x_on_bound = x + p + + # Reflected direction will cross first either feasible region or trust + # region boundary. + _, to_tr = intersect_trust_region(p_h, r_h, Delta) + to_bound, _ = step_size_to_bound(x_on_bound, r, lb, ub) + + # Find lower and upper bounds on a step size along the reflected + # direction, considering the strict feasibility requirement. There is no + # single correct way to do that, the chosen approach seems to work best + # on test problems. + r_stride = min(to_bound, to_tr) + if r_stride > 0: + r_stride_l = (1 - theta) * p_stride / r_stride + if r_stride == to_bound: + r_stride_u = theta * to_bound + else: + r_stride_u = to_tr + else: + r_stride_l = 0 + r_stride_u = -1 + + # Check if reflection step is available. + if r_stride_l <= r_stride_u: + a, b, c = build_quadratic_1d(J_h, g_h, r_h, s0=p_h, diag=diag_h) + r_stride, r_value = minimize_quadratic_1d( + a, b, r_stride_l, r_stride_u, c=c) + r_h *= r_stride + r_h += p_h + r = r_h * d + else: + r_value = np.inf + + # Now correct p_h to make it strictly interior. + p *= theta + p_h *= theta + p_value = evaluate_quadratic(J_h, g_h, p_h, diag=diag_h) + + ag_h = -g_h + ag = d * ag_h + + to_tr = Delta / norm(ag_h) + to_bound, _ = step_size_to_bound(x, ag, lb, ub) + if to_bound < to_tr: + ag_stride = theta * to_bound + else: + ag_stride = to_tr + + a, b = build_quadratic_1d(J_h, g_h, ag_h, diag=diag_h) + ag_stride, ag_value = minimize_quadratic_1d(a, b, 0, ag_stride) + ag_h *= ag_stride + ag *= ag_stride + + if p_value < r_value and p_value < ag_value: + return p, p_h, -p_value + elif r_value < p_value and r_value < ag_value: + return r, r_h, -r_value + else: + return ag, ag_h, -ag_value + + +def trf_bounds(fun, jac, x0, f0, J0, lb, ub, ftol, xtol, gtol, max_nfev, + x_scale, loss_function, tr_solver, tr_options, verbose): + x = x0.copy() + + f = f0 + f_true = f.copy() + nfev = 1 + + J = J0 + njev = 1 + m, n = J.shape + + if loss_function is not None: + rho = loss_function(f) + cost = 0.5 * np.sum(rho[0]) + J, f = scale_for_robust_loss_function(J, f, rho) + else: + cost = 0.5 * np.dot(f, f) + + g = compute_grad(J, f) + + jac_scale = isinstance(x_scale, str) and x_scale == 'jac' + if jac_scale: + scale, scale_inv = compute_jac_scale(J) + else: + scale, scale_inv = x_scale, 1 / x_scale + + v, dv = CL_scaling_vector(x, g, lb, ub) + v[dv != 0] *= scale_inv[dv != 0] + Delta = norm(x0 * scale_inv / v**0.5) + if Delta == 0: + Delta = 1.0 + + g_norm = norm(g * v, ord=np.inf) + + f_augmented = np.zeros(m + n) + if tr_solver == 'exact': + J_augmented = np.empty((m + n, n)) + elif tr_solver == 'lsmr': + reg_term = 0.0 + regularize = tr_options.pop('regularize', True) + + if max_nfev is None: + max_nfev = x0.size * 100 + + alpha = 0.0 # "Levenberg-Marquardt" parameter + + termination_status = None + iteration = 0 + step_norm = None + actual_reduction = None + + if verbose == 2: + print_header_nonlinear() + + while True: + v, dv = CL_scaling_vector(x, g, lb, ub) + + g_norm = norm(g * v, ord=np.inf) + if g_norm < gtol: + termination_status = 1 + + if verbose == 2: + print_iteration_nonlinear(iteration, nfev, cost, actual_reduction, + step_norm, g_norm) + + if termination_status is not None or nfev == max_nfev: + break + + # Now compute variables in "hat" space. Here, we also account for + # scaling introduced by `x_scale` parameter. This part is a bit tricky, + # you have to write down the formulas and see how the trust-region + # problem is formulated when the two types of scaling are applied. + # The idea is that first we apply `x_scale` and then apply Coleman-Li + # approach in the new variables. + + # v is recomputed in the variables after applying `x_scale`, note that + # components which were identically 1 not affected. + v[dv != 0] *= scale_inv[dv != 0] + + # Here, we apply two types of scaling. + d = v**0.5 * scale + + # C = diag(g * scale) Jv + diag_h = g * dv * scale + + # After all this has been done, we continue normally. + + # "hat" gradient. + g_h = d * g + + f_augmented[:m] = f + if tr_solver == 'exact': + J_augmented[:m] = J * d + J_h = J_augmented[:m] # Memory view. + J_augmented[m:] = np.diag(diag_h**0.5) + U, s, V = svd(J_augmented, full_matrices=False) + V = V.T + uf = U.T.dot(f_augmented) + elif tr_solver == 'lsmr': + J_h = right_multiplied_operator(J, d) + + if regularize: + a, b = build_quadratic_1d(J_h, g_h, -g_h, diag=diag_h) + to_tr = Delta / norm(g_h) + ag_value = minimize_quadratic_1d(a, b, 0, to_tr)[1] + reg_term = -ag_value / Delta**2 + + lsmr_op = regularized_lsq_operator(J_h, (diag_h + reg_term)**0.5) + gn_h = lsmr(lsmr_op, f_augmented, **tr_options)[0] + S = np.vstack((g_h, gn_h)).T + S, _ = qr(S, mode='economic') + JS = J_h.dot(S) # LinearOperator does dot too. + B_S = np.dot(JS.T, JS) + np.dot(S.T * diag_h, S) + g_S = S.T.dot(g_h) + + # theta controls step back step ratio from the bounds. + theta = max(0.995, 1 - g_norm) + + actual_reduction = -1 + while actual_reduction <= 0 and nfev < max_nfev: + if tr_solver == 'exact': + p_h, alpha, n_iter = solve_lsq_trust_region( + n, m, uf, s, V, Delta, initial_alpha=alpha) + elif tr_solver == 'lsmr': + p_S, _ = solve_trust_region_2d(B_S, g_S, Delta) + p_h = S.dot(p_S) + + p = d * p_h # Trust-region solution in the original space. + step, step_h, predicted_reduction = select_step( + x, J_h, diag_h, g_h, p, p_h, d, Delta, lb, ub, theta) + + x_new = make_strictly_feasible(x + step, lb, ub, rstep=0) + f_new = fun(x_new) + nfev += 1 + + step_h_norm = norm(step_h) + + if not np.all(np.isfinite(f_new)): + Delta = 0.25 * step_h_norm + continue + + # Usual trust-region step quality estimation. + if loss_function is not None: + cost_new = loss_function(f_new, cost_only=True) + else: + cost_new = 0.5 * np.dot(f_new, f_new) + actual_reduction = cost - cost_new + Delta_new, ratio = update_tr_radius( + Delta, actual_reduction, predicted_reduction, + step_h_norm, step_h_norm > 0.95 * Delta) + + step_norm = norm(step) + termination_status = check_termination( + actual_reduction, cost, step_norm, norm(x), ratio, ftol, xtol) + if termination_status is not None: + break + + alpha *= Delta / Delta_new + Delta = Delta_new + + if actual_reduction > 0: + x = x_new + + f = f_new + f_true = f.copy() + + cost = cost_new + + J = jac(x, f) + njev += 1 + + if loss_function is not None: + rho = loss_function(f) + J, f = scale_for_robust_loss_function(J, f, rho) + + g = compute_grad(J, f) + + if jac_scale: + scale, scale_inv = compute_jac_scale(J, scale_inv) + else: + step_norm = 0 + actual_reduction = 0 + + iteration += 1 + + if termination_status is None: + termination_status = 0 + + active_mask = find_active_constraints(x, lb, ub, rtol=xtol) + return OptimizeResult( + x=x, cost=cost, fun=f_true, jac=J, grad=g, optimality=g_norm, + active_mask=active_mask, nfev=nfev, njev=njev, + status=termination_status) + + +def trf_no_bounds(fun, jac, x0, f0, J0, ftol, xtol, gtol, max_nfev, + x_scale, loss_function, tr_solver, tr_options, verbose): + x = x0.copy() + + f = f0 + f_true = f.copy() + nfev = 1 + + J = J0 + njev = 1 + m, n = J.shape + + if loss_function is not None: + rho = loss_function(f) + cost = 0.5 * np.sum(rho[0]) + J, f = scale_for_robust_loss_function(J, f, rho) + else: + cost = 0.5 * np.dot(f, f) + + g = compute_grad(J, f) + + jac_scale = isinstance(x_scale, str) and x_scale == 'jac' + if jac_scale: + scale, scale_inv = compute_jac_scale(J) + else: + scale, scale_inv = x_scale, 1 / x_scale + + Delta = norm(x0 * scale_inv) + if Delta == 0: + Delta = 1.0 + + if tr_solver == 'lsmr': + reg_term = 0 + damp = tr_options.pop('damp', 0.0) + regularize = tr_options.pop('regularize', True) + + if max_nfev is None: + max_nfev = x0.size * 100 + + alpha = 0.0 # "Levenberg-Marquardt" parameter + + termination_status = None + iteration = 0 + step_norm = None + actual_reduction = None + + if verbose == 2: + print_header_nonlinear() + + while True: + g_norm = norm(g, ord=np.inf) + if g_norm < gtol: + termination_status = 1 + + if verbose == 2: + print_iteration_nonlinear(iteration, nfev, cost, actual_reduction, + step_norm, g_norm) + + if termination_status is not None or nfev == max_nfev: + break + + d = scale + g_h = d * g + + if tr_solver == 'exact': + J_h = J * d + U, s, V = svd(J_h, full_matrices=False) + V = V.T + uf = U.T.dot(f) + elif tr_solver == 'lsmr': + J_h = right_multiplied_operator(J, d) + + if regularize: + a, b = build_quadratic_1d(J_h, g_h, -g_h) + to_tr = Delta / norm(g_h) + ag_value = minimize_quadratic_1d(a, b, 0, to_tr)[1] + reg_term = -ag_value / Delta**2 + + damp_full = (damp**2 + reg_term)**0.5 + gn_h = lsmr(J_h, f, damp=damp_full, **tr_options)[0] + S = np.vstack((g_h, gn_h)).T + S, _ = qr(S, mode='economic') + JS = J_h.dot(S) + B_S = np.dot(JS.T, JS) + g_S = S.T.dot(g_h) + + actual_reduction = -1 + while actual_reduction <= 0 and nfev < max_nfev: + if tr_solver == 'exact': + step_h, alpha, n_iter = solve_lsq_trust_region( + n, m, uf, s, V, Delta, initial_alpha=alpha) + elif tr_solver == 'lsmr': + p_S, _ = solve_trust_region_2d(B_S, g_S, Delta) + step_h = S.dot(p_S) + + predicted_reduction = -evaluate_quadratic(J_h, g_h, step_h) + step = d * step_h + x_new = x + step + f_new = fun(x_new) + nfev += 1 + + step_h_norm = norm(step_h) + + if not np.all(np.isfinite(f_new)): + Delta = 0.25 * step_h_norm + continue + + # Usual trust-region step quality estimation. + if loss_function is not None: + cost_new = loss_function(f_new, cost_only=True) + else: + cost_new = 0.5 * np.dot(f_new, f_new) + actual_reduction = cost - cost_new + + Delta_new, ratio = update_tr_radius( + Delta, actual_reduction, predicted_reduction, + step_h_norm, step_h_norm > 0.95 * Delta) + + step_norm = norm(step) + termination_status = check_termination( + actual_reduction, cost, step_norm, norm(x), ratio, ftol, xtol) + if termination_status is not None: + break + + alpha *= Delta / Delta_new + Delta = Delta_new + + if actual_reduction > 0: + x = x_new + + f = f_new + f_true = f.copy() + + cost = cost_new + + J = jac(x, f) + njev += 1 + + if loss_function is not None: + rho = loss_function(f) + J, f = scale_for_robust_loss_function(J, f, rho) + + g = compute_grad(J, f) + + if jac_scale: + scale, scale_inv = compute_jac_scale(J, scale_inv) + else: + step_norm = 0 + actual_reduction = 0 + + iteration += 1 + + if termination_status is None: + termination_status = 0 + + active_mask = np.zeros_like(x) + return OptimizeResult( + x=x, cost=cost, fun=f_true, jac=J, grad=g, optimality=g_norm, + active_mask=active_mask, nfev=nfev, njev=njev, + status=termination_status) diff --git a/parrot/lib/python3.10/site-packages/scipy/optimize/_lsq/trf_linear.py b/parrot/lib/python3.10/site-packages/scipy/optimize/_lsq/trf_linear.py new file mode 100644 index 0000000000000000000000000000000000000000..dd752763179bcf97945c7f34ce6a9e49e85c819e --- /dev/null +++ b/parrot/lib/python3.10/site-packages/scipy/optimize/_lsq/trf_linear.py @@ -0,0 +1,249 @@ +"""The adaptation of Trust Region Reflective algorithm for a linear +least-squares problem.""" +import numpy as np +from numpy.linalg import norm +from scipy.linalg import qr, solve_triangular +from scipy.sparse.linalg import lsmr +from scipy.optimize import OptimizeResult + +from .givens_elimination import givens_elimination +from .common import ( + EPS, step_size_to_bound, find_active_constraints, in_bounds, + make_strictly_feasible, build_quadratic_1d, evaluate_quadratic, + minimize_quadratic_1d, CL_scaling_vector, reflective_transformation, + print_header_linear, print_iteration_linear, compute_grad, + regularized_lsq_operator, right_multiplied_operator) + + +def regularized_lsq_with_qr(m, n, R, QTb, perm, diag, copy_R=True): + """Solve regularized least squares using information from QR-decomposition. + + The initial problem is to solve the following system in a least-squares + sense:: + + A x = b + D x = 0 + + where D is diagonal matrix. The method is based on QR decomposition + of the form A P = Q R, where P is a column permutation matrix, Q is an + orthogonal matrix and R is an upper triangular matrix. + + Parameters + ---------- + m, n : int + Initial shape of A. + R : ndarray, shape (n, n) + Upper triangular matrix from QR decomposition of A. + QTb : ndarray, shape (n,) + First n components of Q^T b. + perm : ndarray, shape (n,) + Array defining column permutation of A, such that ith column of + P is perm[i]-th column of identity matrix. + diag : ndarray, shape (n,) + Array containing diagonal elements of D. + + Returns + ------- + x : ndarray, shape (n,) + Found least-squares solution. + """ + if copy_R: + R = R.copy() + v = QTb.copy() + + givens_elimination(R, v, diag[perm]) + + abs_diag_R = np.abs(np.diag(R)) + threshold = EPS * max(m, n) * np.max(abs_diag_R) + nns, = np.nonzero(abs_diag_R > threshold) + + R = R[np.ix_(nns, nns)] + v = v[nns] + + x = np.zeros(n) + x[perm[nns]] = solve_triangular(R, v) + + return x + + +def backtracking(A, g, x, p, theta, p_dot_g, lb, ub): + """Find an appropriate step size using backtracking line search.""" + alpha = 1 + while True: + x_new, _ = reflective_transformation(x + alpha * p, lb, ub) + step = x_new - x + cost_change = -evaluate_quadratic(A, g, step) + if cost_change > -0.1 * alpha * p_dot_g: + break + alpha *= 0.5 + + active = find_active_constraints(x_new, lb, ub) + if np.any(active != 0): + x_new, _ = reflective_transformation(x + theta * alpha * p, lb, ub) + x_new = make_strictly_feasible(x_new, lb, ub, rstep=0) + step = x_new - x + cost_change = -evaluate_quadratic(A, g, step) + + return x, step, cost_change + + +def select_step(x, A_h, g_h, c_h, p, p_h, d, lb, ub, theta): + """Select the best step according to Trust Region Reflective algorithm.""" + if in_bounds(x + p, lb, ub): + return p + + p_stride, hits = step_size_to_bound(x, p, lb, ub) + r_h = np.copy(p_h) + r_h[hits.astype(bool)] *= -1 + r = d * r_h + + # Restrict step, such that it hits the bound. + p *= p_stride + p_h *= p_stride + x_on_bound = x + p + + # Find the step size along reflected direction. + r_stride_u, _ = step_size_to_bound(x_on_bound, r, lb, ub) + + # Stay interior. + r_stride_l = (1 - theta) * r_stride_u + r_stride_u *= theta + + if r_stride_u > 0: + a, b, c = build_quadratic_1d(A_h, g_h, r_h, s0=p_h, diag=c_h) + r_stride, r_value = minimize_quadratic_1d( + a, b, r_stride_l, r_stride_u, c=c) + r_h = p_h + r_h * r_stride + r = d * r_h + else: + r_value = np.inf + + # Now correct p_h to make it strictly interior. + p_h *= theta + p *= theta + p_value = evaluate_quadratic(A_h, g_h, p_h, diag=c_h) + + ag_h = -g_h + ag = d * ag_h + ag_stride_u, _ = step_size_to_bound(x, ag, lb, ub) + ag_stride_u *= theta + a, b = build_quadratic_1d(A_h, g_h, ag_h, diag=c_h) + ag_stride, ag_value = minimize_quadratic_1d(a, b, 0, ag_stride_u) + ag *= ag_stride + + if p_value < r_value and p_value < ag_value: + return p + elif r_value < p_value and r_value < ag_value: + return r + else: + return ag + + +def trf_linear(A, b, x_lsq, lb, ub, tol, lsq_solver, lsmr_tol, + max_iter, verbose, *, lsmr_maxiter=None): + m, n = A.shape + x, _ = reflective_transformation(x_lsq, lb, ub) + x = make_strictly_feasible(x, lb, ub, rstep=0.1) + + if lsq_solver == 'exact': + QT, R, perm = qr(A, mode='economic', pivoting=True) + QT = QT.T + + if m < n: + R = np.vstack((R, np.zeros((n - m, n)))) + + QTr = np.zeros(n) + k = min(m, n) + elif lsq_solver == 'lsmr': + r_aug = np.zeros(m + n) + auto_lsmr_tol = False + if lsmr_tol is None: + lsmr_tol = 1e-2 * tol + elif lsmr_tol == 'auto': + auto_lsmr_tol = True + + r = A.dot(x) - b + g = compute_grad(A, r) + cost = 0.5 * np.dot(r, r) + initial_cost = cost + + termination_status = None + step_norm = None + cost_change = None + + if max_iter is None: + max_iter = 100 + + if verbose == 2: + print_header_linear() + + for iteration in range(max_iter): + v, dv = CL_scaling_vector(x, g, lb, ub) + g_scaled = g * v + g_norm = norm(g_scaled, ord=np.inf) + if g_norm < tol: + termination_status = 1 + + if verbose == 2: + print_iteration_linear(iteration, cost, cost_change, + step_norm, g_norm) + + if termination_status is not None: + break + + diag_h = g * dv + diag_root_h = diag_h ** 0.5 + d = v ** 0.5 + g_h = d * g + + A_h = right_multiplied_operator(A, d) + if lsq_solver == 'exact': + QTr[:k] = QT.dot(r) + p_h = -regularized_lsq_with_qr(m, n, R * d[perm], QTr, perm, + diag_root_h, copy_R=False) + elif lsq_solver == 'lsmr': + lsmr_op = regularized_lsq_operator(A_h, diag_root_h) + r_aug[:m] = r + if auto_lsmr_tol: + eta = 1e-2 * min(0.5, g_norm) + lsmr_tol = max(EPS, min(0.1, eta * g_norm)) + p_h = -lsmr(lsmr_op, r_aug, maxiter=lsmr_maxiter, + atol=lsmr_tol, btol=lsmr_tol)[0] + + p = d * p_h + + p_dot_g = np.dot(p, g) + if p_dot_g > 0: + termination_status = -1 + + theta = 1 - min(0.005, g_norm) + step = select_step(x, A_h, g_h, diag_h, p, p_h, d, lb, ub, theta) + cost_change = -evaluate_quadratic(A, g, step) + + # Perhaps almost never executed, the idea is that `p` is descent + # direction thus we must find acceptable cost decrease using simple + # "backtracking", otherwise the algorithm's logic would break. + if cost_change < 0: + x, step, cost_change = backtracking( + A, g, x, p, theta, p_dot_g, lb, ub) + else: + x = make_strictly_feasible(x + step, lb, ub, rstep=0) + + step_norm = norm(step) + r = A.dot(x) - b + g = compute_grad(A, r) + + if cost_change < tol * cost: + termination_status = 2 + + cost = 0.5 * np.dot(r, r) + + if termination_status is None: + termination_status = 0 + + active_mask = find_active_constraints(x, lb, ub, rtol=tol) + + return OptimizeResult( + x=x, fun=r, cost=cost, optimality=g_norm, active_mask=active_mask, + nit=iteration + 1, status=termination_status, + initial_cost=initial_cost) diff --git a/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/__pycache__/__init__.cpython-310.pyc b/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/__pycache__/__init__.cpython-310.pyc new file mode 100644 index 0000000000000000000000000000000000000000..2e0fac506dd0fffd0d95b80957f7e91a3ed739f4 Binary files /dev/null and b/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/__pycache__/__init__.cpython-310.pyc differ diff --git a/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/__pycache__/canonical_constraint.cpython-310.pyc b/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/__pycache__/canonical_constraint.cpython-310.pyc new file mode 100644 index 0000000000000000000000000000000000000000..180ec8ee20cc65aa3d73170883f2b4b38dd9928b Binary files /dev/null and b/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/__pycache__/canonical_constraint.cpython-310.pyc differ diff --git a/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/__pycache__/projections.cpython-310.pyc b/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/__pycache__/projections.cpython-310.pyc new file mode 100644 index 0000000000000000000000000000000000000000..b1fbcc1d3533466eee9c989f0774b7f1ff0a86ab Binary files /dev/null and b/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/__pycache__/projections.cpython-310.pyc differ diff --git a/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/equality_constrained_sqp.py b/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/equality_constrained_sqp.py new file mode 100644 index 0000000000000000000000000000000000000000..fb4c05dcdd03fb990d3418220a398e249ef581ee --- /dev/null +++ b/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/equality_constrained_sqp.py @@ -0,0 +1,231 @@ +"""Byrd-Omojokun Trust-Region SQP method.""" + +from scipy.sparse import eye as speye +from .projections import projections +from .qp_subproblem import modified_dogleg, projected_cg, box_intersections +import numpy as np +from numpy.linalg import norm + +__all__ = ['equality_constrained_sqp'] + + +def default_scaling(x): + n, = np.shape(x) + return speye(n) + + +def equality_constrained_sqp(fun_and_constr, grad_and_jac, lagr_hess, + x0, fun0, grad0, constr0, + jac0, stop_criteria, + state, + initial_penalty, + initial_trust_radius, + factorization_method, + trust_lb=None, + trust_ub=None, + scaling=default_scaling): + """Solve nonlinear equality-constrained problem using trust-region SQP. + + Solve optimization problem: + + minimize fun(x) + subject to: constr(x) = 0 + + using Byrd-Omojokun Trust-Region SQP method described in [1]_. Several + implementation details are based on [2]_ and [3]_, p. 549. + + References + ---------- + .. [1] Lalee, Marucha, Jorge Nocedal, and Todd Plantenga. "On the + implementation of an algorithm for large-scale equality + constrained optimization." SIAM Journal on + Optimization 8.3 (1998): 682-706. + .. [2] Byrd, Richard H., Mary E. Hribar, and Jorge Nocedal. + "An interior point algorithm for large-scale nonlinear + programming." SIAM Journal on Optimization 9.4 (1999): 877-900. + .. [3] Nocedal, Jorge, and Stephen J. Wright. "Numerical optimization" + Second Edition (2006). + """ + PENALTY_FACTOR = 0.3 # Rho from formula (3.51), reference [2]_, p.891. + LARGE_REDUCTION_RATIO = 0.9 + INTERMEDIARY_REDUCTION_RATIO = 0.3 + SUFFICIENT_REDUCTION_RATIO = 1e-8 # Eta from reference [2]_, p.892. + TRUST_ENLARGEMENT_FACTOR_L = 7.0 + TRUST_ENLARGEMENT_FACTOR_S = 2.0 + MAX_TRUST_REDUCTION = 0.5 + MIN_TRUST_REDUCTION = 0.1 + SOC_THRESHOLD = 0.1 + TR_FACTOR = 0.8 # Zeta from formula (3.21), reference [2]_, p.885. + BOX_FACTOR = 0.5 + + n, = np.shape(x0) # Number of parameters + + # Set default lower and upper bounds. + if trust_lb is None: + trust_lb = np.full(n, -np.inf) + if trust_ub is None: + trust_ub = np.full(n, np.inf) + + # Initial values + x = np.copy(x0) + trust_radius = initial_trust_radius + penalty = initial_penalty + # Compute Values + f = fun0 + c = grad0 + b = constr0 + A = jac0 + S = scaling(x) + # Get projections + try: + Z, LS, Y = projections(A, factorization_method) + except ValueError as e: + if str(e) == "expected square matrix": + # can be the case if there are more equality + # constraints than independent variables + raise ValueError( + "The 'expected square matrix' error can occur if there are" + " more equality constraints than independent variables." + " Consider how your constraints are set up, or use" + " factorization_method='SVDFactorization'." + ) from e + else: + raise e + + # Compute least-square lagrange multipliers + v = -LS.dot(c) + # Compute Hessian + H = lagr_hess(x, v) + + # Update state parameters + optimality = norm(c + A.T.dot(v), np.inf) + constr_violation = norm(b, np.inf) if len(b) > 0 else 0 + cg_info = {'niter': 0, 'stop_cond': 0, + 'hits_boundary': False} + + last_iteration_failed = False + while not stop_criteria(state, x, last_iteration_failed, + optimality, constr_violation, + trust_radius, penalty, cg_info): + # Normal Step - `dn` + # minimize 1/2*||A dn + b||^2 + # subject to: + # ||dn|| <= TR_FACTOR * trust_radius + # BOX_FACTOR * lb <= dn <= BOX_FACTOR * ub. + dn = modified_dogleg(A, Y, b, + TR_FACTOR*trust_radius, + BOX_FACTOR*trust_lb, + BOX_FACTOR*trust_ub) + + # Tangential Step - `dt` + # Solve the QP problem: + # minimize 1/2 dt.T H dt + dt.T (H dn + c) + # subject to: + # A dt = 0 + # ||dt|| <= sqrt(trust_radius**2 - ||dn||**2) + # lb - dn <= dt <= ub - dn + c_t = H.dot(dn) + c + b_t = np.zeros_like(b) + trust_radius_t = np.sqrt(trust_radius**2 - np.linalg.norm(dn)**2) + lb_t = trust_lb - dn + ub_t = trust_ub - dn + dt, cg_info = projected_cg(H, c_t, Z, Y, b_t, + trust_radius_t, + lb_t, ub_t) + + # Compute update (normal + tangential steps). + d = dn + dt + + # Compute second order model: 1/2 d H d + c.T d + f. + quadratic_model = 1/2*(H.dot(d)).dot(d) + c.T.dot(d) + # Compute linearized constraint: l = A d + b. + linearized_constr = A.dot(d)+b + # Compute new penalty parameter according to formula (3.52), + # reference [2]_, p.891. + vpred = norm(b) - norm(linearized_constr) + # Guarantee `vpred` always positive, + # regardless of roundoff errors. + vpred = max(1e-16, vpred) + previous_penalty = penalty + if quadratic_model > 0: + new_penalty = quadratic_model / ((1-PENALTY_FACTOR)*vpred) + penalty = max(penalty, new_penalty) + # Compute predicted reduction according to formula (3.52), + # reference [2]_, p.891. + predicted_reduction = -quadratic_model + penalty*vpred + + # Compute merit function at current point + merit_function = f + penalty*norm(b) + # Evaluate function and constraints at trial point + x_next = x + S.dot(d) + f_next, b_next = fun_and_constr(x_next) + # Compute merit function at trial point + merit_function_next = f_next + penalty*norm(b_next) + # Compute actual reduction according to formula (3.54), + # reference [2]_, p.892. + actual_reduction = merit_function - merit_function_next + # Compute reduction ratio + reduction_ratio = actual_reduction / predicted_reduction + + # Second order correction (SOC), reference [2]_, p.892. + if reduction_ratio < SUFFICIENT_REDUCTION_RATIO and \ + norm(dn) <= SOC_THRESHOLD * norm(dt): + # Compute second order correction + y = -Y.dot(b_next) + # Make sure increment is inside box constraints + _, t, intersect = box_intersections(d, y, trust_lb, trust_ub) + # Compute tentative point + x_soc = x + S.dot(d + t*y) + f_soc, b_soc = fun_and_constr(x_soc) + # Recompute actual reduction + merit_function_soc = f_soc + penalty*norm(b_soc) + actual_reduction_soc = merit_function - merit_function_soc + # Recompute reduction ratio + reduction_ratio_soc = actual_reduction_soc / predicted_reduction + if intersect and reduction_ratio_soc >= SUFFICIENT_REDUCTION_RATIO: + x_next = x_soc + f_next = f_soc + b_next = b_soc + reduction_ratio = reduction_ratio_soc + + # Readjust trust region step, formula (3.55), reference [2]_, p.892. + if reduction_ratio >= LARGE_REDUCTION_RATIO: + trust_radius = max(TRUST_ENLARGEMENT_FACTOR_L * norm(d), + trust_radius) + elif reduction_ratio >= INTERMEDIARY_REDUCTION_RATIO: + trust_radius = max(TRUST_ENLARGEMENT_FACTOR_S * norm(d), + trust_radius) + # Reduce trust region step, according to reference [3]_, p.696. + elif reduction_ratio < SUFFICIENT_REDUCTION_RATIO: + trust_reduction = ((1-SUFFICIENT_REDUCTION_RATIO) / + (1-reduction_ratio)) + new_trust_radius = trust_reduction * norm(d) + if new_trust_radius >= MAX_TRUST_REDUCTION * trust_radius: + trust_radius *= MAX_TRUST_REDUCTION + elif new_trust_radius >= MIN_TRUST_REDUCTION * trust_radius: + trust_radius = new_trust_radius + else: + trust_radius *= MIN_TRUST_REDUCTION + + # Update iteration + if reduction_ratio >= SUFFICIENT_REDUCTION_RATIO: + x = x_next + f, b = f_next, b_next + c, A = grad_and_jac(x) + S = scaling(x) + # Get projections + Z, LS, Y = projections(A, factorization_method) + # Compute least-square lagrange multipliers + v = -LS.dot(c) + # Compute Hessian + H = lagr_hess(x, v) + # Set Flag + last_iteration_failed = False + # Otimality values + optimality = norm(c + A.T.dot(v), np.inf) + constr_violation = norm(b, np.inf) if len(b) > 0 else 0 + else: + penalty = previous_penalty + last_iteration_failed = True + + return x, state diff --git a/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/minimize_trustregion_constr.py b/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/minimize_trustregion_constr.py new file mode 100644 index 0000000000000000000000000000000000000000..2835ea5445c0eafc303f0cb1ab8543f48b7e3bb9 --- /dev/null +++ b/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/minimize_trustregion_constr.py @@ -0,0 +1,564 @@ +import time +import numpy as np +from scipy.sparse.linalg import LinearOperator +from .._differentiable_functions import VectorFunction +from .._constraints import ( + NonlinearConstraint, LinearConstraint, PreparedConstraint, Bounds, strict_bounds) +from .._hessian_update_strategy import BFGS +from .._optimize import OptimizeResult +from .._differentiable_functions import ScalarFunction +from .equality_constrained_sqp import equality_constrained_sqp +from .canonical_constraint import (CanonicalConstraint, + initial_constraints_as_canonical) +from .tr_interior_point import tr_interior_point +from .report import BasicReport, SQPReport, IPReport + + +TERMINATION_MESSAGES = { + 0: "The maximum number of function evaluations is exceeded.", + 1: "`gtol` termination condition is satisfied.", + 2: "`xtol` termination condition is satisfied.", + 3: "`callback` function requested termination." +} + + +class HessianLinearOperator: + """Build LinearOperator from hessp""" + def __init__(self, hessp, n): + self.hessp = hessp + self.n = n + + def __call__(self, x, *args): + def matvec(p): + return self.hessp(x, p, *args) + + return LinearOperator((self.n, self.n), matvec=matvec) + + +class LagrangianHessian: + """The Hessian of the Lagrangian as LinearOperator. + + The Lagrangian is computed as the objective function plus all the + constraints multiplied with some numbers (Lagrange multipliers). + """ + def __init__(self, n, objective_hess, constraints_hess): + self.n = n + self.objective_hess = objective_hess + self.constraints_hess = constraints_hess + + def __call__(self, x, v_eq=np.empty(0), v_ineq=np.empty(0)): + H_objective = self.objective_hess(x) + H_constraints = self.constraints_hess(x, v_eq, v_ineq) + + def matvec(p): + return H_objective.dot(p) + H_constraints.dot(p) + + return LinearOperator((self.n, self.n), matvec) + + +def update_state_sqp(state, x, last_iteration_failed, objective, prepared_constraints, + start_time, tr_radius, constr_penalty, cg_info): + state.nit += 1 + state.nfev = objective.nfev + state.njev = objective.ngev + state.nhev = objective.nhev + state.constr_nfev = [c.fun.nfev if isinstance(c.fun, VectorFunction) else 0 + for c in prepared_constraints] + state.constr_njev = [c.fun.njev if isinstance(c.fun, VectorFunction) else 0 + for c in prepared_constraints] + state.constr_nhev = [c.fun.nhev if isinstance(c.fun, VectorFunction) else 0 + for c in prepared_constraints] + + if not last_iteration_failed: + state.x = x + state.fun = objective.f + state.grad = objective.g + state.v = [c.fun.v for c in prepared_constraints] + state.constr = [c.fun.f for c in prepared_constraints] + state.jac = [c.fun.J for c in prepared_constraints] + # Compute Lagrangian Gradient + state.lagrangian_grad = np.copy(state.grad) + for c in prepared_constraints: + state.lagrangian_grad += c.fun.J.T.dot(c.fun.v) + state.optimality = np.linalg.norm(state.lagrangian_grad, np.inf) + # Compute maximum constraint violation + state.constr_violation = 0 + for i in range(len(prepared_constraints)): + lb, ub = prepared_constraints[i].bounds + c = state.constr[i] + state.constr_violation = np.max([state.constr_violation, + np.max(lb - c), + np.max(c - ub)]) + + state.execution_time = time.time() - start_time + state.tr_radius = tr_radius + state.constr_penalty = constr_penalty + state.cg_niter += cg_info["niter"] + state.cg_stop_cond = cg_info["stop_cond"] + + return state + + +def update_state_ip(state, x, last_iteration_failed, objective, + prepared_constraints, start_time, + tr_radius, constr_penalty, cg_info, + barrier_parameter, barrier_tolerance): + state = update_state_sqp(state, x, last_iteration_failed, objective, + prepared_constraints, start_time, tr_radius, + constr_penalty, cg_info) + state.barrier_parameter = barrier_parameter + state.barrier_tolerance = barrier_tolerance + return state + + +def _minimize_trustregion_constr(fun, x0, args, grad, + hess, hessp, bounds, constraints, + xtol=1e-8, gtol=1e-8, + barrier_tol=1e-8, + sparse_jacobian=None, + callback=None, maxiter=1000, + verbose=0, finite_diff_rel_step=None, + initial_constr_penalty=1.0, initial_tr_radius=1.0, + initial_barrier_parameter=0.1, + initial_barrier_tolerance=0.1, + factorization_method=None, + disp=False): + """Minimize a scalar function subject to constraints. + + Parameters + ---------- + gtol : float, optional + Tolerance for termination by the norm of the Lagrangian gradient. + The algorithm will terminate when both the infinity norm (i.e., max + abs value) of the Lagrangian gradient and the constraint violation + are smaller than ``gtol``. Default is 1e-8. + xtol : float, optional + Tolerance for termination by the change of the independent variable. + The algorithm will terminate when ``tr_radius < xtol``, where + ``tr_radius`` is the radius of the trust region used in the algorithm. + Default is 1e-8. + barrier_tol : float, optional + Threshold on the barrier parameter for the algorithm termination. + When inequality constraints are present, the algorithm will terminate + only when the barrier parameter is less than `barrier_tol`. + Default is 1e-8. + sparse_jacobian : {bool, None}, optional + Determines how to represent Jacobians of the constraints. If bool, + then Jacobians of all the constraints will be converted to the + corresponding format. If None (default), then Jacobians won't be + converted, but the algorithm can proceed only if they all have the + same format. + initial_tr_radius: float, optional + Initial trust radius. The trust radius gives the maximum distance + between solution points in consecutive iterations. It reflects the + trust the algorithm puts in the local approximation of the optimization + problem. For an accurate local approximation the trust-region should be + large and for an approximation valid only close to the current point it + should be a small one. The trust radius is automatically updated throughout + the optimization process, with ``initial_tr_radius`` being its initial value. + Default is 1 (recommended in [1]_, p. 19). + initial_constr_penalty : float, optional + Initial constraints penalty parameter. The penalty parameter is used for + balancing the requirements of decreasing the objective function + and satisfying the constraints. It is used for defining the merit function: + ``merit_function(x) = fun(x) + constr_penalty * constr_norm_l2(x)``, + where ``constr_norm_l2(x)`` is the l2 norm of a vector containing all + the constraints. The merit function is used for accepting or rejecting + trial points and ``constr_penalty`` weights the two conflicting goals + of reducing objective function and constraints. The penalty is automatically + updated throughout the optimization process, with + ``initial_constr_penalty`` being its initial value. Default is 1 + (recommended in [1]_, p 19). + initial_barrier_parameter, initial_barrier_tolerance: float, optional + Initial barrier parameter and initial tolerance for the barrier subproblem. + Both are used only when inequality constraints are present. For dealing with + optimization problems ``min_x f(x)`` subject to inequality constraints + ``c(x) <= 0`` the algorithm introduces slack variables, solving the problem + ``min_(x,s) f(x) + barrier_parameter*sum(ln(s))`` subject to the equality + constraints ``c(x) + s = 0`` instead of the original problem. This subproblem + is solved for decreasing values of ``barrier_parameter`` and with decreasing + tolerances for the termination, starting with ``initial_barrier_parameter`` + for the barrier parameter and ``initial_barrier_tolerance`` for the + barrier tolerance. Default is 0.1 for both values (recommended in [1]_ p. 19). + Also note that ``barrier_parameter`` and ``barrier_tolerance`` are updated + with the same prefactor. + factorization_method : string or None, optional + Method to factorize the Jacobian of the constraints. Use None (default) + for the auto selection or one of: + + - 'NormalEquation' (requires scikit-sparse) + - 'AugmentedSystem' + - 'QRFactorization' + - 'SVDFactorization' + + The methods 'NormalEquation' and 'AugmentedSystem' can be used only + with sparse constraints. The projections required by the algorithm + will be computed using, respectively, the normal equation and the + augmented system approaches explained in [1]_. 'NormalEquation' + computes the Cholesky factorization of ``A A.T`` and 'AugmentedSystem' + performs the LU factorization of an augmented system. They usually + provide similar results. 'AugmentedSystem' is used by default for + sparse matrices. + + The methods 'QRFactorization' and 'SVDFactorization' can be used + only with dense constraints. They compute the required projections + using, respectively, QR and SVD factorizations. The 'SVDFactorization' + method can cope with Jacobian matrices with deficient row rank and will + be used whenever other factorization methods fail (which may imply the + conversion of sparse matrices to a dense format when required). + By default, 'QRFactorization' is used for dense matrices. + finite_diff_rel_step : None or array_like, optional + Relative step size for the finite difference approximation. + maxiter : int, optional + Maximum number of algorithm iterations. Default is 1000. + verbose : {0, 1, 2}, optional + Level of algorithm's verbosity: + + * 0 (default) : work silently. + * 1 : display a termination report. + * 2 : display progress during iterations. + * 3 : display progress during iterations (more complete report). + + disp : bool, optional + If True (default), then `verbose` will be set to 1 if it was 0. + + Returns + ------- + `OptimizeResult` with the fields documented below. Note the following: + + 1. All values corresponding to the constraints are ordered as they + were passed to the solver. And values corresponding to `bounds` + constraints are put *after* other constraints. + 2. All numbers of function, Jacobian or Hessian evaluations correspond + to numbers of actual Python function calls. It means, for example, + that if a Jacobian is estimated by finite differences, then the + number of Jacobian evaluations will be zero and the number of + function evaluations will be incremented by all calls during the + finite difference estimation. + + x : ndarray, shape (n,) + Solution found. + optimality : float + Infinity norm of the Lagrangian gradient at the solution. + constr_violation : float + Maximum constraint violation at the solution. + fun : float + Objective function at the solution. + grad : ndarray, shape (n,) + Gradient of the objective function at the solution. + lagrangian_grad : ndarray, shape (n,) + Gradient of the Lagrangian function at the solution. + nit : int + Total number of iterations. + nfev : integer + Number of the objective function evaluations. + njev : integer + Number of the objective function gradient evaluations. + nhev : integer + Number of the objective function Hessian evaluations. + cg_niter : int + Total number of the conjugate gradient method iterations. + method : {'equality_constrained_sqp', 'tr_interior_point'} + Optimization method used. + constr : list of ndarray + List of constraint values at the solution. + jac : list of {ndarray, sparse matrix} + List of the Jacobian matrices of the constraints at the solution. + v : list of ndarray + List of the Lagrange multipliers for the constraints at the solution. + For an inequality constraint a positive multiplier means that the upper + bound is active, a negative multiplier means that the lower bound is + active and if a multiplier is zero it means the constraint is not + active. + constr_nfev : list of int + Number of constraint evaluations for each of the constraints. + constr_njev : list of int + Number of Jacobian matrix evaluations for each of the constraints. + constr_nhev : list of int + Number of Hessian evaluations for each of the constraints. + tr_radius : float + Radius of the trust region at the last iteration. + constr_penalty : float + Penalty parameter at the last iteration, see `initial_constr_penalty`. + barrier_tolerance : float + Tolerance for the barrier subproblem at the last iteration. + Only for problems with inequality constraints. + barrier_parameter : float + Barrier parameter at the last iteration. Only for problems + with inequality constraints. + execution_time : float + Total execution time. + message : str + Termination message. + status : {0, 1, 2, 3} + Termination status: + + * 0 : The maximum number of function evaluations is exceeded. + * 1 : `gtol` termination condition is satisfied. + * 2 : `xtol` termination condition is satisfied. + * 3 : `callback` function requested termination. + + cg_stop_cond : int + Reason for CG subproblem termination at the last iteration: + + * 0 : CG subproblem not evaluated. + * 1 : Iteration limit was reached. + * 2 : Reached the trust-region boundary. + * 3 : Negative curvature detected. + * 4 : Tolerance was satisfied. + + References + ---------- + .. [1] Conn, A. R., Gould, N. I., & Toint, P. L. + Trust region methods. 2000. Siam. pp. 19. + """ + x0 = np.atleast_1d(x0).astype(float) + n_vars = np.size(x0) + if hess is None: + if callable(hessp): + hess = HessianLinearOperator(hessp, n_vars) + else: + hess = BFGS() + if disp and verbose == 0: + verbose = 1 + + if bounds is not None: + modified_lb = np.nextafter(bounds.lb, -np.inf, where=bounds.lb > -np.inf) + modified_ub = np.nextafter(bounds.ub, np.inf, where=bounds.ub < np.inf) + modified_lb = np.where(np.isfinite(bounds.lb), modified_lb, bounds.lb) + modified_ub = np.where(np.isfinite(bounds.ub), modified_ub, bounds.ub) + bounds = Bounds(modified_lb, modified_ub, keep_feasible=bounds.keep_feasible) + finite_diff_bounds = strict_bounds(bounds.lb, bounds.ub, + bounds.keep_feasible, n_vars) + else: + finite_diff_bounds = (-np.inf, np.inf) + + # Define Objective Function + objective = ScalarFunction(fun, x0, args, grad, hess, + finite_diff_rel_step, finite_diff_bounds) + + # Put constraints in list format when needed. + if isinstance(constraints, (NonlinearConstraint, LinearConstraint)): + constraints = [constraints] + + # Prepare constraints. + prepared_constraints = [ + PreparedConstraint(c, x0, sparse_jacobian, finite_diff_bounds) + for c in constraints] + + # Check that all constraints are either sparse or dense. + n_sparse = sum(c.fun.sparse_jacobian for c in prepared_constraints) + if 0 < n_sparse < len(prepared_constraints): + raise ValueError("All constraints must have the same kind of the " + "Jacobian --- either all sparse or all dense. " + "You can set the sparsity globally by setting " + "`sparse_jacobian` to either True of False.") + if prepared_constraints: + sparse_jacobian = n_sparse > 0 + + if bounds is not None: + if sparse_jacobian is None: + sparse_jacobian = True + prepared_constraints.append(PreparedConstraint(bounds, x0, + sparse_jacobian)) + + # Concatenate initial constraints to the canonical form. + c_eq0, c_ineq0, J_eq0, J_ineq0 = initial_constraints_as_canonical( + n_vars, prepared_constraints, sparse_jacobian) + + # Prepare all canonical constraints and concatenate it into one. + canonical_all = [CanonicalConstraint.from_PreparedConstraint(c) + for c in prepared_constraints] + + if len(canonical_all) == 0: + canonical = CanonicalConstraint.empty(n_vars) + elif len(canonical_all) == 1: + canonical = canonical_all[0] + else: + canonical = CanonicalConstraint.concatenate(canonical_all, + sparse_jacobian) + + # Generate the Hessian of the Lagrangian. + lagrangian_hess = LagrangianHessian(n_vars, objective.hess, canonical.hess) + + # Choose appropriate method + if canonical.n_ineq == 0: + method = 'equality_constrained_sqp' + else: + method = 'tr_interior_point' + + # Construct OptimizeResult + state = OptimizeResult( + nit=0, nfev=0, njev=0, nhev=0, + cg_niter=0, cg_stop_cond=0, + fun=objective.f, grad=objective.g, + lagrangian_grad=np.copy(objective.g), + constr=[c.fun.f for c in prepared_constraints], + jac=[c.fun.J for c in prepared_constraints], + constr_nfev=[0 for c in prepared_constraints], + constr_njev=[0 for c in prepared_constraints], + constr_nhev=[0 for c in prepared_constraints], + v=[c.fun.v for c in prepared_constraints], + method=method) + + # Start counting + start_time = time.time() + + # Define stop criteria + if method == 'equality_constrained_sqp': + def stop_criteria(state, x, last_iteration_failed, + optimality, constr_violation, + tr_radius, constr_penalty, cg_info): + state = update_state_sqp(state, x, last_iteration_failed, + objective, prepared_constraints, + start_time, tr_radius, constr_penalty, + cg_info) + if verbose == 2: + BasicReport.print_iteration(state.nit, + state.nfev, + state.cg_niter, + state.fun, + state.tr_radius, + state.optimality, + state.constr_violation) + elif verbose > 2: + SQPReport.print_iteration(state.nit, + state.nfev, + state.cg_niter, + state.fun, + state.tr_radius, + state.optimality, + state.constr_violation, + state.constr_penalty, + state.cg_stop_cond) + state.status = None + state.niter = state.nit # Alias for callback (backward-compatibility) + if callback is not None: + callback_stop = False + try: + callback_stop = callback(state) + except StopIteration: + callback_stop = True + if callback_stop: + state.status = 3 + return True + if state.optimality < gtol and state.constr_violation < gtol: + state.status = 1 + elif state.tr_radius < xtol: + state.status = 2 + elif state.nit >= maxiter: + state.status = 0 + return state.status in (0, 1, 2, 3) + elif method == 'tr_interior_point': + def stop_criteria(state, x, last_iteration_failed, tr_radius, + constr_penalty, cg_info, barrier_parameter, + barrier_tolerance): + state = update_state_ip(state, x, last_iteration_failed, + objective, prepared_constraints, + start_time, tr_radius, constr_penalty, + cg_info, barrier_parameter, barrier_tolerance) + if verbose == 2: + BasicReport.print_iteration(state.nit, + state.nfev, + state.cg_niter, + state.fun, + state.tr_radius, + state.optimality, + state.constr_violation) + elif verbose > 2: + IPReport.print_iteration(state.nit, + state.nfev, + state.cg_niter, + state.fun, + state.tr_radius, + state.optimality, + state.constr_violation, + state.constr_penalty, + state.barrier_parameter, + state.cg_stop_cond) + state.status = None + state.niter = state.nit # Alias for callback (backward compatibility) + if callback is not None: + callback_stop = False + try: + callback_stop = callback(state) + except StopIteration: + callback_stop = True + if callback_stop: + state.status = 3 + return True + if state.optimality < gtol and state.constr_violation < gtol: + state.status = 1 + elif (state.tr_radius < xtol + and state.barrier_parameter < barrier_tol): + state.status = 2 + elif state.nit >= maxiter: + state.status = 0 + return state.status in (0, 1, 2, 3) + + if verbose == 2: + BasicReport.print_header() + elif verbose > 2: + if method == 'equality_constrained_sqp': + SQPReport.print_header() + elif method == 'tr_interior_point': + IPReport.print_header() + + # Call inferior function to do the optimization + if method == 'equality_constrained_sqp': + def fun_and_constr(x): + f = objective.fun(x) + c_eq, _ = canonical.fun(x) + return f, c_eq + + def grad_and_jac(x): + g = objective.grad(x) + J_eq, _ = canonical.jac(x) + return g, J_eq + + _, result = equality_constrained_sqp( + fun_and_constr, grad_and_jac, lagrangian_hess, + x0, objective.f, objective.g, + c_eq0, J_eq0, + stop_criteria, state, + initial_constr_penalty, initial_tr_radius, + factorization_method) + + elif method == 'tr_interior_point': + _, result = tr_interior_point( + objective.fun, objective.grad, lagrangian_hess, + n_vars, canonical.n_ineq, canonical.n_eq, + canonical.fun, canonical.jac, + x0, objective.f, objective.g, + c_ineq0, J_ineq0, c_eq0, J_eq0, + stop_criteria, + canonical.keep_feasible, + xtol, state, initial_barrier_parameter, + initial_barrier_tolerance, + initial_constr_penalty, initial_tr_radius, + factorization_method) + + # Status 3 occurs when the callback function requests termination, + # this is assumed to not be a success. + result.success = True if result.status in (1, 2) else False + result.message = TERMINATION_MESSAGES[result.status] + + # Alias (for backward compatibility with 1.1.0) + result.niter = result.nit + + if verbose == 2: + BasicReport.print_footer() + elif verbose > 2: + if method == 'equality_constrained_sqp': + SQPReport.print_footer() + elif method == 'tr_interior_point': + IPReport.print_footer() + if verbose >= 1: + print(result.message) + print("Number of iterations: {}, function evaluations: {}, " + "CG iterations: {}, optimality: {:.2e}, " + "constraint violation: {:.2e}, execution time: {:4.2} s." + .format(result.nit, result.nfev, result.cg_niter, + result.optimality, result.constr_violation, + result.execution_time)) + return result diff --git a/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/projections.py b/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/projections.py new file mode 100644 index 0000000000000000000000000000000000000000..a07b836bdbad688a265ae34ce91a361fd5050eb1 --- /dev/null +++ b/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/projections.py @@ -0,0 +1,407 @@ +"""Basic linear factorizations needed by the solver.""" + +from scipy.sparse import (bmat, csc_matrix, eye, issparse) +from scipy.sparse.linalg import LinearOperator +import scipy.linalg +import scipy.sparse.linalg +try: + from sksparse.cholmod import cholesky_AAt + sksparse_available = True +except ImportError: + import warnings + sksparse_available = False +import numpy as np +from warnings import warn + +__all__ = [ + 'orthogonality', + 'projections', +] + + +def orthogonality(A, g): + """Measure orthogonality between a vector and the null space of a matrix. + + Compute a measure of orthogonality between the null space + of the (possibly sparse) matrix ``A`` and a given vector ``g``. + + The formula is a simplified (and cheaper) version of formula (3.13) + from [1]_. + ``orth = norm(A g, ord=2)/(norm(A, ord='fro')*norm(g, ord=2))``. + + References + ---------- + .. [1] Gould, Nicholas IM, Mary E. Hribar, and Jorge Nocedal. + "On the solution of equality constrained quadratic + programming problems arising in optimization." + SIAM Journal on Scientific Computing 23.4 (2001): 1376-1395. + """ + # Compute vector norms + norm_g = np.linalg.norm(g) + # Compute Froebnius norm of the matrix A + if issparse(A): + norm_A = scipy.sparse.linalg.norm(A, ord='fro') + else: + norm_A = np.linalg.norm(A, ord='fro') + + # Check if norms are zero + if norm_g == 0 or norm_A == 0: + return 0 + + norm_A_g = np.linalg.norm(A.dot(g)) + # Orthogonality measure + orth = norm_A_g / (norm_A*norm_g) + return orth + + +def normal_equation_projections(A, m, n, orth_tol, max_refin, tol): + """Return linear operators for matrix A using ``NormalEquation`` approach. + """ + # Cholesky factorization + factor = cholesky_AAt(A) + + # z = x - A.T inv(A A.T) A x + def null_space(x): + v = factor(A.dot(x)) + z = x - A.T.dot(v) + + # Iterative refinement to improve roundoff + # errors described in [2]_, algorithm 5.1. + k = 0 + while orthogonality(A, z) > orth_tol: + if k >= max_refin: + break + # z_next = z - A.T inv(A A.T) A z + v = factor(A.dot(z)) + z = z - A.T.dot(v) + k += 1 + + return z + + # z = inv(A A.T) A x + def least_squares(x): + return factor(A.dot(x)) + + # z = A.T inv(A A.T) x + def row_space(x): + return A.T.dot(factor(x)) + + return null_space, least_squares, row_space + + +def augmented_system_projections(A, m, n, orth_tol, max_refin, tol): + """Return linear operators for matrix A - ``AugmentedSystem``.""" + # Form augmented system + K = csc_matrix(bmat([[eye(n), A.T], [A, None]])) + # LU factorization + # TODO: Use a symmetric indefinite factorization + # to solve the system twice as fast (because + # of the symmetry). + try: + solve = scipy.sparse.linalg.factorized(K) + except RuntimeError: + warn("Singular Jacobian matrix. Using dense SVD decomposition to " + "perform the factorizations.", + stacklevel=3) + return svd_factorization_projections(A.toarray(), + m, n, orth_tol, + max_refin, tol) + + # z = x - A.T inv(A A.T) A x + # is computed solving the extended system: + # [I A.T] * [ z ] = [x] + # [A O ] [aux] [0] + def null_space(x): + # v = [x] + # [0] + v = np.hstack([x, np.zeros(m)]) + # lu_sol = [ z ] + # [aux] + lu_sol = solve(v) + z = lu_sol[:n] + + # Iterative refinement to improve roundoff + # errors described in [2]_, algorithm 5.2. + k = 0 + while orthogonality(A, z) > orth_tol: + if k >= max_refin: + break + # new_v = [x] - [I A.T] * [ z ] + # [0] [A O ] [aux] + new_v = v - K.dot(lu_sol) + # [I A.T] * [delta z ] = new_v + # [A O ] [delta aux] + lu_update = solve(new_v) + # [ z ] += [delta z ] + # [aux] [delta aux] + lu_sol += lu_update + z = lu_sol[:n] + k += 1 + + # return z = x - A.T inv(A A.T) A x + return z + + # z = inv(A A.T) A x + # is computed solving the extended system: + # [I A.T] * [aux] = [x] + # [A O ] [ z ] [0] + def least_squares(x): + # v = [x] + # [0] + v = np.hstack([x, np.zeros(m)]) + # lu_sol = [aux] + # [ z ] + lu_sol = solve(v) + # return z = inv(A A.T) A x + return lu_sol[n:m+n] + + # z = A.T inv(A A.T) x + # is computed solving the extended system: + # [I A.T] * [ z ] = [0] + # [A O ] [aux] [x] + def row_space(x): + # v = [0] + # [x] + v = np.hstack([np.zeros(n), x]) + # lu_sol = [ z ] + # [aux] + lu_sol = solve(v) + # return z = A.T inv(A A.T) x + return lu_sol[:n] + + return null_space, least_squares, row_space + + +def qr_factorization_projections(A, m, n, orth_tol, max_refin, tol): + """Return linear operators for matrix A using ``QRFactorization`` approach. + """ + # QRFactorization + Q, R, P = scipy.linalg.qr(A.T, pivoting=True, mode='economic') + + if np.linalg.norm(R[-1, :], np.inf) < tol: + warn('Singular Jacobian matrix. Using SVD decomposition to ' + + 'perform the factorizations.', + stacklevel=3) + return svd_factorization_projections(A, m, n, + orth_tol, + max_refin, + tol) + + # z = x - A.T inv(A A.T) A x + def null_space(x): + # v = P inv(R) Q.T x + aux1 = Q.T.dot(x) + aux2 = scipy.linalg.solve_triangular(R, aux1, lower=False) + v = np.zeros(m) + v[P] = aux2 + z = x - A.T.dot(v) + + # Iterative refinement to improve roundoff + # errors described in [2]_, algorithm 5.1. + k = 0 + while orthogonality(A, z) > orth_tol: + if k >= max_refin: + break + # v = P inv(R) Q.T x + aux1 = Q.T.dot(z) + aux2 = scipy.linalg.solve_triangular(R, aux1, lower=False) + v[P] = aux2 + # z_next = z - A.T v + z = z - A.T.dot(v) + k += 1 + + return z + + # z = inv(A A.T) A x + def least_squares(x): + # z = P inv(R) Q.T x + aux1 = Q.T.dot(x) + aux2 = scipy.linalg.solve_triangular(R, aux1, lower=False) + z = np.zeros(m) + z[P] = aux2 + return z + + # z = A.T inv(A A.T) x + def row_space(x): + # z = Q inv(R.T) P.T x + aux1 = x[P] + aux2 = scipy.linalg.solve_triangular(R, aux1, + lower=False, + trans='T') + z = Q.dot(aux2) + return z + + return null_space, least_squares, row_space + + +def svd_factorization_projections(A, m, n, orth_tol, max_refin, tol): + """Return linear operators for matrix A using ``SVDFactorization`` approach. + """ + # SVD Factorization + U, s, Vt = scipy.linalg.svd(A, full_matrices=False) + + # Remove dimensions related with very small singular values + U = U[:, s > tol] + Vt = Vt[s > tol, :] + s = s[s > tol] + + # z = x - A.T inv(A A.T) A x + def null_space(x): + # v = U 1/s V.T x = inv(A A.T) A x + aux1 = Vt.dot(x) + aux2 = 1/s*aux1 + v = U.dot(aux2) + z = x - A.T.dot(v) + + # Iterative refinement to improve roundoff + # errors described in [2]_, algorithm 5.1. + k = 0 + while orthogonality(A, z) > orth_tol: + if k >= max_refin: + break + # v = U 1/s V.T x = inv(A A.T) A x + aux1 = Vt.dot(z) + aux2 = 1/s*aux1 + v = U.dot(aux2) + # z_next = z - A.T v + z = z - A.T.dot(v) + k += 1 + + return z + + # z = inv(A A.T) A x + def least_squares(x): + # z = U 1/s V.T x = inv(A A.T) A x + aux1 = Vt.dot(x) + aux2 = 1/s*aux1 + z = U.dot(aux2) + return z + + # z = A.T inv(A A.T) x + def row_space(x): + # z = V 1/s U.T x + aux1 = U.T.dot(x) + aux2 = 1/s*aux1 + z = Vt.T.dot(aux2) + return z + + return null_space, least_squares, row_space + + +def projections(A, method=None, orth_tol=1e-12, max_refin=3, tol=1e-15): + """Return three linear operators related with a given matrix A. + + Parameters + ---------- + A : sparse matrix (or ndarray), shape (m, n) + Matrix ``A`` used in the projection. + method : string, optional + Method used for compute the given linear + operators. Should be one of: + + - 'NormalEquation': The operators + will be computed using the + so-called normal equation approach + explained in [1]_. In order to do + so the Cholesky factorization of + ``(A A.T)`` is computed. Exclusive + for sparse matrices. + - 'AugmentedSystem': The operators + will be computed using the + so-called augmented system approach + explained in [1]_. Exclusive + for sparse matrices. + - 'QRFactorization': Compute projections + using QR factorization. Exclusive for + dense matrices. + - 'SVDFactorization': Compute projections + using SVD factorization. Exclusive for + dense matrices. + + orth_tol : float, optional + Tolerance for iterative refinements. + max_refin : int, optional + Maximum number of iterative refinements. + tol : float, optional + Tolerance for singular values. + + Returns + ------- + Z : LinearOperator, shape (n, n) + Null-space operator. For a given vector ``x``, + the null space operator is equivalent to apply + a projection matrix ``P = I - A.T inv(A A.T) A`` + to the vector. It can be shown that this is + equivalent to project ``x`` into the null space + of A. + LS : LinearOperator, shape (m, n) + Least-squares operator. For a given vector ``x``, + the least-squares operator is equivalent to apply a + pseudoinverse matrix ``pinv(A.T) = inv(A A.T) A`` + to the vector. It can be shown that this vector + ``pinv(A.T) x`` is the least_square solution to + ``A.T y = x``. + Y : LinearOperator, shape (n, m) + Row-space operator. For a given vector ``x``, + the row-space operator is equivalent to apply a + projection matrix ``Q = A.T inv(A A.T)`` + to the vector. It can be shown that this + vector ``y = Q x`` the minimum norm solution + of ``A y = x``. + + Notes + ----- + Uses iterative refinements described in [1] + during the computation of ``Z`` in order to + cope with the possibility of large roundoff errors. + + References + ---------- + .. [1] Gould, Nicholas IM, Mary E. Hribar, and Jorge Nocedal. + "On the solution of equality constrained quadratic + programming problems arising in optimization." + SIAM Journal on Scientific Computing 23.4 (2001): 1376-1395. + """ + m, n = np.shape(A) + + # The factorization of an empty matrix + # only works for the sparse representation. + if m*n == 0: + A = csc_matrix(A) + + # Check Argument + if issparse(A): + if method is None: + method = "AugmentedSystem" + if method not in ("NormalEquation", "AugmentedSystem"): + raise ValueError("Method not allowed for sparse matrix.") + if method == "NormalEquation" and not sksparse_available: + warnings.warn("Only accepts 'NormalEquation' option when " + "scikit-sparse is available. Using " + "'AugmentedSystem' option instead.", + ImportWarning, stacklevel=3) + method = 'AugmentedSystem' + else: + if method is None: + method = "QRFactorization" + if method not in ("QRFactorization", "SVDFactorization"): + raise ValueError("Method not allowed for dense array.") + + if method == 'NormalEquation': + null_space, least_squares, row_space \ + = normal_equation_projections(A, m, n, orth_tol, max_refin, tol) + elif method == 'AugmentedSystem': + null_space, least_squares, row_space \ + = augmented_system_projections(A, m, n, orth_tol, max_refin, tol) + elif method == "QRFactorization": + null_space, least_squares, row_space \ + = qr_factorization_projections(A, m, n, orth_tol, max_refin, tol) + elif method == "SVDFactorization": + null_space, least_squares, row_space \ + = svd_factorization_projections(A, m, n, orth_tol, max_refin, tol) + + Z = LinearOperator((n, n), null_space) + LS = LinearOperator((m, n), least_squares) + Y = LinearOperator((n, m), row_space) + + return Z, LS, Y diff --git a/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/qp_subproblem.py b/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/qp_subproblem.py new file mode 100644 index 0000000000000000000000000000000000000000..a039a7738c283f90f30fd7c4583bf9e1a8f559d5 --- /dev/null +++ b/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/qp_subproblem.py @@ -0,0 +1,637 @@ +"""Equality-constrained quadratic programming solvers.""" + +from scipy.sparse import (linalg, bmat, csc_matrix) +from math import copysign +import numpy as np +from numpy.linalg import norm + +__all__ = [ + 'eqp_kktfact', + 'sphere_intersections', + 'box_intersections', + 'box_sphere_intersections', + 'inside_box_boundaries', + 'modified_dogleg', + 'projected_cg' +] + + +# For comparison with the projected CG +def eqp_kktfact(H, c, A, b): + """Solve equality-constrained quadratic programming (EQP) problem. + + Solve ``min 1/2 x.T H x + x.t c`` subject to ``A x + b = 0`` + using direct factorization of the KKT system. + + Parameters + ---------- + H : sparse matrix, shape (n, n) + Hessian matrix of the EQP problem. + c : array_like, shape (n,) + Gradient of the quadratic objective function. + A : sparse matrix + Jacobian matrix of the EQP problem. + b : array_like, shape (m,) + Right-hand side of the constraint equation. + + Returns + ------- + x : array_like, shape (n,) + Solution of the KKT problem. + lagrange_multipliers : ndarray, shape (m,) + Lagrange multipliers of the KKT problem. + """ + n, = np.shape(c) # Number of parameters + m, = np.shape(b) # Number of constraints + + # Karush-Kuhn-Tucker matrix of coefficients. + # Defined as in Nocedal/Wright "Numerical + # Optimization" p.452 in Eq. (16.4). + kkt_matrix = csc_matrix(bmat([[H, A.T], [A, None]])) + # Vector of coefficients. + kkt_vec = np.hstack([-c, -b]) + + # TODO: Use a symmetric indefinite factorization + # to solve the system twice as fast (because + # of the symmetry). + lu = linalg.splu(kkt_matrix) + kkt_sol = lu.solve(kkt_vec) + x = kkt_sol[:n] + lagrange_multipliers = -kkt_sol[n:n+m] + + return x, lagrange_multipliers + + +def sphere_intersections(z, d, trust_radius, + entire_line=False): + """Find the intersection between segment (or line) and spherical constraints. + + Find the intersection between the segment (or line) defined by the + parametric equation ``x(t) = z + t*d`` and the ball + ``||x|| <= trust_radius``. + + Parameters + ---------- + z : array_like, shape (n,) + Initial point. + d : array_like, shape (n,) + Direction. + trust_radius : float + Ball radius. + entire_line : bool, optional + When ``True``, the function returns the intersection between the line + ``x(t) = z + t*d`` (``t`` can assume any value) and the ball + ``||x|| <= trust_radius``. When ``False``, the function returns the intersection + between the segment ``x(t) = z + t*d``, ``0 <= t <= 1``, and the ball. + + Returns + ------- + ta, tb : float + The line/segment ``x(t) = z + t*d`` is inside the ball for + for ``ta <= t <= tb``. + intersect : bool + When ``True``, there is a intersection between the line/segment + and the sphere. On the other hand, when ``False``, there is no + intersection. + """ + # Special case when d=0 + if norm(d) == 0: + return 0, 0, False + # Check for inf trust_radius + if np.isinf(trust_radius): + if entire_line: + ta = -np.inf + tb = np.inf + else: + ta = 0 + tb = 1 + intersect = True + return ta, tb, intersect + + a = np.dot(d, d) + b = 2 * np.dot(z, d) + c = np.dot(z, z) - trust_radius**2 + discriminant = b*b - 4*a*c + if discriminant < 0: + intersect = False + return 0, 0, intersect + sqrt_discriminant = np.sqrt(discriminant) + + # The following calculation is mathematically + # equivalent to: + # ta = (-b - sqrt_discriminant) / (2*a) + # tb = (-b + sqrt_discriminant) / (2*a) + # but produce smaller round off errors. + # Look at Matrix Computation p.97 + # for a better justification. + aux = b + copysign(sqrt_discriminant, b) + ta = -aux / (2*a) + tb = -2*c / aux + ta, tb = sorted([ta, tb]) + + if entire_line: + intersect = True + else: + # Checks to see if intersection happens + # within vectors length. + if tb < 0 or ta > 1: + intersect = False + ta = 0 + tb = 0 + else: + intersect = True + # Restrict intersection interval + # between 0 and 1. + ta = max(0, ta) + tb = min(1, tb) + + return ta, tb, intersect + + +def box_intersections(z, d, lb, ub, + entire_line=False): + """Find the intersection between segment (or line) and box constraints. + + Find the intersection between the segment (or line) defined by the + parametric equation ``x(t) = z + t*d`` and the rectangular box + ``lb <= x <= ub``. + + Parameters + ---------- + z : array_like, shape (n,) + Initial point. + d : array_like, shape (n,) + Direction. + lb : array_like, shape (n,) + Lower bounds to each one of the components of ``x``. Used + to delimit the rectangular box. + ub : array_like, shape (n, ) + Upper bounds to each one of the components of ``x``. Used + to delimit the rectangular box. + entire_line : bool, optional + When ``True``, the function returns the intersection between the line + ``x(t) = z + t*d`` (``t`` can assume any value) and the rectangular + box. When ``False``, the function returns the intersection between the segment + ``x(t) = z + t*d``, ``0 <= t <= 1``, and the rectangular box. + + Returns + ------- + ta, tb : float + The line/segment ``x(t) = z + t*d`` is inside the box for + for ``ta <= t <= tb``. + intersect : bool + When ``True``, there is a intersection between the line (or segment) + and the rectangular box. On the other hand, when ``False``, there is no + intersection. + """ + # Make sure it is a numpy array + z = np.asarray(z) + d = np.asarray(d) + lb = np.asarray(lb) + ub = np.asarray(ub) + # Special case when d=0 + if norm(d) == 0: + return 0, 0, False + + # Get values for which d==0 + zero_d = (d == 0) + # If the boundaries are not satisfied for some coordinate + # for which "d" is zero, there is no box-line intersection. + if (z[zero_d] < lb[zero_d]).any() or (z[zero_d] > ub[zero_d]).any(): + intersect = False + return 0, 0, intersect + # Remove values for which d is zero + not_zero_d = np.logical_not(zero_d) + z = z[not_zero_d] + d = d[not_zero_d] + lb = lb[not_zero_d] + ub = ub[not_zero_d] + + # Find a series of intervals (t_lb[i], t_ub[i]). + t_lb = (lb-z) / d + t_ub = (ub-z) / d + # Get the intersection of all those intervals. + ta = max(np.minimum(t_lb, t_ub)) + tb = min(np.maximum(t_lb, t_ub)) + + # Check if intersection is feasible + if ta <= tb: + intersect = True + else: + intersect = False + # Checks to see if intersection happens within vectors length. + if not entire_line: + if tb < 0 or ta > 1: + intersect = False + ta = 0 + tb = 0 + else: + # Restrict intersection interval between 0 and 1. + ta = max(0, ta) + tb = min(1, tb) + + return ta, tb, intersect + + +def box_sphere_intersections(z, d, lb, ub, trust_radius, + entire_line=False, + extra_info=False): + """Find the intersection between segment (or line) and box/sphere constraints. + + Find the intersection between the segment (or line) defined by the + parametric equation ``x(t) = z + t*d``, the rectangular box + ``lb <= x <= ub`` and the ball ``||x|| <= trust_radius``. + + Parameters + ---------- + z : array_like, shape (n,) + Initial point. + d : array_like, shape (n,) + Direction. + lb : array_like, shape (n,) + Lower bounds to each one of the components of ``x``. Used + to delimit the rectangular box. + ub : array_like, shape (n, ) + Upper bounds to each one of the components of ``x``. Used + to delimit the rectangular box. + trust_radius : float + Ball radius. + entire_line : bool, optional + When ``True``, the function returns the intersection between the line + ``x(t) = z + t*d`` (``t`` can assume any value) and the constraints. + When ``False``, the function returns the intersection between the segment + ``x(t) = z + t*d``, ``0 <= t <= 1`` and the constraints. + extra_info : bool, optional + When ``True``, the function returns ``intersect_sphere`` and ``intersect_box``. + + Returns + ------- + ta, tb : float + The line/segment ``x(t) = z + t*d`` is inside the rectangular box and + inside the ball for ``ta <= t <= tb``. + intersect : bool + When ``True``, there is a intersection between the line (or segment) + and both constraints. On the other hand, when ``False``, there is no + intersection. + sphere_info : dict, optional + Dictionary ``{ta, tb, intersect}`` containing the interval ``[ta, tb]`` + for which the line intercepts the ball. And a boolean value indicating + whether the sphere is intersected by the line. + box_info : dict, optional + Dictionary ``{ta, tb, intersect}`` containing the interval ``[ta, tb]`` + for which the line intercepts the box. And a boolean value indicating + whether the box is intersected by the line. + """ + ta_b, tb_b, intersect_b = box_intersections(z, d, lb, ub, + entire_line) + ta_s, tb_s, intersect_s = sphere_intersections(z, d, + trust_radius, + entire_line) + ta = np.maximum(ta_b, ta_s) + tb = np.minimum(tb_b, tb_s) + if intersect_b and intersect_s and ta <= tb: + intersect = True + else: + intersect = False + + if extra_info: + sphere_info = {'ta': ta_s, 'tb': tb_s, 'intersect': intersect_s} + box_info = {'ta': ta_b, 'tb': tb_b, 'intersect': intersect_b} + return ta, tb, intersect, sphere_info, box_info + else: + return ta, tb, intersect + + +def inside_box_boundaries(x, lb, ub): + """Check if lb <= x <= ub.""" + return (lb <= x).all() and (x <= ub).all() + + +def reinforce_box_boundaries(x, lb, ub): + """Return clipped value of x""" + return np.minimum(np.maximum(x, lb), ub) + + +def modified_dogleg(A, Y, b, trust_radius, lb, ub): + """Approximately minimize ``1/2*|| A x + b ||^2`` inside trust-region. + + Approximately solve the problem of minimizing ``1/2*|| A x + b ||^2`` + subject to ``||x|| < Delta`` and ``lb <= x <= ub`` using a modification + of the classical dogleg approach. + + Parameters + ---------- + A : LinearOperator (or sparse matrix or ndarray), shape (m, n) + Matrix ``A`` in the minimization problem. It should have + dimension ``(m, n)`` such that ``m < n``. + Y : LinearOperator (or sparse matrix or ndarray), shape (n, m) + LinearOperator that apply the projection matrix + ``Q = A.T inv(A A.T)`` to the vector. The obtained vector + ``y = Q x`` being the minimum norm solution of ``A y = x``. + b : array_like, shape (m,) + Vector ``b``in the minimization problem. + trust_radius: float + Trust radius to be considered. Delimits a sphere boundary + to the problem. + lb : array_like, shape (n,) + Lower bounds to each one of the components of ``x``. + It is expected that ``lb <= 0``, otherwise the algorithm + may fail. If ``lb[i] = -Inf``, the lower + bound for the ith component is just ignored. + ub : array_like, shape (n, ) + Upper bounds to each one of the components of ``x``. + It is expected that ``ub >= 0``, otherwise the algorithm + may fail. If ``ub[i] = Inf``, the upper bound for the ith + component is just ignored. + + Returns + ------- + x : array_like, shape (n,) + Solution to the problem. + + Notes + ----- + Based on implementations described in pp. 885-886 from [1]_. + + References + ---------- + .. [1] Byrd, Richard H., Mary E. Hribar, and Jorge Nocedal. + "An interior point algorithm for large-scale nonlinear + programming." SIAM Journal on Optimization 9.4 (1999): 877-900. + """ + # Compute minimum norm minimizer of 1/2*|| A x + b ||^2. + newton_point = -Y.dot(b) + # Check for interior point + if inside_box_boundaries(newton_point, lb, ub) \ + and norm(newton_point) <= trust_radius: + x = newton_point + return x + + # Compute gradient vector ``g = A.T b`` + g = A.T.dot(b) + # Compute Cauchy point + # `cauchy_point = g.T g / (g.T A.T A g)``. + A_g = A.dot(g) + cauchy_point = -np.dot(g, g) / np.dot(A_g, A_g) * g + # Origin + origin_point = np.zeros_like(cauchy_point) + + # Check the segment between cauchy_point and newton_point + # for a possible solution. + z = cauchy_point + p = newton_point - cauchy_point + _, alpha, intersect = box_sphere_intersections(z, p, lb, ub, + trust_radius) + if intersect: + x1 = z + alpha*p + else: + # Check the segment between the origin and cauchy_point + # for a possible solution. + z = origin_point + p = cauchy_point + _, alpha, _ = box_sphere_intersections(z, p, lb, ub, + trust_radius) + x1 = z + alpha*p + + # Check the segment between origin and newton_point + # for a possible solution. + z = origin_point + p = newton_point + _, alpha, _ = box_sphere_intersections(z, p, lb, ub, + trust_radius) + x2 = z + alpha*p + + # Return the best solution among x1 and x2. + if norm(A.dot(x1) + b) < norm(A.dot(x2) + b): + return x1 + else: + return x2 + + +def projected_cg(H, c, Z, Y, b, trust_radius=np.inf, + lb=None, ub=None, tol=None, + max_iter=None, max_infeasible_iter=None, + return_all=False): + """Solve EQP problem with projected CG method. + + Solve equality-constrained quadratic programming problem + ``min 1/2 x.T H x + x.t c`` subject to ``A x + b = 0`` and, + possibly, to trust region constraints ``||x|| < trust_radius`` + and box constraints ``lb <= x <= ub``. + + Parameters + ---------- + H : LinearOperator (or sparse matrix or ndarray), shape (n, n) + Operator for computing ``H v``. + c : array_like, shape (n,) + Gradient of the quadratic objective function. + Z : LinearOperator (or sparse matrix or ndarray), shape (n, n) + Operator for projecting ``x`` into the null space of A. + Y : LinearOperator, sparse matrix, ndarray, shape (n, m) + Operator that, for a given a vector ``b``, compute smallest + norm solution of ``A x + b = 0``. + b : array_like, shape (m,) + Right-hand side of the constraint equation. + trust_radius : float, optional + Trust radius to be considered. By default, uses ``trust_radius=inf``, + which means no trust radius at all. + lb : array_like, shape (n,), optional + Lower bounds to each one of the components of ``x``. + If ``lb[i] = -Inf`` the lower bound for the i-th + component is just ignored (default). + ub : array_like, shape (n, ), optional + Upper bounds to each one of the components of ``x``. + If ``ub[i] = Inf`` the upper bound for the i-th + component is just ignored (default). + tol : float, optional + Tolerance used to interrupt the algorithm. + max_iter : int, optional + Maximum algorithm iterations. Where ``max_inter <= n-m``. + By default, uses ``max_iter = n-m``. + max_infeasible_iter : int, optional + Maximum infeasible (regarding box constraints) iterations the + algorithm is allowed to take. + By default, uses ``max_infeasible_iter = n-m``. + return_all : bool, optional + When ``true``, return the list of all vectors through the iterations. + + Returns + ------- + x : array_like, shape (n,) + Solution of the EQP problem. + info : Dict + Dictionary containing the following: + + - niter : Number of iterations. + - stop_cond : Reason for algorithm termination: + 1. Iteration limit was reached; + 2. Reached the trust-region boundary; + 3. Negative curvature detected; + 4. Tolerance was satisfied. + - allvecs : List containing all intermediary vectors (optional). + - hits_boundary : True if the proposed step is on the boundary + of the trust region. + + Notes + ----- + Implementation of Algorithm 6.2 on [1]_. + + In the absence of spherical and box constraints, for sufficient + iterations, the method returns a truly optimal result. + In the presence of those constraints, the value returned is only + a inexpensive approximation of the optimal value. + + References + ---------- + .. [1] Gould, Nicholas IM, Mary E. Hribar, and Jorge Nocedal. + "On the solution of equality constrained quadratic + programming problems arising in optimization." + SIAM Journal on Scientific Computing 23.4 (2001): 1376-1395. + """ + CLOSE_TO_ZERO = 1e-25 + + n, = np.shape(c) # Number of parameters + m, = np.shape(b) # Number of constraints + + # Initial Values + x = Y.dot(-b) + r = Z.dot(H.dot(x) + c) + g = Z.dot(r) + p = -g + + # Store ``x`` value + if return_all: + allvecs = [x] + # Values for the first iteration + H_p = H.dot(p) + rt_g = norm(g)**2 # g.T g = r.T Z g = r.T g (ref [1]_ p.1389) + + # If x > trust-region the problem does not have a solution. + tr_distance = trust_radius - norm(x) + if tr_distance < 0: + raise ValueError("Trust region problem does not have a solution.") + # If x == trust_radius, then x is the solution + # to the optimization problem, since x is the + # minimum norm solution to Ax=b. + elif tr_distance < CLOSE_TO_ZERO: + info = {'niter': 0, 'stop_cond': 2, 'hits_boundary': True} + if return_all: + allvecs.append(x) + info['allvecs'] = allvecs + return x, info + + # Set default tolerance + if tol is None: + tol = max(min(0.01 * np.sqrt(rt_g), 0.1 * rt_g), CLOSE_TO_ZERO) + # Set default lower and upper bounds + if lb is None: + lb = np.full(n, -np.inf) + if ub is None: + ub = np.full(n, np.inf) + # Set maximum iterations + if max_iter is None: + max_iter = n-m + max_iter = min(max_iter, n-m) + # Set maximum infeasible iterations + if max_infeasible_iter is None: + max_infeasible_iter = n-m + + hits_boundary = False + stop_cond = 1 + counter = 0 + last_feasible_x = np.zeros_like(x) + k = 0 + for i in range(max_iter): + # Stop criteria - Tolerance : r.T g < tol + if rt_g < tol: + stop_cond = 4 + break + k += 1 + # Compute curvature + pt_H_p = H_p.dot(p) + # Stop criteria - Negative curvature + if pt_H_p <= 0: + if np.isinf(trust_radius): + raise ValueError("Negative curvature not allowed " + "for unrestricted problems.") + else: + # Find intersection with constraints + _, alpha, intersect = box_sphere_intersections( + x, p, lb, ub, trust_radius, entire_line=True) + # Update solution + if intersect: + x = x + alpha*p + # Reinforce variables are inside box constraints. + # This is only necessary because of roundoff errors. + x = reinforce_box_boundaries(x, lb, ub) + # Attribute information + stop_cond = 3 + hits_boundary = True + break + + # Get next step + alpha = rt_g / pt_H_p + x_next = x + alpha*p + + # Stop criteria - Hits boundary + if np.linalg.norm(x_next) >= trust_radius: + # Find intersection with box constraints + _, theta, intersect = box_sphere_intersections(x, alpha*p, lb, ub, + trust_radius) + # Update solution + if intersect: + x = x + theta*alpha*p + # Reinforce variables are inside box constraints. + # This is only necessary because of roundoff errors. + x = reinforce_box_boundaries(x, lb, ub) + # Attribute information + stop_cond = 2 + hits_boundary = True + break + + # Check if ``x`` is inside the box and start counter if it is not. + if inside_box_boundaries(x_next, lb, ub): + counter = 0 + else: + counter += 1 + # Whenever outside box constraints keep looking for intersections. + if counter > 0: + _, theta, intersect = box_sphere_intersections(x, alpha*p, lb, ub, + trust_radius) + if intersect: + last_feasible_x = x + theta*alpha*p + # Reinforce variables are inside box constraints. + # This is only necessary because of roundoff errors. + last_feasible_x = reinforce_box_boundaries(last_feasible_x, + lb, ub) + counter = 0 + # Stop after too many infeasible (regarding box constraints) iteration. + if counter > max_infeasible_iter: + break + # Store ``x_next`` value + if return_all: + allvecs.append(x_next) + + # Update residual + r_next = r + alpha*H_p + # Project residual g+ = Z r+ + g_next = Z.dot(r_next) + # Compute conjugate direction step d + rt_g_next = norm(g_next)**2 # g.T g = r.T g (ref [1]_ p.1389) + beta = rt_g_next / rt_g + p = - g_next + beta*p + # Prepare for next iteration + x = x_next + g = g_next + r = g_next + rt_g = norm(g)**2 # g.T g = r.T Z g = r.T g (ref [1]_ p.1389) + H_p = H.dot(p) + + if not inside_box_boundaries(x, lb, ub): + x = last_feasible_x + hits_boundary = True + info = {'niter': k, 'stop_cond': stop_cond, + 'hits_boundary': hits_boundary} + if return_all: + info['allvecs'] = allvecs + return x, info diff --git a/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/report.py b/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/report.py new file mode 100644 index 0000000000000000000000000000000000000000..5c3b2fb4ef5c90da78ae3f181159140e87393dcf --- /dev/null +++ b/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/report.py @@ -0,0 +1,51 @@ +"""Progress report printers.""" + +from __future__ import annotations + +class ReportBase: + COLUMN_NAMES: list[str] = NotImplemented + COLUMN_WIDTHS: list[int] = NotImplemented + ITERATION_FORMATS: list[str] = NotImplemented + + @classmethod + def print_header(cls): + fmt = ("|" + + "|".join([f"{{:^{x}}}" for x in cls.COLUMN_WIDTHS]) + + "|") + separators = ['-' * x for x in cls.COLUMN_WIDTHS] + print(fmt.format(*cls.COLUMN_NAMES)) + print(fmt.format(*separators)) + + @classmethod + def print_iteration(cls, *args): + iteration_format = [f"{{:{x}}}" for x in cls.ITERATION_FORMATS] + fmt = "|" + "|".join(iteration_format) + "|" + print(fmt.format(*args)) + + @classmethod + def print_footer(cls): + print() + + +class BasicReport(ReportBase): + COLUMN_NAMES = ["niter", "f evals", "CG iter", "obj func", "tr radius", + "opt", "c viol"] + COLUMN_WIDTHS = [7, 7, 7, 13, 10, 10, 10] + ITERATION_FORMATS = ["^7", "^7", "^7", "^+13.4e", + "^10.2e", "^10.2e", "^10.2e"] + + +class SQPReport(ReportBase): + COLUMN_NAMES = ["niter", "f evals", "CG iter", "obj func", "tr radius", + "opt", "c viol", "penalty", "CG stop"] + COLUMN_WIDTHS = [7, 7, 7, 13, 10, 10, 10, 10, 7] + ITERATION_FORMATS = ["^7", "^7", "^7", "^+13.4e", "^10.2e", "^10.2e", + "^10.2e", "^10.2e", "^7"] + + +class IPReport(ReportBase): + COLUMN_NAMES = ["niter", "f evals", "CG iter", "obj func", "tr radius", + "opt", "c viol", "penalty", "barrier param", "CG stop"] + COLUMN_WIDTHS = [7, 7, 7, 13, 10, 10, 10, 10, 13, 7] + ITERATION_FORMATS = ["^7", "^7", "^7", "^+13.4e", "^10.2e", "^10.2e", + "^10.2e", "^10.2e", "^13.2e", "^7"] diff --git a/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/tests/__init__.py b/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/tests/__init__.py new file mode 100644 index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 diff --git a/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/tests/__pycache__/__init__.cpython-310.pyc b/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/tests/__pycache__/__init__.cpython-310.pyc new file mode 100644 index 0000000000000000000000000000000000000000..375959df5e14dfd3fe370a18c5c72d676b0e5369 Binary files /dev/null and b/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/tests/__pycache__/__init__.cpython-310.pyc differ diff --git a/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/tests/__pycache__/test_projections.cpython-310.pyc b/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/tests/__pycache__/test_projections.cpython-310.pyc new file mode 100644 index 0000000000000000000000000000000000000000..4d8a6c6951caa6d58123fe8e5ad43853e4f4a540 Binary files /dev/null and b/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/tests/__pycache__/test_projections.cpython-310.pyc differ diff --git a/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/tests/__pycache__/test_report.cpython-310.pyc b/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/tests/__pycache__/test_report.cpython-310.pyc new file mode 100644 index 0000000000000000000000000000000000000000..4d0381bc8cfb0862fab3a34d04ac142fb33e4be4 Binary files /dev/null and b/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/tests/__pycache__/test_report.cpython-310.pyc differ diff --git a/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/tests/test_canonical_constraint.py b/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/tests/test_canonical_constraint.py new file mode 100644 index 0000000000000000000000000000000000000000..452b327d02da3b3bd3fab9592bdef4d56d6aff57 --- /dev/null +++ b/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/tests/test_canonical_constraint.py @@ -0,0 +1,296 @@ +import numpy as np +from numpy.testing import assert_array_equal, assert_equal +from scipy.optimize._constraints import (NonlinearConstraint, Bounds, + PreparedConstraint) +from scipy.optimize._trustregion_constr.canonical_constraint \ + import CanonicalConstraint, initial_constraints_as_canonical + + +def create_quadratic_function(n, m, rng): + a = rng.rand(m) + A = rng.rand(m, n) + H = rng.rand(m, n, n) + HT = np.transpose(H, (1, 2, 0)) + + def fun(x): + return a + A.dot(x) + 0.5 * H.dot(x).dot(x) + + def jac(x): + return A + H.dot(x) + + def hess(x, v): + return HT.dot(v) + + return fun, jac, hess + + +def test_bounds_cases(): + # Test 1: no constraints. + user_constraint = Bounds(-np.inf, np.inf) + x0 = np.array([-1, 2]) + prepared_constraint = PreparedConstraint(user_constraint, x0, False) + c = CanonicalConstraint.from_PreparedConstraint(prepared_constraint) + + assert_equal(c.n_eq, 0) + assert_equal(c.n_ineq, 0) + + c_eq, c_ineq = c.fun(x0) + assert_array_equal(c_eq, []) + assert_array_equal(c_ineq, []) + + J_eq, J_ineq = c.jac(x0) + assert_array_equal(J_eq, np.empty((0, 2))) + assert_array_equal(J_ineq, np.empty((0, 2))) + + assert_array_equal(c.keep_feasible, []) + + # Test 2: infinite lower bound. + user_constraint = Bounds(-np.inf, [0, np.inf, 1], [False, True, True]) + x0 = np.array([-1, -2, -3], dtype=float) + prepared_constraint = PreparedConstraint(user_constraint, x0, False) + c = CanonicalConstraint.from_PreparedConstraint(prepared_constraint) + + assert_equal(c.n_eq, 0) + assert_equal(c.n_ineq, 2) + + c_eq, c_ineq = c.fun(x0) + assert_array_equal(c_eq, []) + assert_array_equal(c_ineq, [-1, -4]) + + J_eq, J_ineq = c.jac(x0) + assert_array_equal(J_eq, np.empty((0, 3))) + assert_array_equal(J_ineq, np.array([[1, 0, 0], [0, 0, 1]])) + + assert_array_equal(c.keep_feasible, [False, True]) + + # Test 3: infinite upper bound. + user_constraint = Bounds([0, 1, -np.inf], np.inf, [True, False, True]) + x0 = np.array([1, 2, 3], dtype=float) + prepared_constraint = PreparedConstraint(user_constraint, x0, False) + c = CanonicalConstraint.from_PreparedConstraint(prepared_constraint) + + assert_equal(c.n_eq, 0) + assert_equal(c.n_ineq, 2) + + c_eq, c_ineq = c.fun(x0) + assert_array_equal(c_eq, []) + assert_array_equal(c_ineq, [-1, -1]) + + J_eq, J_ineq = c.jac(x0) + assert_array_equal(J_eq, np.empty((0, 3))) + assert_array_equal(J_ineq, np.array([[-1, 0, 0], [0, -1, 0]])) + + assert_array_equal(c.keep_feasible, [True, False]) + + # Test 4: interval constraint. + user_constraint = Bounds([-1, -np.inf, 2, 3], [1, np.inf, 10, 3], + [False, True, True, True]) + x0 = np.array([0, 10, 8, 5]) + prepared_constraint = PreparedConstraint(user_constraint, x0, False) + c = CanonicalConstraint.from_PreparedConstraint(prepared_constraint) + + assert_equal(c.n_eq, 1) + assert_equal(c.n_ineq, 4) + + c_eq, c_ineq = c.fun(x0) + assert_array_equal(c_eq, [2]) + assert_array_equal(c_ineq, [-1, -2, -1, -6]) + + J_eq, J_ineq = c.jac(x0) + assert_array_equal(J_eq, [[0, 0, 0, 1]]) + assert_array_equal(J_ineq, [[1, 0, 0, 0], + [0, 0, 1, 0], + [-1, 0, 0, 0], + [0, 0, -1, 0]]) + + assert_array_equal(c.keep_feasible, [False, True, False, True]) + + +def test_nonlinear_constraint(): + n = 3 + m = 5 + rng = np.random.RandomState(0) + x0 = rng.rand(n) + + fun, jac, hess = create_quadratic_function(n, m, rng) + f = fun(x0) + J = jac(x0) + + lb = [-10, 3, -np.inf, -np.inf, -5] + ub = [10, 3, np.inf, 3, np.inf] + user_constraint = NonlinearConstraint( + fun, lb, ub, jac, hess, [True, False, False, True, False]) + + for sparse_jacobian in [False, True]: + prepared_constraint = PreparedConstraint(user_constraint, x0, + sparse_jacobian) + c = CanonicalConstraint.from_PreparedConstraint(prepared_constraint) + + assert_array_equal(c.n_eq, 1) + assert_array_equal(c.n_ineq, 4) + + c_eq, c_ineq = c.fun(x0) + assert_array_equal(c_eq, [f[1] - lb[1]]) + assert_array_equal(c_ineq, [f[3] - ub[3], lb[4] - f[4], + f[0] - ub[0], lb[0] - f[0]]) + + J_eq, J_ineq = c.jac(x0) + if sparse_jacobian: + J_eq = J_eq.toarray() + J_ineq = J_ineq.toarray() + + assert_array_equal(J_eq, J[1, None]) + assert_array_equal(J_ineq, np.vstack((J[3], -J[4], J[0], -J[0]))) + + v_eq = rng.rand(c.n_eq) + v_ineq = rng.rand(c.n_ineq) + v = np.zeros(m) + v[1] = v_eq[0] + v[3] = v_ineq[0] + v[4] = -v_ineq[1] + v[0] = v_ineq[2] - v_ineq[3] + assert_array_equal(c.hess(x0, v_eq, v_ineq), hess(x0, v)) + + assert_array_equal(c.keep_feasible, [True, False, True, True]) + + +def test_concatenation(): + rng = np.random.RandomState(0) + n = 4 + x0 = rng.rand(n) + + f1 = x0 + J1 = np.eye(n) + lb1 = [-1, -np.inf, -2, 3] + ub1 = [1, np.inf, np.inf, 3] + bounds = Bounds(lb1, ub1, [False, False, True, False]) + + fun, jac, hess = create_quadratic_function(n, 5, rng) + f2 = fun(x0) + J2 = jac(x0) + lb2 = [-10, 3, -np.inf, -np.inf, -5] + ub2 = [10, 3, np.inf, 5, np.inf] + nonlinear = NonlinearConstraint( + fun, lb2, ub2, jac, hess, [True, False, False, True, False]) + + for sparse_jacobian in [False, True]: + bounds_prepared = PreparedConstraint(bounds, x0, sparse_jacobian) + nonlinear_prepared = PreparedConstraint(nonlinear, x0, sparse_jacobian) + + c1 = CanonicalConstraint.from_PreparedConstraint(bounds_prepared) + c2 = CanonicalConstraint.from_PreparedConstraint(nonlinear_prepared) + c = CanonicalConstraint.concatenate([c1, c2], sparse_jacobian) + + assert_equal(c.n_eq, 2) + assert_equal(c.n_ineq, 7) + + c_eq, c_ineq = c.fun(x0) + assert_array_equal(c_eq, [f1[3] - lb1[3], f2[1] - lb2[1]]) + assert_array_equal(c_ineq, [lb1[2] - f1[2], f1[0] - ub1[0], + lb1[0] - f1[0], f2[3] - ub2[3], + lb2[4] - f2[4], f2[0] - ub2[0], + lb2[0] - f2[0]]) + + J_eq, J_ineq = c.jac(x0) + if sparse_jacobian: + J_eq = J_eq.toarray() + J_ineq = J_ineq.toarray() + + assert_array_equal(J_eq, np.vstack((J1[3], J2[1]))) + assert_array_equal(J_ineq, np.vstack((-J1[2], J1[0], -J1[0], J2[3], + -J2[4], J2[0], -J2[0]))) + + v_eq = rng.rand(c.n_eq) + v_ineq = rng.rand(c.n_ineq) + v = np.zeros(5) + v[1] = v_eq[1] + v[3] = v_ineq[3] + v[4] = -v_ineq[4] + v[0] = v_ineq[5] - v_ineq[6] + H = c.hess(x0, v_eq, v_ineq).dot(np.eye(n)) + assert_array_equal(H, hess(x0, v)) + + assert_array_equal(c.keep_feasible, + [True, False, False, True, False, True, True]) + + +def test_empty(): + x = np.array([1, 2, 3]) + c = CanonicalConstraint.empty(3) + assert_equal(c.n_eq, 0) + assert_equal(c.n_ineq, 0) + + c_eq, c_ineq = c.fun(x) + assert_array_equal(c_eq, []) + assert_array_equal(c_ineq, []) + + J_eq, J_ineq = c.jac(x) + assert_array_equal(J_eq, np.empty((0, 3))) + assert_array_equal(J_ineq, np.empty((0, 3))) + + H = c.hess(x, None, None).toarray() + assert_array_equal(H, np.zeros((3, 3))) + + +def test_initial_constraints_as_canonical(): + # rng is only used to generate the coefficients of the quadratic + # function that is used by the nonlinear constraint. + rng = np.random.RandomState(0) + + x0 = np.array([0.5, 0.4, 0.3, 0.2]) + n = len(x0) + + lb1 = [-1, -np.inf, -2, 3] + ub1 = [1, np.inf, np.inf, 3] + bounds = Bounds(lb1, ub1, [False, False, True, False]) + + fun, jac, hess = create_quadratic_function(n, 5, rng) + lb2 = [-10, 3, -np.inf, -np.inf, -5] + ub2 = [10, 3, np.inf, 5, np.inf] + nonlinear = NonlinearConstraint( + fun, lb2, ub2, jac, hess, [True, False, False, True, False]) + + for sparse_jacobian in [False, True]: + bounds_prepared = PreparedConstraint(bounds, x0, sparse_jacobian) + nonlinear_prepared = PreparedConstraint(nonlinear, x0, sparse_jacobian) + + f1 = bounds_prepared.fun.f + J1 = bounds_prepared.fun.J + f2 = nonlinear_prepared.fun.f + J2 = nonlinear_prepared.fun.J + + c_eq, c_ineq, J_eq, J_ineq = initial_constraints_as_canonical( + n, [bounds_prepared, nonlinear_prepared], sparse_jacobian) + + assert_array_equal(c_eq, [f1[3] - lb1[3], f2[1] - lb2[1]]) + assert_array_equal(c_ineq, [lb1[2] - f1[2], f1[0] - ub1[0], + lb1[0] - f1[0], f2[3] - ub2[3], + lb2[4] - f2[4], f2[0] - ub2[0], + lb2[0] - f2[0]]) + + if sparse_jacobian: + J1 = J1.toarray() + J2 = J2.toarray() + J_eq = J_eq.toarray() + J_ineq = J_ineq.toarray() + + assert_array_equal(J_eq, np.vstack((J1[3], J2[1]))) + assert_array_equal(J_ineq, np.vstack((-J1[2], J1[0], -J1[0], J2[3], + -J2[4], J2[0], -J2[0]))) + + +def test_initial_constraints_as_canonical_empty(): + n = 3 + for sparse_jacobian in [False, True]: + c_eq, c_ineq, J_eq, J_ineq = initial_constraints_as_canonical( + n, [], sparse_jacobian) + + assert_array_equal(c_eq, []) + assert_array_equal(c_ineq, []) + + if sparse_jacobian: + J_eq = J_eq.toarray() + J_ineq = J_ineq.toarray() + + assert_array_equal(J_eq, np.empty((0, n))) + assert_array_equal(J_ineq, np.empty((0, n))) diff --git a/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/tests/test_projections.py b/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/tests/test_projections.py new file mode 100644 index 0000000000000000000000000000000000000000..6ff3c39d649d0ac663d9b71bb906f1daac021118 --- /dev/null +++ b/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/tests/test_projections.py @@ -0,0 +1,214 @@ +import numpy as np +import scipy.linalg +from scipy.sparse import csc_matrix +from scipy.optimize._trustregion_constr.projections \ + import projections, orthogonality +from numpy.testing import (TestCase, assert_array_almost_equal, + assert_equal, assert_allclose) + +try: + from sksparse.cholmod import cholesky_AAt # noqa: F401 + sksparse_available = True + available_sparse_methods = ("NormalEquation", "AugmentedSystem") +except ImportError: + sksparse_available = False + available_sparse_methods = ("AugmentedSystem",) +available_dense_methods = ('QRFactorization', 'SVDFactorization') + + +class TestProjections(TestCase): + + def test_nullspace_and_least_squares_sparse(self): + A_dense = np.array([[1, 2, 3, 4, 0, 5, 0, 7], + [0, 8, 7, 0, 1, 5, 9, 0], + [1, 0, 0, 0, 0, 1, 2, 3]]) + At_dense = A_dense.T + A = csc_matrix(A_dense) + test_points = ([1, 2, 3, 4, 5, 6, 7, 8], + [1, 10, 3, 0, 1, 6, 7, 8], + [1.12, 10, 0, 0, 100000, 6, 0.7, 8]) + + for method in available_sparse_methods: + Z, LS, _ = projections(A, method) + for z in test_points: + # Test if x is in the null_space + x = Z.matvec(z) + assert_array_almost_equal(A.dot(x), 0) + # Test orthogonality + assert_array_almost_equal(orthogonality(A, x), 0) + # Test if x is the least square solution + x = LS.matvec(z) + x2 = scipy.linalg.lstsq(At_dense, z)[0] + assert_array_almost_equal(x, x2) + + def test_iterative_refinements_sparse(self): + A_dense = np.array([[1, 2, 3, 4, 0, 5, 0, 7], + [0, 8, 7, 0, 1, 5, 9, 0], + [1, 0, 0, 0, 0, 1, 2, 3]]) + A = csc_matrix(A_dense) + test_points = ([1, 2, 3, 4, 5, 6, 7, 8], + [1, 10, 3, 0, 1, 6, 7, 8], + [1.12, 10, 0, 0, 100000, 6, 0.7, 8], + [1, 0, 0, 0, 0, 1, 2, 3+1e-10]) + + for method in available_sparse_methods: + Z, LS, _ = projections(A, method, orth_tol=1e-18, max_refin=100) + for z in test_points: + # Test if x is in the null_space + x = Z.matvec(z) + atol = 1e-13 * abs(x).max() + assert_allclose(A.dot(x), 0, atol=atol) + # Test orthogonality + assert_allclose(orthogonality(A, x), 0, atol=1e-13) + + def test_rowspace_sparse(self): + A_dense = np.array([[1, 2, 3, 4, 0, 5, 0, 7], + [0, 8, 7, 0, 1, 5, 9, 0], + [1, 0, 0, 0, 0, 1, 2, 3]]) + A = csc_matrix(A_dense) + test_points = ([1, 2, 3], + [1, 10, 3], + [1.12, 10, 0]) + + for method in available_sparse_methods: + _, _, Y = projections(A, method) + for z in test_points: + # Test if x is solution of A x = z + x = Y.matvec(z) + assert_array_almost_equal(A.dot(x), z) + # Test if x is in the return row space of A + A_ext = np.vstack((A_dense, x)) + assert_equal(np.linalg.matrix_rank(A_dense), + np.linalg.matrix_rank(A_ext)) + + def test_nullspace_and_least_squares_dense(self): + A = np.array([[1, 2, 3, 4, 0, 5, 0, 7], + [0, 8, 7, 0, 1, 5, 9, 0], + [1, 0, 0, 0, 0, 1, 2, 3]]) + At = A.T + test_points = ([1, 2, 3, 4, 5, 6, 7, 8], + [1, 10, 3, 0, 1, 6, 7, 8], + [1.12, 10, 0, 0, 100000, 6, 0.7, 8]) + + for method in available_dense_methods: + Z, LS, _ = projections(A, method) + for z in test_points: + # Test if x is in the null_space + x = Z.matvec(z) + assert_array_almost_equal(A.dot(x), 0) + # Test orthogonality + assert_array_almost_equal(orthogonality(A, x), 0) + # Test if x is the least square solution + x = LS.matvec(z) + x2 = scipy.linalg.lstsq(At, z)[0] + assert_array_almost_equal(x, x2) + + def test_compare_dense_and_sparse(self): + D = np.diag(range(1, 101)) + A = np.hstack([D, D, D, D]) + A_sparse = csc_matrix(A) + np.random.seed(0) + + Z, LS, Y = projections(A) + Z_sparse, LS_sparse, Y_sparse = projections(A_sparse) + for k in range(20): + z = np.random.normal(size=(400,)) + assert_array_almost_equal(Z.dot(z), Z_sparse.dot(z)) + assert_array_almost_equal(LS.dot(z), LS_sparse.dot(z)) + x = np.random.normal(size=(100,)) + assert_array_almost_equal(Y.dot(x), Y_sparse.dot(x)) + + def test_compare_dense_and_sparse2(self): + D1 = np.diag([-1.7, 1, 0.5]) + D2 = np.diag([1, -0.6, -0.3]) + D3 = np.diag([-0.3, -1.5, 2]) + A = np.hstack([D1, D2, D3]) + A_sparse = csc_matrix(A) + np.random.seed(0) + + Z, LS, Y = projections(A) + Z_sparse, LS_sparse, Y_sparse = projections(A_sparse) + for k in range(1): + z = np.random.normal(size=(9,)) + assert_array_almost_equal(Z.dot(z), Z_sparse.dot(z)) + assert_array_almost_equal(LS.dot(z), LS_sparse.dot(z)) + x = np.random.normal(size=(3,)) + assert_array_almost_equal(Y.dot(x), Y_sparse.dot(x)) + + def test_iterative_refinements_dense(self): + A = np.array([[1, 2, 3, 4, 0, 5, 0, 7], + [0, 8, 7, 0, 1, 5, 9, 0], + [1, 0, 0, 0, 0, 1, 2, 3]]) + test_points = ([1, 2, 3, 4, 5, 6, 7, 8], + [1, 10, 3, 0, 1, 6, 7, 8], + [1, 0, 0, 0, 0, 1, 2, 3+1e-10]) + + for method in available_dense_methods: + Z, LS, _ = projections(A, method, orth_tol=1e-18, max_refin=10) + for z in test_points: + # Test if x is in the null_space + x = Z.matvec(z) + assert_allclose(A.dot(x), 0, rtol=0, atol=2.5e-14) + # Test orthogonality + assert_allclose(orthogonality(A, x), 0, rtol=0, atol=5e-16) + + def test_rowspace_dense(self): + A = np.array([[1, 2, 3, 4, 0, 5, 0, 7], + [0, 8, 7, 0, 1, 5, 9, 0], + [1, 0, 0, 0, 0, 1, 2, 3]]) + test_points = ([1, 2, 3], + [1, 10, 3], + [1.12, 10, 0]) + + for method in available_dense_methods: + _, _, Y = projections(A, method) + for z in test_points: + # Test if x is solution of A x = z + x = Y.matvec(z) + assert_array_almost_equal(A.dot(x), z) + # Test if x is in the return row space of A + A_ext = np.vstack((A, x)) + assert_equal(np.linalg.matrix_rank(A), + np.linalg.matrix_rank(A_ext)) + + +class TestOrthogonality(TestCase): + + def test_dense_matrix(self): + A = np.array([[1, 2, 3, 4, 0, 5, 0, 7], + [0, 8, 7, 0, 1, 5, 9, 0], + [1, 0, 0, 0, 0, 1, 2, 3]]) + test_vectors = ([-1.98931144, -1.56363389, + -0.84115584, 2.2864762, + 5.599141, 0.09286976, + 1.37040802, -0.28145812], + [697.92794044, -4091.65114008, + -3327.42316335, 836.86906951, + 99434.98929065, -1285.37653682, + -4109.21503806, 2935.29289083]) + test_expected_orth = (0, 0) + + for i in range(len(test_vectors)): + x = test_vectors[i] + orth = test_expected_orth[i] + assert_array_almost_equal(orthogonality(A, x), orth) + + def test_sparse_matrix(self): + A = np.array([[1, 2, 3, 4, 0, 5, 0, 7], + [0, 8, 7, 0, 1, 5, 9, 0], + [1, 0, 0, 0, 0, 1, 2, 3]]) + A = csc_matrix(A) + test_vectors = ([-1.98931144, -1.56363389, + -0.84115584, 2.2864762, + 5.599141, 0.09286976, + 1.37040802, -0.28145812], + [697.92794044, -4091.65114008, + -3327.42316335, 836.86906951, + 99434.98929065, -1285.37653682, + -4109.21503806, 2935.29289083]) + test_expected_orth = (0, 0) + + for i in range(len(test_vectors)): + x = test_vectors[i] + orth = test_expected_orth[i] + assert_array_almost_equal(orthogonality(A, x), orth) diff --git a/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/tests/test_qp_subproblem.py b/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/tests/test_qp_subproblem.py new file mode 100644 index 0000000000000000000000000000000000000000..e0235caace6c19563efc31fdf4b8e41d9d81819b --- /dev/null +++ b/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/tests/test_qp_subproblem.py @@ -0,0 +1,645 @@ +import numpy as np +from scipy.sparse import csc_matrix +from scipy.optimize._trustregion_constr.qp_subproblem \ + import (eqp_kktfact, + projected_cg, + box_intersections, + sphere_intersections, + box_sphere_intersections, + modified_dogleg) +from scipy.optimize._trustregion_constr.projections \ + import projections +from numpy.testing import TestCase, assert_array_almost_equal, assert_equal +import pytest + + +class TestEQPDirectFactorization(TestCase): + + # From Example 16.2 Nocedal/Wright "Numerical + # Optimization" p.452. + def test_nocedal_example(self): + H = csc_matrix([[6, 2, 1], + [2, 5, 2], + [1, 2, 4]]) + A = csc_matrix([[1, 0, 1], + [0, 1, 1]]) + c = np.array([-8, -3, -3]) + b = -np.array([3, 0]) + x, lagrange_multipliers = eqp_kktfact(H, c, A, b) + assert_array_almost_equal(x, [2, -1, 1]) + assert_array_almost_equal(lagrange_multipliers, [3, -2]) + + +class TestSphericalBoundariesIntersections(TestCase): + + def test_2d_sphere_constraints(self): + # Interior inicial point + ta, tb, intersect = sphere_intersections([0, 0], + [1, 0], 0.5) + assert_array_almost_equal([ta, tb], [0, 0.5]) + assert_equal(intersect, True) + + # No intersection between line and circle + ta, tb, intersect = sphere_intersections([2, 0], + [0, 1], 1) + assert_equal(intersect, False) + + # Outside initial point pointing toward outside the circle + ta, tb, intersect = sphere_intersections([2, 0], + [1, 0], 1) + assert_equal(intersect, False) + + # Outside initial point pointing toward inside the circle + ta, tb, intersect = sphere_intersections([2, 0], + [-1, 0], 1.5) + assert_array_almost_equal([ta, tb], [0.5, 1]) + assert_equal(intersect, True) + + # Initial point on the boundary + ta, tb, intersect = sphere_intersections([2, 0], + [1, 0], 2) + assert_array_almost_equal([ta, tb], [0, 0]) + assert_equal(intersect, True) + + def test_2d_sphere_constraints_line_intersections(self): + # Interior initial point + ta, tb, intersect = sphere_intersections([0, 0], + [1, 0], 0.5, + entire_line=True) + assert_array_almost_equal([ta, tb], [-0.5, 0.5]) + assert_equal(intersect, True) + + # No intersection between line and circle + ta, tb, intersect = sphere_intersections([2, 0], + [0, 1], 1, + entire_line=True) + assert_equal(intersect, False) + + # Outside initial point pointing toward outside the circle + ta, tb, intersect = sphere_intersections([2, 0], + [1, 0], 1, + entire_line=True) + assert_array_almost_equal([ta, tb], [-3, -1]) + assert_equal(intersect, True) + + # Outside initial point pointing toward inside the circle + ta, tb, intersect = sphere_intersections([2, 0], + [-1, 0], 1.5, + entire_line=True) + assert_array_almost_equal([ta, tb], [0.5, 3.5]) + assert_equal(intersect, True) + + # Initial point on the boundary + ta, tb, intersect = sphere_intersections([2, 0], + [1, 0], 2, + entire_line=True) + assert_array_almost_equal([ta, tb], [-4, 0]) + assert_equal(intersect, True) + + +class TestBoxBoundariesIntersections(TestCase): + + def test_2d_box_constraints(self): + # Box constraint in the direction of vector d + ta, tb, intersect = box_intersections([2, 0], [0, 2], + [1, 1], [3, 3]) + assert_array_almost_equal([ta, tb], [0.5, 1]) + assert_equal(intersect, True) + + # Negative direction + ta, tb, intersect = box_intersections([2, 0], [0, 2], + [1, -3], [3, -1]) + assert_equal(intersect, False) + + # Some constraints are absent (set to +/- inf) + ta, tb, intersect = box_intersections([2, 0], [0, 2], + [-np.inf, 1], + [np.inf, np.inf]) + assert_array_almost_equal([ta, tb], [0.5, 1]) + assert_equal(intersect, True) + + # Intersect on the face of the box + ta, tb, intersect = box_intersections([1, 0], [0, 1], + [1, 1], [3, 3]) + assert_array_almost_equal([ta, tb], [1, 1]) + assert_equal(intersect, True) + + # Interior initial point + ta, tb, intersect = box_intersections([0, 0], [4, 4], + [-2, -3], [3, 2]) + assert_array_almost_equal([ta, tb], [0, 0.5]) + assert_equal(intersect, True) + + # No intersection between line and box constraints + ta, tb, intersect = box_intersections([2, 0], [0, 2], + [-3, -3], [-1, -1]) + assert_equal(intersect, False) + ta, tb, intersect = box_intersections([2, 0], [0, 2], + [-3, 3], [-1, 1]) + assert_equal(intersect, False) + ta, tb, intersect = box_intersections([2, 0], [0, 2], + [-3, -np.inf], + [-1, np.inf]) + assert_equal(intersect, False) + ta, tb, intersect = box_intersections([0, 0], [1, 100], + [1, 1], [3, 3]) + assert_equal(intersect, False) + ta, tb, intersect = box_intersections([0.99, 0], [0, 2], + [1, 1], [3, 3]) + assert_equal(intersect, False) + + # Initial point on the boundary + ta, tb, intersect = box_intersections([2, 2], [0, 1], + [-2, -2], [2, 2]) + assert_array_almost_equal([ta, tb], [0, 0]) + assert_equal(intersect, True) + + def test_2d_box_constraints_entire_line(self): + # Box constraint in the direction of vector d + ta, tb, intersect = box_intersections([2, 0], [0, 2], + [1, 1], [3, 3], + entire_line=True) + assert_array_almost_equal([ta, tb], [0.5, 1.5]) + assert_equal(intersect, True) + + # Negative direction + ta, tb, intersect = box_intersections([2, 0], [0, 2], + [1, -3], [3, -1], + entire_line=True) + assert_array_almost_equal([ta, tb], [-1.5, -0.5]) + assert_equal(intersect, True) + + # Some constraints are absent (set to +/- inf) + ta, tb, intersect = box_intersections([2, 0], [0, 2], + [-np.inf, 1], + [np.inf, np.inf], + entire_line=True) + assert_array_almost_equal([ta, tb], [0.5, np.inf]) + assert_equal(intersect, True) + + # Intersect on the face of the box + ta, tb, intersect = box_intersections([1, 0], [0, 1], + [1, 1], [3, 3], + entire_line=True) + assert_array_almost_equal([ta, tb], [1, 3]) + assert_equal(intersect, True) + + # Interior initial pointoint + ta, tb, intersect = box_intersections([0, 0], [4, 4], + [-2, -3], [3, 2], + entire_line=True) + assert_array_almost_equal([ta, tb], [-0.5, 0.5]) + assert_equal(intersect, True) + + # No intersection between line and box constraints + ta, tb, intersect = box_intersections([2, 0], [0, 2], + [-3, -3], [-1, -1], + entire_line=True) + assert_equal(intersect, False) + ta, tb, intersect = box_intersections([2, 0], [0, 2], + [-3, 3], [-1, 1], + entire_line=True) + assert_equal(intersect, False) + ta, tb, intersect = box_intersections([2, 0], [0, 2], + [-3, -np.inf], + [-1, np.inf], + entire_line=True) + assert_equal(intersect, False) + ta, tb, intersect = box_intersections([0, 0], [1, 100], + [1, 1], [3, 3], + entire_line=True) + assert_equal(intersect, False) + ta, tb, intersect = box_intersections([0.99, 0], [0, 2], + [1, 1], [3, 3], + entire_line=True) + assert_equal(intersect, False) + + # Initial point on the boundary + ta, tb, intersect = box_intersections([2, 2], [0, 1], + [-2, -2], [2, 2], + entire_line=True) + assert_array_almost_equal([ta, tb], [-4, 0]) + assert_equal(intersect, True) + + def test_3d_box_constraints(self): + # Simple case + ta, tb, intersect = box_intersections([1, 1, 0], [0, 0, 1], + [1, 1, 1], [3, 3, 3]) + assert_array_almost_equal([ta, tb], [1, 1]) + assert_equal(intersect, True) + + # Negative direction + ta, tb, intersect = box_intersections([1, 1, 0], [0, 0, -1], + [1, 1, 1], [3, 3, 3]) + assert_equal(intersect, False) + + # Interior point + ta, tb, intersect = box_intersections([2, 2, 2], [0, -1, 1], + [1, 1, 1], [3, 3, 3]) + assert_array_almost_equal([ta, tb], [0, 1]) + assert_equal(intersect, True) + + def test_3d_box_constraints_entire_line(self): + # Simple case + ta, tb, intersect = box_intersections([1, 1, 0], [0, 0, 1], + [1, 1, 1], [3, 3, 3], + entire_line=True) + assert_array_almost_equal([ta, tb], [1, 3]) + assert_equal(intersect, True) + + # Negative direction + ta, tb, intersect = box_intersections([1, 1, 0], [0, 0, -1], + [1, 1, 1], [3, 3, 3], + entire_line=True) + assert_array_almost_equal([ta, tb], [-3, -1]) + assert_equal(intersect, True) + + # Interior point + ta, tb, intersect = box_intersections([2, 2, 2], [0, -1, 1], + [1, 1, 1], [3, 3, 3], + entire_line=True) + assert_array_almost_equal([ta, tb], [-1, 1]) + assert_equal(intersect, True) + + +class TestBoxSphereBoundariesIntersections(TestCase): + + def test_2d_box_constraints(self): + # Both constraints are active + ta, tb, intersect = box_sphere_intersections([1, 1], [-2, 2], + [-1, -2], [1, 2], 2, + entire_line=False) + assert_array_almost_equal([ta, tb], [0, 0.5]) + assert_equal(intersect, True) + + # None of the constraints are active + ta, tb, intersect = box_sphere_intersections([1, 1], [-1, 1], + [-1, -3], [1, 3], 10, + entire_line=False) + assert_array_almost_equal([ta, tb], [0, 1]) + assert_equal(intersect, True) + + # Box constraints are active + ta, tb, intersect = box_sphere_intersections([1, 1], [-4, 4], + [-1, -3], [1, 3], 10, + entire_line=False) + assert_array_almost_equal([ta, tb], [0, 0.5]) + assert_equal(intersect, True) + + # Spherical constraints are active + ta, tb, intersect = box_sphere_intersections([1, 1], [-4, 4], + [-1, -3], [1, 3], 2, + entire_line=False) + assert_array_almost_equal([ta, tb], [0, 0.25]) + assert_equal(intersect, True) + + # Infeasible problems + ta, tb, intersect = box_sphere_intersections([2, 2], [-4, 4], + [-1, -3], [1, 3], 2, + entire_line=False) + assert_equal(intersect, False) + ta, tb, intersect = box_sphere_intersections([1, 1], [-4, 4], + [2, 4], [2, 4], 2, + entire_line=False) + assert_equal(intersect, False) + + def test_2d_box_constraints_entire_line(self): + # Both constraints are active + ta, tb, intersect = box_sphere_intersections([1, 1], [-2, 2], + [-1, -2], [1, 2], 2, + entire_line=True) + assert_array_almost_equal([ta, tb], [0, 0.5]) + assert_equal(intersect, True) + + # None of the constraints are active + ta, tb, intersect = box_sphere_intersections([1, 1], [-1, 1], + [-1, -3], [1, 3], 10, + entire_line=True) + assert_array_almost_equal([ta, tb], [0, 2]) + assert_equal(intersect, True) + + # Box constraints are active + ta, tb, intersect = box_sphere_intersections([1, 1], [-4, 4], + [-1, -3], [1, 3], 10, + entire_line=True) + assert_array_almost_equal([ta, tb], [0, 0.5]) + assert_equal(intersect, True) + + # Spherical constraints are active + ta, tb, intersect = box_sphere_intersections([1, 1], [-4, 4], + [-1, -3], [1, 3], 2, + entire_line=True) + assert_array_almost_equal([ta, tb], [0, 0.25]) + assert_equal(intersect, True) + + # Infeasible problems + ta, tb, intersect = box_sphere_intersections([2, 2], [-4, 4], + [-1, -3], [1, 3], 2, + entire_line=True) + assert_equal(intersect, False) + ta, tb, intersect = box_sphere_intersections([1, 1], [-4, 4], + [2, 4], [2, 4], 2, + entire_line=True) + assert_equal(intersect, False) + + +class TestModifiedDogleg(TestCase): + + def test_cauchypoint_equalsto_newtonpoint(self): + A = np.array([[1, 8]]) + b = np.array([-16]) + _, _, Y = projections(A) + newton_point = np.array([0.24615385, 1.96923077]) + + # Newton point inside boundaries + x = modified_dogleg(A, Y, b, 2, [-np.inf, -np.inf], [np.inf, np.inf]) + assert_array_almost_equal(x, newton_point) + + # Spherical constraint active + x = modified_dogleg(A, Y, b, 1, [-np.inf, -np.inf], [np.inf, np.inf]) + assert_array_almost_equal(x, newton_point/np.linalg.norm(newton_point)) + + # Box constraints active + x = modified_dogleg(A, Y, b, 2, [-np.inf, -np.inf], [0.1, np.inf]) + assert_array_almost_equal(x, (newton_point/newton_point[0]) * 0.1) + + def test_3d_example(self): + A = np.array([[1, 8, 1], + [4, 2, 2]]) + b = np.array([-16, 2]) + Z, LS, Y = projections(A) + + newton_point = np.array([-1.37090909, 2.23272727, -0.49090909]) + cauchy_point = np.array([0.11165723, 1.73068711, 0.16748585]) + origin = np.zeros_like(newton_point) + + # newton_point inside boundaries + x = modified_dogleg(A, Y, b, 3, [-np.inf, -np.inf, -np.inf], + [np.inf, np.inf, np.inf]) + assert_array_almost_equal(x, newton_point) + + # line between cauchy_point and newton_point contains best point + # (spherical constraint is active). + x = modified_dogleg(A, Y, b, 2, [-np.inf, -np.inf, -np.inf], + [np.inf, np.inf, np.inf]) + z = cauchy_point + d = newton_point-cauchy_point + t = ((x-z)/(d)) + assert_array_almost_equal(t, np.full(3, 0.40807330)) + assert_array_almost_equal(np.linalg.norm(x), 2) + + # line between cauchy_point and newton_point contains best point + # (box constraint is active). + x = modified_dogleg(A, Y, b, 5, [-1, -np.inf, -np.inf], + [np.inf, np.inf, np.inf]) + z = cauchy_point + d = newton_point-cauchy_point + t = ((x-z)/(d)) + assert_array_almost_equal(t, np.full(3, 0.7498195)) + assert_array_almost_equal(x[0], -1) + + # line between origin and cauchy_point contains best point + # (spherical constraint is active). + x = modified_dogleg(A, Y, b, 1, [-np.inf, -np.inf, -np.inf], + [np.inf, np.inf, np.inf]) + z = origin + d = cauchy_point + t = ((x-z)/(d)) + assert_array_almost_equal(t, np.full(3, 0.573936265)) + assert_array_almost_equal(np.linalg.norm(x), 1) + + # line between origin and newton_point contains best point + # (box constraint is active). + x = modified_dogleg(A, Y, b, 2, [-np.inf, -np.inf, -np.inf], + [np.inf, 1, np.inf]) + z = origin + d = newton_point + t = ((x-z)/(d)) + assert_array_almost_equal(t, np.full(3, 0.4478827364)) + assert_array_almost_equal(x[1], 1) + + +class TestProjectCG(TestCase): + + # From Example 16.2 Nocedal/Wright "Numerical + # Optimization" p.452. + def test_nocedal_example(self): + H = csc_matrix([[6, 2, 1], + [2, 5, 2], + [1, 2, 4]]) + A = csc_matrix([[1, 0, 1], + [0, 1, 1]]) + c = np.array([-8, -3, -3]) + b = -np.array([3, 0]) + Z, _, Y = projections(A) + x, info = projected_cg(H, c, Z, Y, b) + assert_equal(info["stop_cond"], 4) + assert_equal(info["hits_boundary"], False) + assert_array_almost_equal(x, [2, -1, 1]) + + def test_compare_with_direct_fact(self): + H = csc_matrix([[6, 2, 1, 3], + [2, 5, 2, 4], + [1, 2, 4, 5], + [3, 4, 5, 7]]) + A = csc_matrix([[1, 0, 1, 0], + [0, 1, 1, 1]]) + c = np.array([-2, -3, -3, 1]) + b = -np.array([3, 0]) + Z, _, Y = projections(A) + x, info = projected_cg(H, c, Z, Y, b, tol=0) + x_kkt, _ = eqp_kktfact(H, c, A, b) + assert_equal(info["stop_cond"], 1) + assert_equal(info["hits_boundary"], False) + assert_array_almost_equal(x, x_kkt) + + def test_trust_region_infeasible(self): + H = csc_matrix([[6, 2, 1, 3], + [2, 5, 2, 4], + [1, 2, 4, 5], + [3, 4, 5, 7]]) + A = csc_matrix([[1, 0, 1, 0], + [0, 1, 1, 1]]) + c = np.array([-2, -3, -3, 1]) + b = -np.array([3, 0]) + trust_radius = 1 + Z, _, Y = projections(A) + with pytest.raises(ValueError): + projected_cg(H, c, Z, Y, b, trust_radius=trust_radius) + + def test_trust_region_barely_feasible(self): + H = csc_matrix([[6, 2, 1, 3], + [2, 5, 2, 4], + [1, 2, 4, 5], + [3, 4, 5, 7]]) + A = csc_matrix([[1, 0, 1, 0], + [0, 1, 1, 1]]) + c = np.array([-2, -3, -3, 1]) + b = -np.array([3, 0]) + trust_radius = 2.32379000772445021283 + Z, _, Y = projections(A) + x, info = projected_cg(H, c, Z, Y, b, + tol=0, + trust_radius=trust_radius) + assert_equal(info["stop_cond"], 2) + assert_equal(info["hits_boundary"], True) + assert_array_almost_equal(np.linalg.norm(x), trust_radius) + assert_array_almost_equal(x, -Y.dot(b)) + + def test_hits_boundary(self): + H = csc_matrix([[6, 2, 1, 3], + [2, 5, 2, 4], + [1, 2, 4, 5], + [3, 4, 5, 7]]) + A = csc_matrix([[1, 0, 1, 0], + [0, 1, 1, 1]]) + c = np.array([-2, -3, -3, 1]) + b = -np.array([3, 0]) + trust_radius = 3 + Z, _, Y = projections(A) + x, info = projected_cg(H, c, Z, Y, b, + tol=0, + trust_radius=trust_radius) + assert_equal(info["stop_cond"], 2) + assert_equal(info["hits_boundary"], True) + assert_array_almost_equal(np.linalg.norm(x), trust_radius) + + def test_negative_curvature_unconstrained(self): + H = csc_matrix([[1, 2, 1, 3], + [2, 0, 2, 4], + [1, 2, 0, 2], + [3, 4, 2, 0]]) + A = csc_matrix([[1, 0, 1, 0], + [0, 1, 0, 1]]) + c = np.array([-2, -3, -3, 1]) + b = -np.array([3, 0]) + Z, _, Y = projections(A) + with pytest.raises(ValueError): + projected_cg(H, c, Z, Y, b, tol=0) + + def test_negative_curvature(self): + H = csc_matrix([[1, 2, 1, 3], + [2, 0, 2, 4], + [1, 2, 0, 2], + [3, 4, 2, 0]]) + A = csc_matrix([[1, 0, 1, 0], + [0, 1, 0, 1]]) + c = np.array([-2, -3, -3, 1]) + b = -np.array([3, 0]) + Z, _, Y = projections(A) + trust_radius = 1000 + x, info = projected_cg(H, c, Z, Y, b, + tol=0, + trust_radius=trust_radius) + assert_equal(info["stop_cond"], 3) + assert_equal(info["hits_boundary"], True) + assert_array_almost_equal(np.linalg.norm(x), trust_radius) + + # The box constraints are inactive at the solution but + # are active during the iterations. + def test_inactive_box_constraints(self): + H = csc_matrix([[6, 2, 1, 3], + [2, 5, 2, 4], + [1, 2, 4, 5], + [3, 4, 5, 7]]) + A = csc_matrix([[1, 0, 1, 0], + [0, 1, 1, 1]]) + c = np.array([-2, -3, -3, 1]) + b = -np.array([3, 0]) + Z, _, Y = projections(A) + x, info = projected_cg(H, c, Z, Y, b, + tol=0, + lb=[0.5, -np.inf, + -np.inf, -np.inf], + return_all=True) + x_kkt, _ = eqp_kktfact(H, c, A, b) + assert_equal(info["stop_cond"], 1) + assert_equal(info["hits_boundary"], False) + assert_array_almost_equal(x, x_kkt) + + # The box constraints active and the termination is + # by maximum iterations (infeasible interaction). + def test_active_box_constraints_maximum_iterations_reached(self): + H = csc_matrix([[6, 2, 1, 3], + [2, 5, 2, 4], + [1, 2, 4, 5], + [3, 4, 5, 7]]) + A = csc_matrix([[1, 0, 1, 0], + [0, 1, 1, 1]]) + c = np.array([-2, -3, -3, 1]) + b = -np.array([3, 0]) + Z, _, Y = projections(A) + x, info = projected_cg(H, c, Z, Y, b, + tol=0, + lb=[0.8, -np.inf, + -np.inf, -np.inf], + return_all=True) + assert_equal(info["stop_cond"], 1) + assert_equal(info["hits_boundary"], True) + assert_array_almost_equal(A.dot(x), -b) + assert_array_almost_equal(x[0], 0.8) + + # The box constraints are active and the termination is + # because it hits boundary (without infeasible interaction). + def test_active_box_constraints_hits_boundaries(self): + H = csc_matrix([[6, 2, 1, 3], + [2, 5, 2, 4], + [1, 2, 4, 5], + [3, 4, 5, 7]]) + A = csc_matrix([[1, 0, 1, 0], + [0, 1, 1, 1]]) + c = np.array([-2, -3, -3, 1]) + b = -np.array([3, 0]) + trust_radius = 3 + Z, _, Y = projections(A) + x, info = projected_cg(H, c, Z, Y, b, + tol=0, + ub=[np.inf, np.inf, 1.6, np.inf], + trust_radius=trust_radius, + return_all=True) + assert_equal(info["stop_cond"], 2) + assert_equal(info["hits_boundary"], True) + assert_array_almost_equal(x[2], 1.6) + + # The box constraints are active and the termination is + # because it hits boundary (infeasible interaction). + def test_active_box_constraints_hits_boundaries_infeasible_iter(self): + H = csc_matrix([[6, 2, 1, 3], + [2, 5, 2, 4], + [1, 2, 4, 5], + [3, 4, 5, 7]]) + A = csc_matrix([[1, 0, 1, 0], + [0, 1, 1, 1]]) + c = np.array([-2, -3, -3, 1]) + b = -np.array([3, 0]) + trust_radius = 4 + Z, _, Y = projections(A) + x, info = projected_cg(H, c, Z, Y, b, + tol=0, + ub=[np.inf, 0.1, np.inf, np.inf], + trust_radius=trust_radius, + return_all=True) + assert_equal(info["stop_cond"], 2) + assert_equal(info["hits_boundary"], True) + assert_array_almost_equal(x[1], 0.1) + + # The box constraints are active and the termination is + # because it hits boundary (no infeasible interaction). + def test_active_box_constraints_negative_curvature(self): + H = csc_matrix([[1, 2, 1, 3], + [2, 0, 2, 4], + [1, 2, 0, 2], + [3, 4, 2, 0]]) + A = csc_matrix([[1, 0, 1, 0], + [0, 1, 0, 1]]) + c = np.array([-2, -3, -3, 1]) + b = -np.array([3, 0]) + Z, _, Y = projections(A) + trust_radius = 1000 + x, info = projected_cg(H, c, Z, Y, b, + tol=0, + ub=[np.inf, np.inf, 100, np.inf], + trust_radius=trust_radius) + assert_equal(info["stop_cond"], 3) + assert_equal(info["hits_boundary"], True) + assert_array_almost_equal(x[2], 100) diff --git a/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/tests/test_report.py b/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/tests/test_report.py new file mode 100644 index 0000000000000000000000000000000000000000..c82796fea723ab043971564306d4b76bdf9f0380 --- /dev/null +++ b/parrot/lib/python3.10/site-packages/scipy/optimize/_trustregion_constr/tests/test_report.py @@ -0,0 +1,34 @@ +import pytest +import numpy as np +from scipy.optimize import minimize, Bounds + +def test_gh10880(): + # checks that verbose reporting works with trust-constr for + # bound-contrained problems + bnds = Bounds(1, 2) + opts = {'maxiter': 1000, 'verbose': 2} + minimize(lambda x: x**2, x0=2., method='trust-constr', + bounds=bnds, options=opts) + + opts = {'maxiter': 1000, 'verbose': 3} + minimize(lambda x: x**2, x0=2., method='trust-constr', + bounds=bnds, options=opts) + +@pytest.mark.xslow +def test_gh12922(): + # checks that verbose reporting works with trust-constr for + # general constraints + def objective(x): + return np.array([(np.sum((x+1)**4))]) + + cons = {'type': 'ineq', 'fun': lambda x: -x[0]**2} + n = 25 + x0 = np.linspace(-5, 5, n) + + opts = {'maxiter': 1000, 'verbose': 2} + minimize(objective, x0=x0, method='trust-constr', + constraints=cons, options=opts) + + opts = {'maxiter': 1000, 'verbose': 3} + minimize(objective, x0=x0, method='trust-constr', + constraints=cons, options=opts) diff --git a/vllm/lib/python3.10/site-packages/transformers/generation/__pycache__/logits_process.cpython-310.pyc b/vllm/lib/python3.10/site-packages/transformers/generation/__pycache__/logits_process.cpython-310.pyc new file mode 100644 index 0000000000000000000000000000000000000000..cf730d32ec8ccbdb8102faeac6c61a75775afbf1 --- /dev/null +++ b/vllm/lib/python3.10/site-packages/transformers/generation/__pycache__/logits_process.cpython-310.pyc @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:83b349ed862ec45f0013a5aed6f0b165e3f93f9b6ac9d35f1ca94c149e36cc55 +size 121302 diff --git a/vllm/lib/python3.10/site-packages/transformers/generation/__pycache__/tf_utils.cpython-310.pyc b/vllm/lib/python3.10/site-packages/transformers/generation/__pycache__/tf_utils.cpython-310.pyc new file mode 100644 index 0000000000000000000000000000000000000000..3a35dc810b282cafb27f0deba82f1a5386e222bf --- /dev/null +++ b/vllm/lib/python3.10/site-packages/transformers/generation/__pycache__/tf_utils.cpython-310.pyc @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9063b9d40ec0d4ed1c0bbefe449b7c458fb783f45c308f4e60f9b2eb15a5c241 +size 103620 diff --git a/vllm/lib/python3.10/site-packages/transformers/models/bloom/__init__.py b/vllm/lib/python3.10/site-packages/transformers/models/bloom/__init__.py new file mode 100644 index 0000000000000000000000000000000000000000..72d1d6e6ca4724235ce46978e5c11e84583d4033 --- /dev/null +++ b/vllm/lib/python3.10/site-packages/transformers/models/bloom/__init__.py @@ -0,0 +1,29 @@ +# Copyright 2024 The HuggingFace Team. All rights reserved. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +from typing import TYPE_CHECKING + +from ...utils import _LazyModule +from ...utils.import_utils import define_import_structure + + +if TYPE_CHECKING: + from .configuration_bloom import * + from .modeling_bloom import * + from .modeling_flax_bloom import * + from .tokenization_bloom_fast import * +else: + import sys + + _file = globals()["__file__"] + sys.modules[__name__] = _LazyModule(__name__, _file, define_import_structure(_file), module_spec=__spec__) diff --git a/vllm/lib/python3.10/site-packages/transformers/models/bloom/__pycache__/__init__.cpython-310.pyc b/vllm/lib/python3.10/site-packages/transformers/models/bloom/__pycache__/__init__.cpython-310.pyc new file mode 100644 index 0000000000000000000000000000000000000000..b68d741b68e16e1b934b7a520531b0a25350081d Binary files /dev/null and b/vllm/lib/python3.10/site-packages/transformers/models/bloom/__pycache__/__init__.cpython-310.pyc differ diff --git a/vllm/lib/python3.10/site-packages/transformers/models/bloom/__pycache__/configuration_bloom.cpython-310.pyc b/vllm/lib/python3.10/site-packages/transformers/models/bloom/__pycache__/configuration_bloom.cpython-310.pyc new file mode 100644 index 0000000000000000000000000000000000000000..13d4be02cdf7dc31af4bf3af530bbc34ead0e3f4 Binary files /dev/null and b/vllm/lib/python3.10/site-packages/transformers/models/bloom/__pycache__/configuration_bloom.cpython-310.pyc differ diff --git a/vllm/lib/python3.10/site-packages/transformers/models/bloom/__pycache__/modeling_flax_bloom.cpython-310.pyc b/vllm/lib/python3.10/site-packages/transformers/models/bloom/__pycache__/modeling_flax_bloom.cpython-310.pyc new file mode 100644 index 0000000000000000000000000000000000000000..1354cc77266698ec4a53b108493f8f37f2764a86 Binary files /dev/null and b/vllm/lib/python3.10/site-packages/transformers/models/bloom/__pycache__/modeling_flax_bloom.cpython-310.pyc differ diff --git a/vllm/lib/python3.10/site-packages/transformers/models/bloom/__pycache__/tokenization_bloom_fast.cpython-310.pyc b/vllm/lib/python3.10/site-packages/transformers/models/bloom/__pycache__/tokenization_bloom_fast.cpython-310.pyc new file mode 100644 index 0000000000000000000000000000000000000000..610bc340969b4a7ced0acf7c2076f53977d9a85f Binary files /dev/null and b/vllm/lib/python3.10/site-packages/transformers/models/bloom/__pycache__/tokenization_bloom_fast.cpython-310.pyc differ diff --git a/vllm/lib/python3.10/site-packages/transformers/models/bloom/configuration_bloom.py b/vllm/lib/python3.10/site-packages/transformers/models/bloom/configuration_bloom.py new file mode 100644 index 0000000000000000000000000000000000000000..ca10c7ce7ed4ef795f8aefad01bcf93bc88099ee --- /dev/null +++ b/vllm/lib/python3.10/site-packages/transformers/models/bloom/configuration_bloom.py @@ -0,0 +1,237 @@ +# coding=utf-8 +# Copyright 2022 the Big Science Workshop and HuggingFace Inc. team. All rights reserved. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +"""Bloom configuration""" + +from collections import OrderedDict +from typing import TYPE_CHECKING, Any, List, Mapping, Optional + +from packaging import version + + +if TYPE_CHECKING: + from ... import PreTrainedTokenizer, TensorType + +from ...configuration_utils import PretrainedConfig +from ...onnx import OnnxConfigWithPast, PatchingSpec +from ...utils import is_torch_available, logging + + +logger = logging.get_logger(__name__) + + +class BloomConfig(PretrainedConfig): + """ + This is the configuration class to store the configuration of a [`BloomModel`]. It is used to instantiate a Bloom + model according to the specified arguments, defining the model architecture. Instantiating a configuration with the + defaults will yield a similar configuration to the Bloom architecture + [bigscience/bloom](https://huggingface.co/bigscience/bloom). + + Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the + documentation from [`PretrainedConfig`] for more information. + + + Args: + vocab_size (`int`, *optional*, defaults to 250880): + Vocabulary size of the Bloom model. Defines the maximum number of different tokens that can be represented + by the `inputs_ids` passed when calling [`BloomModel`]. Check [this + discussion](https://huggingface.co/bigscience/bloom/discussions/120#633d28389addb8530b406c2a) on how the + `vocab_size` has been defined. + hidden_size (`int`, *optional*, defaults to 64): + Dimensionality of the embeddings and hidden states. + n_layer (`int`, *optional*, defaults to 2): + Number of hidden layers in the Transformer encoder. + n_head (`int`, *optional*, defaults to 8): + Number of attention heads for each attention layer in the Transformer encoder. + layer_norm_epsilon (`float`, *optional*, defaults to 1e-5): + The epsilon to use in the layer normalization layers. + initializer_range (`float`, *optional*, defaults to 0.02): + The standard deviation of the truncated_normal_initializer for initializing all weight matrices. + apply_residual_connection_post_layernorm (`bool`, *optional*, defaults to `False`): + If enabled, use the layer norm of the hidden states as the residual in the transformer blocks + hidden_dropout (`float`, *optional*, defaults to 0.1): + Dropout rate of the dropout function on the bias dropout. + attention_dropout (`float`, *optional*, defaults to 0.1): + Dropout rate applied to the attention probs + use_cache (`bool`, *optional*, defaults to `True`): + Whether or not the model should return the last key/values attentions (not used by all models). + pretraining_tp (`int`, *optional*, defaults to `1`): + Experimental feature. Tensor parallelism rank used during pretraining with Megatron. Please refer to [this + document](https://huggingface.co/docs/transformers/parallelism) to understand more about it. This value is + necessary to ensure exact reproducibility of the pretraining results. Please refer to [this + issue](https://github.com/pytorch/pytorch/issues/76232). Note also that this is enabled only when + `slow_but_exact=True`. + slow_but_exact (`bool`, *optional*, defaults to `False`): + Experimental feature. Whether to use slow but exact implementation of the attention mechanism. While + merging the TP rank tensors, due to slicing operations the results may be slightly different between the + model trained on Megatron and our model. Please refer to [this + issue](https://github.com/pytorch/pytorch/issues/76232). A solution to obtain more accurate results is to + enable this feature. Enabling this will hurt the computational time of the inference. Will be probably + resolved in the future once the main model has been fine-tuned with TP_rank=1. + + Example: + + ```python + >>> from transformers import BloomConfig, BloomModel + + >>> # Initializing a Bloom configuration + >>> configuration = BloomConfig() + + >>> # Initializing a model (with random weights) from the configuration + >>> model = BloomModel(configuration) + + >>> # Accessing the model configuration + >>> configuration = model.config + ```""" + + model_type = "bloom" + keys_to_ignore_at_inference = ["past_key_values"] + attribute_map = { + "num_hidden_layers": "n_layer", + "num_attention_heads": "n_head", + } + + def __init__( + self, + vocab_size=250880, + hidden_size=64, + n_layer=2, + n_head=8, + layer_norm_epsilon=1e-5, + initializer_range=0.02, + use_cache=True, + bos_token_id=1, + eos_token_id=2, + apply_residual_connection_post_layernorm=False, + hidden_dropout=0.0, + attention_dropout=0.0, + pretraining_tp=1, # TP rank used when training with megatron + slow_but_exact=False, + **kwargs, + ): + self.vocab_size = vocab_size + # Backward compatibility with n_embed kwarg + n_embed = kwargs.pop("n_embed", None) + self.hidden_size = hidden_size if n_embed is None else n_embed + self.n_layer = n_layer + self.n_head = n_head + self.layer_norm_epsilon = layer_norm_epsilon + self.initializer_range = initializer_range + self.use_cache = use_cache + self.pretraining_tp = pretraining_tp + self.apply_residual_connection_post_layernorm = apply_residual_connection_post_layernorm + self.hidden_dropout = hidden_dropout + self.attention_dropout = attention_dropout + + self.bos_token_id = bos_token_id + self.eos_token_id = eos_token_id + self.slow_but_exact = slow_but_exact + + super().__init__(bos_token_id=bos_token_id, eos_token_id=eos_token_id, **kwargs) + + +class BloomOnnxConfig(OnnxConfigWithPast): + torch_onnx_minimum_version = version.parse("1.12") + + def __init__( + self, + config: PretrainedConfig, + task: str = "default", + patching_specs: List[PatchingSpec] = None, + use_past: bool = False, + ): + super().__init__(config, task=task, patching_specs=patching_specs, use_past=use_past) + if not getattr(self._config, "pad_token_id", None): + # TODO: how to do that better? + self._config.pad_token_id = 0 + + @property + def inputs(self) -> Mapping[str, Mapping[int, str]]: + common_inputs = OrderedDict({"input_ids": {0: "batch", 1: "sequence"}}) + if self.use_past: + # BLOOM stores values on dynamic axis 2. For more details see: https://github.com/huggingface/transformers/pull/18344 + self.fill_with_past_key_values_(common_inputs, direction="inputs", inverted_values_shape=True) + common_inputs["attention_mask"] = {0: "batch", 1: "past_sequence + sequence"} + else: + common_inputs["attention_mask"] = {0: "batch", 1: "sequence"} + + return common_inputs + + @property + def num_layers(self) -> int: + return self._config.n_layer + + @property + def num_attention_heads(self) -> int: + return self._config.n_head + + @property + def atol_for_validation(self) -> float: + return 1e-3 + + def generate_dummy_inputs( + self, + tokenizer: "PreTrainedTokenizer", + batch_size: int = -1, + seq_length: int = -1, + is_pair: bool = False, + framework: Optional["TensorType"] = None, + ) -> Mapping[str, Any]: + common_inputs = super(OnnxConfigWithPast, self).generate_dummy_inputs( + tokenizer, batch_size=batch_size, seq_length=seq_length, is_pair=is_pair, framework=framework + ) + + # We need to order the input in the way they appears in the forward() + ordered_inputs = OrderedDict({"input_ids": common_inputs["input_ids"]}) + + # Need to add the past_keys + if self.use_past: + if not is_torch_available(): + raise ValueError("Cannot generate dummy past_keys inputs without PyTorch installed.") + else: + import torch + + batch, seqlen = common_inputs["input_ids"].shape + # Not using the same length for past_key_values + past_key_values_length = seqlen + 2 + head_dim = self._config.hidden_size // self.num_attention_heads + past_key_shape = ( + batch * self.num_attention_heads, + head_dim, + past_key_values_length, + ) + past_value_shape = ( + batch * self.num_attention_heads, + past_key_values_length, + head_dim, + ) + ordered_inputs["past_key_values"] = [ + (torch.zeros(past_key_shape), torch.zeros(past_value_shape)) for _ in range(self.num_layers) + ] + + ordered_inputs["attention_mask"] = common_inputs["attention_mask"] + if self.use_past: + mask_dtype = ordered_inputs["attention_mask"].dtype + ordered_inputs["attention_mask"] = torch.cat( + [ordered_inputs["attention_mask"], torch.ones(batch, past_key_values_length, dtype=mask_dtype)], dim=1 + ) + + return ordered_inputs + + @property + def default_onnx_opset(self) -> int: + return 13 + + +__all__ = ["BloomConfig", "BloomOnnxConfig"] diff --git a/vllm/lib/python3.10/site-packages/transformers/models/bloom/modeling_bloom.py b/vllm/lib/python3.10/site-packages/transformers/models/bloom/modeling_bloom.py new file mode 100644 index 0000000000000000000000000000000000000000..f91cd85779a315f99f6caa5275941509e03192d1 --- /dev/null +++ b/vllm/lib/python3.10/site-packages/transformers/models/bloom/modeling_bloom.py @@ -0,0 +1,1398 @@ +# coding=utf-8 +# Copyright 2022 HuggingFace Inc. team and BigScience workshop. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +"""PyTorch BLOOM model.""" + +import math +import warnings +from typing import Optional, Tuple, Union + +import torch +import torch.utils.checkpoint +from torch import nn +from torch.nn import BCEWithLogitsLoss, CrossEntropyLoss, LayerNorm, MSELoss +from torch.nn import functional as F + +from ...cache_utils import Cache, DynamicCache, StaticCache +from ...file_utils import add_code_sample_docstrings, add_start_docstrings, add_start_docstrings_to_model_forward +from ...generation import GenerationMixin +from ...modeling_attn_mask_utils import AttentionMaskConverter +from ...modeling_outputs import ( + BaseModelOutputWithPastAndCrossAttentions, + CausalLMOutputWithCrossAttentions, + QuestionAnsweringModelOutput, + SequenceClassifierOutputWithPast, + TokenClassifierOutput, +) +from ...modeling_utils import PreTrainedModel +from ...utils import ( + is_torch_flex_attn_available, + logging, +) +from .configuration_bloom import BloomConfig + + +if is_torch_flex_attn_available(): + from torch.nn.attention.flex_attention import BlockMask + + from ...integrations.flex_attention import make_flex_block_causal_mask + + +logger = logging.get_logger(__name__) + +_CHECKPOINT_FOR_DOC = "bigscience/bloom-560m" +_CONFIG_FOR_DOC = "BloomConfig" + + +def build_alibi_tensor(attention_mask: torch.Tensor, num_heads: int, dtype: torch.dtype) -> torch.Tensor: + """ + Link to paper: https://arxiv.org/abs/2108.12409 Alibi tensor is not causal as the original paper mentions, it + relies on a translation invariance of softmax for quick implementation: with l being a tensor, and a fixed value + `softmax(l+a) = softmax(l)`. Based on + https://github.com/ofirpress/attention_with_linear_biases/blob/a35aaca144e0eb6b789dfcb46784c4b8e31b7983/fairseq/models/transformer.py#L742 + TODO @thomasw21 this doesn't work as nicely due to the masking strategy, and so masking varies slightly. + + Args: + Returns tensor shaped (batch_size * num_heads, 1, max_seq_len) + attention_mask (`torch.Tensor`): + Token-wise attention mask, this should be of shape (batch_size, max_seq_len). + num_heads (`int`): + number of heads + dtype (`torch.dtype`, *optional*, default=`torch.bfloat16`): + dtype of the output tensor + """ + batch_size, seq_length = attention_mask.shape + closest_power_of_2 = 2 ** math.floor(math.log2(num_heads)) + base = torch.tensor( + 2 ** (-(2 ** -(math.log2(closest_power_of_2) - 3))), device=attention_mask.device, dtype=torch.float32 + ) + powers = torch.arange(1, 1 + closest_power_of_2, device=attention_mask.device, dtype=torch.int32) + slopes = torch.pow(base, powers) + + if closest_power_of_2 != num_heads: + extra_base = torch.tensor( + 2 ** (-(2 ** -(math.log2(2 * closest_power_of_2) - 3))), device=attention_mask.device, dtype=torch.float32 + ) + num_remaining_heads = min(closest_power_of_2, num_heads - closest_power_of_2) + extra_powers = torch.arange(1, 1 + 2 * num_remaining_heads, 2, device=attention_mask.device, dtype=torch.int32) + slopes = torch.cat([slopes, torch.pow(extra_base, extra_powers)], dim=0) + + # Note: alibi will added to the attention bias that will be applied to the query, key product of attention + # => therefore alibi will have to be of shape (batch_size, num_heads, query_length, key_length) + # => here we set (batch_size=1, num_heads=num_heads, query_length=1, key_length=max_length) + # => the query_length dimension will then be broadcasted correctly + # This is more or less identical to T5's relative position bias: + # https://github.com/huggingface/transformers/blob/f681437203baa7671de3174b0fa583c349d9d5e1/src/transformers/models/t5/modeling_t5.py#L527 + arange_tensor = ((attention_mask.cumsum(dim=-1) - 1) * attention_mask)[:, None, :] + alibi = slopes[..., None] * arange_tensor + return alibi.reshape(batch_size * num_heads, 1, seq_length).to(dtype) + + +def dropout_add(x: torch.Tensor, residual: torch.Tensor, prob: float, training: bool) -> torch.Tensor: + """ + Dropout add function + + Args: + x (`torch.tensor`): + input tensor + residual (`torch.tensor`): + residual tensor + prob (`float`): + dropout probability + training (`bool`): + training mode + """ + out = F.dropout(x, p=prob, training=training) + out = residual + out + return out + + +def bloom_gelu_forward(x: torch.Tensor) -> torch.Tensor: + """ + Custom bias GELU function. Adapted from Megatron-DeepSpeed code. Here we use a simple implementation (inference) to + make the model jitable. + + Args: + x (`torch.tensor`): + input hidden states + """ + return x * 0.5 * (1.0 + torch.tanh(0.79788456 * x * (1 + 0.044715 * x * x))) + + +def bloom_gelu_back(g: torch.Tensor, x: torch.Tensor) -> torch.Tensor: + """ + gradient of tanh approximation of gelu gradient of actual gelu is: 0.5 * (1. + torch.erf(x * 0.70710678)) + + 0.3989423 * x * torch.exp(-0.5 * x * x) + + Args: + g (`torch.tensor`): + gradient output tensor + x (`torch.tensor`): + input tensor + """ + x = x[0] # x is a tuple of 1 element, needs to unpack it first + tanh_out = torch.tanh(0.79788456 * x * (1 + 0.044715 * x * x)) + # sqrt(2/pi) * 3 * 0.044715 -> 0.1070322243 + ff = 0.5 * x * ((1 - tanh_out * tanh_out) * (0.79788456 + 0.1070322243 * x * x)) + 0.5 * (1 + tanh_out) + return ff * g + + +class GeLUFunction(torch.autograd.Function): + @staticmethod + def forward(ctx, input: torch.Tensor) -> torch.Tensor: + ctx.save_for_backward(input) + return bloom_gelu_forward(input) + + @staticmethod + def backward(ctx, grad_output: torch.Tensor) -> torch.Tensor: + input = ctx.saved_tensors + tmp = bloom_gelu_back(grad_output, input) + return tmp + + +class BloomGelu(nn.Module): + """ + BloomBiasGelu wrapper function that make use of the simple function on inference mode to make the model + torchscriptable and use the autograd function in training mode to get the accurate results of the gradients Partly + copied from Megatron-DeepSpeed code and adapted for our needs + + See here why autograd functions are not torchscriptable: https://github.com/pytorch/pytorch/issues/22329 + """ + + def __init__(self): + super().__init__() + + def forward(self, x: torch.Tensor) -> torch.Tensor: + if self.training: + return GeLUFunction.apply(x) + else: + return bloom_gelu_forward(x) + + +class BloomAttention(nn.Module): + def __init__(self, config: BloomConfig, layer_idx: Optional[int] = None): + super().__init__() + + self.pretraining_tp = config.pretraining_tp + self.slow_but_exact = config.slow_but_exact + + self.hidden_size = config.hidden_size + self.num_heads = config.n_head + self.head_dim = self.hidden_size // self.num_heads + self.split_size = self.hidden_size + self.hidden_dropout = config.hidden_dropout + + if self.head_dim * self.num_heads != self.hidden_size: + raise ValueError( + f"`hidden_size` must be divisible by num_heads (got `hidden_size`: {self.hidden_size} and `num_heads`:" + f" {self.num_heads})." + ) + + # Layer-wise attention scaling + self.inv_norm_factor = 1.0 / math.sqrt(self.head_dim) + self.beta = 1.0 + self.layer_idx = layer_idx + if layer_idx is None: + logger.warning_once( + f"Instantiating {self.__class__.__name__} without passing a `layer_idx` is not recommended and will " + "lead to errors during the forward call if caching is used. Please make sure to provide a `layer_idx` " + "when creating this class." + ) + + self.query_key_value = nn.Linear(self.hidden_size, 3 * self.hidden_size, bias=True) + self.dense = nn.Linear(self.hidden_size, self.hidden_size) + self.attention_dropout = nn.Dropout(config.attention_dropout) + + def _reshape(self, fused_qkv: torch.Tensor) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]: + """ + Split the last dimension into (num_heads, head_dim) and reshapes to (bs, heads, len, dim) shape + without making any copies, results share same memory storage as `fused_qkv` + + Args: + fused_qkv (`torch.tensor`): [batch_size, seq_length, num_heads * 3 * head_dim] + + Returns: + query: [batch_size, num_heads, seq_length, head_dim] + key: [batch_size, num_heads, seq_length, head_dim] + value: [batch_size, num_heads, seq_length, head_dim] + """ + batch_size, seq_length, three_times_hidden_size = fused_qkv.shape + fused_qkv = fused_qkv.view(batch_size, seq_length, self.num_heads, 3, self.head_dim) + query_layer = fused_qkv[..., 0, :].transpose(1, 2) + key_layer = fused_qkv[..., 1, :].transpose(1, 2) + value_layer = fused_qkv[..., 2, :].transpose(1, 2) + return query_layer, key_layer, value_layer + + def _merge_heads(self, x: torch.Tensor) -> torch.Tensor: + """ + Merge heads together over the last dimension + + Args: + x (`torch.tensor`): [batch_size * num_heads, seq_length, head_dim] + + Returns: + torch.tensor: [batch_size, seq_length, num_heads * head_dim] + """ + # What we want to achieve is: + # batch_size * num_heads, seq_length, head_dim -> batch_size, seq_length, num_heads * head_dim + batch_size_and_num_heads, seq_length, _ = x.shape + batch_size = batch_size_and_num_heads // self.num_heads + + # First view to decompose the batch size + # batch_size * num_heads, seq_length, head_dim -> batch_size, num_heads, seq_length, head_dim + x = x.view(batch_size, self.num_heads, seq_length, self.head_dim) + + # batch_size, num_heads, seq_length, head_dim -> batch_size, seq_length, num_heads, head_dim + x = x.permute(0, 2, 1, 3) + + # batch_size, seq_length, num_heads, head_dim -> batch_size, seq_length, num_heads * head_dim + return x.reshape(batch_size, seq_length, self.num_heads * self.head_dim) + + def forward( + self, + hidden_states: torch.Tensor, + residual: torch.Tensor, + alibi: torch.Tensor, + attention_mask: torch.Tensor, + layer_past: Optional[Cache] = None, + head_mask: Optional[torch.Tensor] = None, + use_cache: bool = False, + output_attentions: bool = False, + cache_position: Optional[torch.LongTensor] = None, + ): + batch_size, q_length, _ = hidden_states.shape + fused_qkv = self.query_key_value(hidden_states) # [batch_size, seq_length, 3 x hidden_size] + # 3 x [batch_size, num_heads, seq_length, head_dim] + query_layer, key_layer, value_layer = self._reshape(fused_qkv) + + if layer_past is not None: + cache_kwargs = {"cache_position": cache_position} + key_layer, value_layer = layer_past.update(key_layer, value_layer, self.layer_idx, cache_kwargs) + + # reshape qkv for further computations + query_layer = query_layer.reshape(batch_size * self.num_heads, -1, self.head_dim) + key_layer = key_layer.reshape(batch_size * self.num_heads, -1, self.head_dim).transpose(-1, -2) + value_layer = value_layer.reshape(batch_size * self.num_heads, -1, self.head_dim) + + # [batch_size * num_heads, q_length, kv_length] + attention_scores = alibi.baddbmm( + batch1=query_layer, + batch2=key_layer, + beta=self.beta, + alpha=self.inv_norm_factor, + ) + + # change view to [batch_size, num_heads, q_length, kv_length] + attn_weights = attention_scores.view(batch_size, self.num_heads, q_length, -1) + if attention_mask is not None: # no matter the length, we just slice it + causal_mask = attention_mask[:, :, :, : key_layer.shape[-1]] + attn_weights = attn_weights + causal_mask + + # cast attention scores to fp32, compute scaled softmax and cast back to initial dtype + attention_probs = F.softmax(attn_weights, dim=-1, dtype=torch.float32).to(query_layer.dtype) + + # [batch_size, num_heads, q_length, kv_length] + attention_probs = self.attention_dropout(attention_probs) + + if head_mask is not None: + attention_probs = attention_probs * head_mask + + # change view [batch_size x num_heads, q_length, kv_length] + attention_probs_reshaped = attention_probs.view(batch_size * self.num_heads, q_length, -1) + + # matmul: [batch_size * num_heads, q_length, head_dim] + context_layer = torch.bmm(attention_probs_reshaped, value_layer) + + # change view [batch_size, q_length, num_heads * head_dim] + context_layer = self._merge_heads(context_layer) + + # aggregate results across tp ranks. See here: https://github.com/pytorch/pytorch/issues/76232 + if self.pretraining_tp > 1 and self.slow_but_exact: + slices = self.hidden_size / self.pretraining_tp + output_tensor = torch.zeros_like(context_layer) + for i in range(self.pretraining_tp): + output_tensor = output_tensor + F.linear( + context_layer[:, :, int(i * slices) : int((i + 1) * slices)], + self.dense.weight[:, int(i * slices) : int((i + 1) * slices)], + ) + else: + output_tensor = self.dense(context_layer) + + output_tensor = dropout_add(output_tensor, residual, self.hidden_dropout, self.training) + + outputs = (output_tensor, layer_past) + if output_attentions: + outputs += (attention_probs,) + + return outputs + + +class BloomMLP(nn.Module): + def __init__(self, config: BloomConfig): + super().__init__() + hidden_size = config.hidden_size + + self.pretraining_tp = config.pretraining_tp + self.slow_but_exact = config.slow_but_exact + self.dense_h_to_4h = nn.Linear(hidden_size, 4 * hidden_size) + self.gelu_impl = BloomGelu() + self.dense_4h_to_h = nn.Linear(4 * hidden_size, hidden_size) + self.hidden_dropout = config.hidden_dropout + + def forward(self, hidden_states: torch.Tensor, residual: torch.Tensor) -> torch.Tensor: + hidden_states = self.gelu_impl(self.dense_h_to_4h(hidden_states)) + + if self.pretraining_tp > 1 and self.slow_but_exact: + intermediate_output = torch.zeros_like(residual) + slices = self.dense_4h_to_h.weight.shape[-1] / self.pretraining_tp + for i in range(self.pretraining_tp): + intermediate_output = intermediate_output + F.linear( + hidden_states[:, :, int(i * slices) : int((i + 1) * slices)], + self.dense_4h_to_h.weight[:, int(i * slices) : int((i + 1) * slices)], + ) + else: + intermediate_output = self.dense_4h_to_h(hidden_states) + + output = dropout_add(intermediate_output, residual, self.hidden_dropout, self.training) + + return output + + +class BloomBlock(nn.Module): + def __init__(self, config: BloomConfig, layer_idx: Optional[int] = None): + super().__init__() + hidden_size = config.hidden_size + + self.input_layernorm = LayerNorm(hidden_size, eps=config.layer_norm_epsilon) + self.num_heads = config.n_head + self.self_attention = BloomAttention(config, layer_idx) + self.post_attention_layernorm = LayerNorm(hidden_size, eps=config.layer_norm_epsilon) + + self.mlp = BloomMLP(config) + + self.apply_residual_connection_post_layernorm = config.apply_residual_connection_post_layernorm + self.hidden_dropout = config.hidden_dropout + + def forward( + self, + hidden_states: torch.Tensor, + alibi: torch.Tensor, + attention_mask: torch.Tensor, + layer_past: Optional[Cache] = None, + head_mask: Optional[torch.Tensor] = None, + use_cache: bool = False, + output_attentions: bool = False, + cache_position: Optional[torch.LongTensor] = None, + ): + # hidden_states: [batch_size, seq_length, hidden_size] + + # Layer norm at the beginning of the transformer layer. + layernorm_output = self.input_layernorm(hidden_states) + + # Layer norm post the self attention. + if self.apply_residual_connection_post_layernorm: + residual = layernorm_output + else: + residual = hidden_states + + # Self attention. + attn_outputs = self.self_attention( + layernorm_output, + residual, + layer_past=layer_past, + attention_mask=attention_mask, + alibi=alibi, + head_mask=head_mask, + use_cache=use_cache, + output_attentions=output_attentions, + cache_position=cache_position, + ) + + attention_output = attn_outputs[0] + + outputs = attn_outputs[1:] + + layernorm_output = self.post_attention_layernorm(attention_output) + + # Get residual + if self.apply_residual_connection_post_layernorm: + residual = layernorm_output + else: + residual = attention_output + + # MLP. + output = self.mlp(layernorm_output, residual) + + if use_cache: + outputs = (output,) + outputs + else: + outputs = (output,) + outputs[1:] + + return outputs # hidden_states, past_kv, attentions + + +class BloomPreTrainedModel(PreTrainedModel): + config_class = BloomConfig + base_model_prefix = "transformer" + supports_gradient_checkpointing = True + _no_split_modules = ["BloomBlock"] + _skip_keys_device_placement = "past_key_values" + _supports_cache_class = True + _supports_static_cache = True + _supports_quantized_cache = True + + def __init__(self, *inputs, **kwargs): + super().__init__(*inputs, **kwargs) + + def _init_weights(self, module: nn.Module): + """Initialize the weights.""" + if isinstance(module, nn.Linear): + # Slightly different from the TF version which uses truncated_normal for initialization + # cf https://github.com/pytorch/pytorch/pull/5617 + module.weight.data.normal_(mean=0.0, std=self.config.initializer_range) + if module.bias is not None: + module.bias.data.zero_() + elif isinstance(module, nn.Embedding): + module.weight.data.normal_(mean=0.0, std=self.config.initializer_range) + if module.padding_idx is not None: + module.weight.data[module.padding_idx].zero_() + elif isinstance(module, LayerNorm): + module.bias.data.zero_() + module.weight.data.fill_(1.0) + + +BLOOM_START_DOCSTRING = r""" + + This model inherits from [`PreTrainedModel`]. Check the superclass documentation for the generic methods the + library implements for all its model (such as downloading or saving, resizing the input embeddings etc.) + + This model is also a PyTorch [torch.nn.Module](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) subclass. + Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage + and behavior. + + Parameters: + config ([`BloomConfig`]): Model configuration class with all the parameters of the model. + Initializing with a config file does not load the weights associated with the model, only the + configuration. Check out the [`~PreTrainedModel.from_pretrained`] method to load the model weights. +""" + +BLOOM_INPUTS_DOCSTRING = r""" + Args: + input_ids (`torch.LongTensor` of shape `(batch_size, input_ids_length)`): + `input_ids_length` = `sequence_length` if `past_key_values` is `None` else `past_key_values[0][0].shape[2]` + (`sequence_length` of input past key value states). Indices of input sequence tokens in the vocabulary. + + If `past_key_values` is used, only `input_ids` that do not have their past calculated should be passed as + `input_ids`. + + Indices can be obtained using [`AutoTokenizer`]. See [`PreTrainedTokenizer.encode`] and + [`PreTrainedTokenizer.__call__`] for details. + + [What are input IDs?](../glossary#input-ids) + past_key_values (`Cache` or `tuple(tuple(torch.FloatTensor))`, *optional*): + Pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention + blocks) that can be used to speed up sequential decoding. This typically consists in the `past_key_values` + returned by the model at a previous stage of decoding, when `use_cache=True` or `config.use_cache=True`. + + Two formats are allowed: + - a [`~cache_utils.Cache`] instance, see our + [kv cache guide](https://huggingface.co/docs/transformers/en/kv_cache); + - Tuple of `tuple(torch.FloatTensor)` of length `config.n_layers`, with each tuple having 2 tensors of + shape `(batch_size, num_heads, sequence_length, embed_size_per_head)`). This is also known as the legacy + cache format. + + The model will output the same cache format that is fed as input. If no `past_key_values` are passed, the + legacy cache format will be returned. + + If `past_key_values` are used, the user can optionally input only the last `input_ids` (those that don't + have their past key value states given to this model) of shape `(batch_size, 1)` instead of all `input_ids` + of shape `(batch_size, sequence_length)`. + attention_mask (`torch.FloatTensor` of shape `(batch_size, sequence_length)`, *optional*): + Mask to avoid performing attention on padding token indices. Mask values selected in `[0, 1]`: + + - 1 for tokens that are **not masked**, + - 0 for tokens that are **masked**. + + [What are attention masks?](../glossary#attention-mask) + head_mask (`torch.FloatTensor` of shape `(num_heads,)` or `(num_layers, num_heads)`, *optional*): + Mask to nullify selected heads of the self-attention modules. Mask values selected in `[0, 1]`: + + - 1 indicates the head is **not masked**, + - 0 indicates the head is **masked**. + + inputs_embeds (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`, *optional*): + Optionally, instead of passing `input_ids` you can choose to directly pass an embedded representation. This + is useful if you want more control over how to convert `input_ids` indices into associated vectors than the + model's internal embedding lookup matrix. + + If `past_key_values` is used, optionally only the last `inputs_embeds` have to be input (see + `past_key_values`). + use_cache (`bool`, *optional*): + If set to `True`, `past_key_values` key value states are returned and can be used to speed up decoding (see + `past_key_values`). + output_attentions (`bool`, *optional*): + Whether or not to return the attentions tensors of all attention layers. See `attentions` under returned + tensors for more detail. + output_hidden_states (`bool`, *optional*): + Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for + more detail. + return_dict (`bool`, *optional*): + Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple. + cache_position (`torch.LongTensor` of shape `(sequence_length)`, *optional*): + Indices depicting the position of the input sequence tokens in the sequence. Contrarily to `position_ids`, + this tensor is not affected by padding. It is used to update the cache in the correct position and to infer + the complete sequence length. +""" + + +@add_start_docstrings( + "The bare Bloom Model transformer outputting raw hidden-states without any specific head on top.", + BLOOM_START_DOCSTRING, +) +class BloomModel(BloomPreTrainedModel): + def __init__(self, config: BloomConfig): + super().__init__(config) + + self.embed_dim = config.hidden_size + self.num_heads = config.n_head + + # Embedding + LN Embedding + self.word_embeddings = nn.Embedding(config.vocab_size, self.embed_dim) + self.word_embeddings_layernorm = LayerNorm(self.embed_dim, eps=config.layer_norm_epsilon) + + # Transformer blocks + self.h = nn.ModuleList([BloomBlock(config, layer_idx=i) for i in range(config.num_hidden_layers)]) + + # Final Layer Norm + self.ln_f = LayerNorm(self.embed_dim, eps=config.layer_norm_epsilon) + + self.gradient_checkpointing = False + + # Initialize weights and apply final processing + self.post_init() + + def build_alibi_tensor(self, attention_mask: torch.Tensor, num_heads: int, dtype: torch.dtype) -> torch.Tensor: + return build_alibi_tensor(attention_mask, num_heads, dtype) + + def get_input_embeddings(self): + return self.word_embeddings + + def set_input_embeddings(self, new_embeddings: torch.Tensor): + self.word_embeddings = new_embeddings + + @add_start_docstrings_to_model_forward(BLOOM_INPUTS_DOCSTRING) + @add_code_sample_docstrings( + checkpoint=_CHECKPOINT_FOR_DOC, + output_type=BaseModelOutputWithPastAndCrossAttentions, + config_class=_CONFIG_FOR_DOC, + ) + def forward( + self, + input_ids: Optional[torch.LongTensor] = None, + past_key_values: Optional[Union[Cache, Tuple[Tuple[torch.Tensor, torch.Tensor], ...]]] = None, + attention_mask: Optional[torch.Tensor] = None, + head_mask: Optional[torch.LongTensor] = None, + inputs_embeds: Optional[torch.LongTensor] = None, + use_cache: Optional[bool] = None, + output_attentions: Optional[bool] = None, + output_hidden_states: Optional[bool] = None, + return_dict: Optional[bool] = None, + cache_position: Optional[torch.LongTensor] = None, + **deprecated_arguments, + ) -> Union[Tuple[torch.Tensor, ...], BaseModelOutputWithPastAndCrossAttentions]: + if deprecated_arguments.pop("position_ids", False) is not False: + # `position_ids` could have been `torch.Tensor` or `None` so defaulting pop to `False` allows to detect if users were passing explicitly `None` + warnings.warn( + "`position_ids` have no functionality in BLOOM and will be removed in v5.0.0. You can safely ignore" + " passing `position_ids`.", + FutureWarning, + ) + if len(deprecated_arguments) > 0: + raise ValueError(f"Got unexpected arguments: {deprecated_arguments}") + + output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions + output_hidden_states = ( + output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states + ) + use_cache = use_cache if use_cache is not None else self.config.use_cache + return_dict = return_dict if return_dict is not None else self.config.use_return_dict + + if (input_ids is None) ^ (inputs_embeds is not None): + raise ValueError("You must specify exactly one of input_ids or inputs_embeds") + + if self.gradient_checkpointing and self.training and use_cache: + logger.warning_once( + "`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`..." + ) + use_cache = False + + if inputs_embeds is None: + inputs_embeds = self.word_embeddings(input_ids) + + # kept for BC (non `Cache` `past_key_values` inputs) + return_legacy_cache = False + if use_cache and not isinstance(past_key_values, Cache): + return_legacy_cache = True + if past_key_values is None: + past_key_values = DynamicCache() + else: + past_key_values = DynamicCache.from_legacy_cache(past_key_values) + logger.warning_once( + "We detected that you are passing `past_key_values` as a tuple of tuples. This is deprecated and " + "will be removed in v4.47. Please convert your cache or use an appropriate `Cache` class " + "(https://huggingface.co/docs/transformers/kv_cache#legacy-cache-format)" + ) + + batch_size, seq_length, _ = inputs_embeds.shape + past_length = past_key_values.get_seq_length() if past_key_values is not None else 0 + seq_length_with_past = seq_length + past_length + if cache_position is None: + cache_position = torch.arange(past_length, past_length + seq_length, device=inputs_embeds.device) + + # Prepare head mask if needed + # 1.0 in head_mask indicate we keep the head + # attention_probs has shape batch_size x num_heads x N x N + # head_mask has shape n_layer x batch x num_heads x N x N + head_mask = self.get_head_mask(head_mask, self.config.n_layer) + hidden_states = self.word_embeddings_layernorm(inputs_embeds) + + next_decoder_cache = None + all_self_attentions = () if output_attentions else None + all_hidden_states = () if output_hidden_states else None + + # Compute alibi tensor: check build_alibi_tensor documentation + if attention_mask is None: + attention_mask = torch.ones((batch_size, seq_length_with_past), device=hidden_states.device) + else: + attention_mask = attention_mask.to(hidden_states.device) + + alibi = self.build_alibi_tensor(attention_mask, self.num_heads, dtype=hidden_states.dtype) + causal_mask = self._update_causal_mask( + attention_mask, inputs_embeds, cache_position, past_key_values, output_attentions + ) + + for i, block in enumerate(self.h): + if output_hidden_states: + all_hidden_states = all_hidden_states + (hidden_states,) + + if self.gradient_checkpointing and self.training: + outputs = self._gradient_checkpointing_func( + block.__call__, + hidden_states, + alibi, + causal_mask, + past_key_values, + head_mask[i], + use_cache, + output_attentions, + cache_position, + ) + else: + outputs = block( + hidden_states, + layer_past=past_key_values, + attention_mask=causal_mask, + head_mask=head_mask[i], + use_cache=use_cache, + output_attentions=output_attentions, + alibi=alibi, + cache_position=cache_position, + ) + + hidden_states = outputs[0] + if use_cache: + next_decoder_cache = outputs[1] + + if output_attentions: + all_self_attentions = all_self_attentions + (outputs[2 if use_cache else 1],) + + # Add last hidden state + hidden_states = self.ln_f(hidden_states) + + if output_hidden_states: + all_hidden_states = all_hidden_states + (hidden_states,) + + next_cache = next_decoder_cache if use_cache else None + if return_legacy_cache: + next_cache = next_cache.to_legacy_cache() + + if not return_dict: + return tuple( + v for v in [hidden_states, next_cache, all_hidden_states, all_self_attentions] if v is not None + ) + + return BaseModelOutputWithPastAndCrossAttentions( + last_hidden_state=hidden_states, + past_key_values=next_cache, + hidden_states=all_hidden_states, + attentions=all_self_attentions, + ) + + # Copied from transformers.models.llama.modeling_llama.LlamaModel._update_causal_mask + def _update_causal_mask( + self, + attention_mask: torch.Tensor, + input_tensor: torch.Tensor, + cache_position: torch.Tensor, + past_key_values: Cache, + output_attentions: bool = False, + ): + if self.config._attn_implementation == "flash_attention_2": + if attention_mask is not None and (attention_mask == 0.0).any(): + return attention_mask + return None + if self.config._attn_implementation == "flex_attention": + if isinstance(attention_mask, torch.Tensor): + attention_mask = make_flex_block_causal_mask(attention_mask) + if isinstance(attention_mask, BlockMask): + return attention_mask + + # For SDPA, when possible, we will rely on its `is_causal` argument instead of its `attn_mask` argument, in + # order to dispatch on Flash Attention 2. This feature is not compatible with static cache, as SDPA will fail + # to infer the attention mask. + past_seen_tokens = past_key_values.get_seq_length() if past_key_values is not None else 0 + using_static_cache = isinstance(past_key_values, StaticCache) + + # When output attentions is True, sdpa implementation's forward method calls the eager implementation's forward + if self.config._attn_implementation == "sdpa" and not using_static_cache and not output_attentions: + if AttentionMaskConverter._ignore_causal_mask_sdpa( + attention_mask, + inputs_embeds=input_tensor, + past_key_values_length=past_seen_tokens, + is_training=self.training, + ): + return None + + dtype, device = input_tensor.dtype, input_tensor.device + sequence_length = input_tensor.shape[1] + if using_static_cache: + target_length = past_key_values.get_max_cache_shape() + else: + target_length = ( + attention_mask.shape[-1] + if isinstance(attention_mask, torch.Tensor) + else past_seen_tokens + sequence_length + 1 + ) + + # In case the provided `attention` mask is 2D, we generate a causal mask here (4D). + causal_mask = self._prepare_4d_causal_attention_mask_with_cache_position( + attention_mask, + sequence_length=sequence_length, + target_length=target_length, + dtype=dtype, + device=device, + cache_position=cache_position, + batch_size=input_tensor.shape[0], + ) + + if ( + self.config._attn_implementation == "sdpa" + and attention_mask is not None + and attention_mask.device.type in ["cuda", "xpu"] + and not output_attentions + ): + # Attend to all tokens in fully masked rows in the causal_mask, for example the relevant first rows when + # using left padding. This is required by F.scaled_dot_product_attention memory-efficient attention path. + # Details: https://github.com/pytorch/pytorch/issues/110213 + min_dtype = torch.finfo(dtype).min + causal_mask = AttentionMaskConverter._unmask_unattended(causal_mask, min_dtype) + + return causal_mask + + @staticmethod + # Copied from transformers.models.llama.modeling_llama.LlamaModel._prepare_4d_causal_attention_mask_with_cache_position + def _prepare_4d_causal_attention_mask_with_cache_position( + attention_mask: torch.Tensor, + sequence_length: int, + target_length: int, + dtype: torch.dtype, + device: torch.device, + cache_position: torch.Tensor, + batch_size: int, + **kwargs, + ): + """ + Creates a causal 4D mask of shape `(batch_size, 1, query_length, key_value_length)` from a 2D mask of shape + `(batch_size, key_value_length)`, or if the input `attention_mask` is already 4D, do nothing. + + Args: + attention_mask (`torch.Tensor`): + A 2D attention mask of shape `(batch_size, key_value_length)` or a 4D attention mask of shape + `(batch_size, 1, query_length, key_value_length)`. + sequence_length (`int`): + The sequence length being processed. + target_length (`int`): + The target length: when generating with static cache, the mask should be as long as the static cache, + to account for the 0 padding, the part of the cache that is not filled yet. + dtype (`torch.dtype`): + The dtype to use for the 4D attention mask. + device (`torch.device`): + The device to place the 4D attention mask on. + cache_position (`torch.Tensor`): + Indices depicting the position of the input sequence tokens in the sequence. + batch_size (`torch.Tensor`): + Batch size. + """ + if attention_mask is not None and attention_mask.dim() == 4: + # In this case we assume that the mask comes already in inverted form and requires no inversion or slicing. + causal_mask = attention_mask + else: + min_dtype = torch.finfo(dtype).min + causal_mask = torch.full( + (sequence_length, target_length), fill_value=min_dtype, dtype=dtype, device=device + ) + if sequence_length != 1: + causal_mask = torch.triu(causal_mask, diagonal=1) + causal_mask *= torch.arange(target_length, device=device) > cache_position.reshape(-1, 1) + causal_mask = causal_mask[None, None, :, :].expand(batch_size, 1, -1, -1) + if attention_mask is not None: + causal_mask = causal_mask.clone() # copy to contiguous memory for in-place edit + mask_length = attention_mask.shape[-1] + padding_mask = causal_mask[:, :, :, :mask_length] + attention_mask[:, None, None, :].to( + causal_mask.device + ) + padding_mask = padding_mask == 0 + causal_mask[:, :, :, :mask_length] = causal_mask[:, :, :, :mask_length].masked_fill( + padding_mask, min_dtype + ) + + return causal_mask + + +@add_start_docstrings( + """ + The Bloom Model transformer with a language modeling head on top (linear layer with weights tied to the input + embeddings). + """, + BLOOM_START_DOCSTRING, +) +class BloomForCausalLM(BloomPreTrainedModel, GenerationMixin): + _tied_weights_keys = ["lm_head.weight"] + + def __init__(self, config: BloomConfig): + super().__init__(config) + self.transformer = BloomModel(config) + self.lm_head = nn.Linear(config.hidden_size, config.vocab_size, bias=False) + + # Initialize weights and apply final processing + self.post_init() + + def get_output_embeddings(self): + return self.lm_head + + def set_output_embeddings(self, new_embeddings: torch.Tensor): + self.lm_head = new_embeddings + + def prepare_inputs_for_generation( + self, + input_ids, + past_key_values=None, + attention_mask=None, + inputs_embeds=None, + cache_position=None, + use_cache=True, + **kwargs, + ): + # Overwriten because of the fixed-shape attention mask creation + + # If we have cache: let's slice `input_ids` through `cache_position`, to keep only the unprocessed tokens + # Exception 1: when passing input_embeds, input_ids may be missing entries + # Exception 2: some generation methods do special slicing of input_ids, so we don't need to do it here + # Exception 3: with synced GPUs cache_position may go out of bounds, but we only want dummy token in that case. + # (we can't check exception 3 while compiling) + # Exception 4: If input_embeds are passed then slice it through `cache_position`, to keep only the unprocessed tokens and + # generate the first token for each sequence. Later use the generated Input ids for continuation. + if past_key_values is not None: + if inputs_embeds is not None and input_ids.shape[1] == 0: # Exception 4 + inputs_embeds = inputs_embeds[:, -cache_position.shape[0] :] + elif ( + inputs_embeds is not None # Exception 1 + or cache_position[-1] >= input_ids.shape[1] # Exception 3 + ): + input_ids = input_ids[:, -cache_position.shape[0] :] + elif input_ids.shape[1] != cache_position.shape[0]: # Default case (the "else", a no op, is Exception 2) + input_ids = input_ids[:, cache_position] + + # if `inputs_embeds` are passed, we only want to use them in the 1st generation step + if inputs_embeds is not None and len(cache_position) == inputs_embeds.shape[1]: + model_inputs = {"inputs_embeds": inputs_embeds, "input_ids": None} + else: + # This `clone` call is needed to avoid recapturing cuda graphs with `torch.compile`'s `mode="reduce-overhead`, as otherwise the + # input `position_ids` would have various stride during the decoding. Here, simply using `.contiguous()` is not sufficient as in + # the batch size = 1 case, `position_ids` is already contiguous but with varying stride which retriggers a capture. + model_inputs = {"input_ids": input_ids.clone(memory_format=torch.contiguous_format), "inputs_embeds": None} + + # This part differs from other models because BLOOM needs a 2D mask to construct alibi tensor + # The only difference is the usage of 2D instead of 4D mask, but the shape will be static + if isinstance(past_key_values, StaticCache) and attention_mask is not None: + target_length = past_key_values.get_max_cache_shape() + batch_size, seq_length = attention_mask.shape + diff = target_length - seq_length + + new_attn_mask = torch.zeros(batch_size, diff, device=attention_mask.device, dtype=attention_mask.dtype) + attention_mask = torch.cat( + [attention_mask, new_attn_mask], + dim=-1, + ) + + model_inputs.update( + { + "cache_position": cache_position, + "past_key_values": past_key_values, + "use_cache": use_cache, + "attention_mask": attention_mask, + } + ) + return model_inputs + + @add_start_docstrings_to_model_forward(BLOOM_INPUTS_DOCSTRING) + @add_code_sample_docstrings( + checkpoint=_CHECKPOINT_FOR_DOC, + output_type=CausalLMOutputWithCrossAttentions, + config_class=_CONFIG_FOR_DOC, + ) + def forward( + self, + input_ids: Optional[torch.LongTensor] = None, + past_key_values: Optional[Union[Cache, Tuple[Tuple[torch.Tensor, torch.Tensor], ...]]] = None, + attention_mask: Optional[torch.Tensor] = None, + head_mask: Optional[torch.Tensor] = None, + inputs_embeds: Optional[torch.Tensor] = None, + labels: Optional[torch.Tensor] = None, + use_cache: Optional[bool] = None, + output_attentions: Optional[bool] = None, + output_hidden_states: Optional[bool] = None, + return_dict: Optional[bool] = None, + cache_position: Optional[torch.LongTensor] = None, + **deprecated_arguments, + ) -> Union[Tuple[torch.Tensor], CausalLMOutputWithCrossAttentions]: + r""" + labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*): + Labels for language modeling. Note that the labels **are shifted** inside the model, i.e. you can set + `labels = input_ids` Indices are selected in `[-100, 0, ..., config.vocab_size]` All labels set to `-100` + are ignored (masked), the loss is only computed for labels in `[0, ..., config.vocab_size]` + """ + # Bloom has deprecated kwargs, so we need to pop num_items_in_batch explicitly + num_items_in_batch = deprecated_arguments.pop("num_items_in_batch", None) + if deprecated_arguments.pop("position_ids", False) is not False: + # `position_ids` could have been `torch.Tensor` or `None` so defaulting pop to `False` allows to detect if users were passing explicitly `None` + warnings.warn( + "`position_ids` have no functionality in BLOOM and will be removed in v5.0.0. You can safely ignore" + " passing `position_ids`.", + FutureWarning, + ) + if len(deprecated_arguments) > 0: + raise ValueError(f"Got unexpected arguments: {deprecated_arguments}") + + return_dict = return_dict if return_dict is not None else self.config.use_return_dict + + transformer_outputs = self.transformer( + input_ids, + past_key_values=past_key_values, + attention_mask=attention_mask, + head_mask=head_mask, + inputs_embeds=inputs_embeds, + use_cache=use_cache, + output_attentions=output_attentions, + output_hidden_states=output_hidden_states, + return_dict=return_dict, + cache_position=cache_position, + ) + hidden_states = transformer_outputs[0] + + lm_logits = self.lm_head(hidden_states) + + loss = None + if labels is not None: + # move labels to correct device to enable model parallelism + labels = labels.to(lm_logits.device) + # Flatten the tokens + loss = self.loss_function( + lm_logits, + labels, + vocab_size=self.config.vocab_size, + num_items_in_batch=num_items_in_batch, + ) + + if not return_dict: + output = (lm_logits,) + transformer_outputs[1:] + return ((loss,) + output) if loss is not None else output + + return CausalLMOutputWithCrossAttentions( + loss=loss, + logits=lm_logits, + past_key_values=transformer_outputs.past_key_values, + hidden_states=transformer_outputs.hidden_states, + attentions=transformer_outputs.attentions, + ) + + def _reorder_cache( + self, past: Tuple[Tuple[torch.Tensor, torch.Tensor], ...], beam_idx: torch.LongTensor + ) -> Tuple[Tuple[torch.Tensor, torch.Tensor], ...]: + """ + This function is used to re-order the `past_key_values` cache if [`~PreTrainedModel.beam_search`] or + [`~PreTrainedModel.beam_sample`] is called. This is required to match `past_key_values` with the correct + beam_idx at every generation step. + + Output shares the same memory storage as `past`. + """ + # Get a copy of `beam_idx` on all the devices where we need those indices. + device_to_beam_idx = { + past_state.device: beam_idx.to(past_state.device) for layer_past in past for past_state in layer_past + } + reordered_past = tuple( + ( + layer_past[0].index_select(0, device_to_beam_idx[layer_past[0].device]), + layer_past[1].index_select(0, device_to_beam_idx[layer_past[0].device]), + ) + for layer_past in past + ) + return reordered_past + + +@add_start_docstrings( + """ + The Bloom Model transformer with a sequence classification head on top (linear layer). + + [`BloomForSequenceClassification`] uses the last token in order to do the classification, as other causal models + (e.g. GPT-1) do. + + Since it does classification on the last token, it requires to know the position of the last token. If a + `pad_token_id` is defined in the configuration, it finds the last token that is not a padding token in each row. If + no `pad_token_id` is defined, it simply takes the last value in each row of the batch. Since it cannot guess the + padding tokens when `inputs_embeds` are passed instead of `input_ids`, it does the same (take the last value in + each row of the batch). + """, + BLOOM_START_DOCSTRING, +) +class BloomForSequenceClassification(BloomPreTrainedModel): + def __init__(self, config: BloomConfig): + super().__init__(config) + self.num_labels = config.num_labels + self.transformer = BloomModel(config) + self.score = nn.Linear(config.hidden_size, config.num_labels, bias=False) + + # Initialize weights and apply final processing + self.post_init() + + @add_start_docstrings_to_model_forward(BLOOM_INPUTS_DOCSTRING) + @add_code_sample_docstrings( + checkpoint=_CHECKPOINT_FOR_DOC, + output_type=SequenceClassifierOutputWithPast, + config_class=_CONFIG_FOR_DOC, + ) + def forward( + self, + input_ids: Optional[torch.LongTensor] = None, + past_key_values: Optional[Union[Cache, Tuple[Tuple[torch.Tensor, torch.Tensor], ...]]] = None, + attention_mask: Optional[torch.Tensor] = None, + head_mask: Optional[torch.Tensor] = None, + inputs_embeds: Optional[torch.Tensor] = None, + labels: Optional[torch.Tensor] = None, + use_cache: Optional[bool] = None, + output_attentions: Optional[bool] = None, + output_hidden_states: Optional[bool] = None, + return_dict: Optional[bool] = None, + **deprecated_arguments, + ) -> Union[Tuple[torch.Tensor], SequenceClassifierOutputWithPast]: + r""" + labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*): + Labels for computing the sequence classification/regression loss. Indices should be in `[0, ..., + config.num_labels - 1]`. If `config.num_labels == 1` a regression loss is computed (Mean-Square loss), If + `config.num_labels > 1` a classification loss is computed (Cross-Entropy). + """ + if deprecated_arguments.pop("position_ids", False) is not False: + # `position_ids` could have been `torch.Tensor` or `None` so defaulting pop to `False` allows to detect if users were passing explicitly `None` + warnings.warn( + "`position_ids` have no functionality in BLOOM and will be removed in v5.0.0. You can safely ignore" + " passing `position_ids`.", + FutureWarning, + ) + if len(deprecated_arguments) > 0: + raise ValueError(f"Got unexpected arguments: {deprecated_arguments}") + + return_dict = return_dict if return_dict is not None else self.config.use_return_dict + + transformer_outputs = self.transformer( + input_ids, + past_key_values=past_key_values, + attention_mask=attention_mask, + head_mask=head_mask, + inputs_embeds=inputs_embeds, + use_cache=use_cache, + output_attentions=output_attentions, + output_hidden_states=output_hidden_states, + return_dict=return_dict, + ) + + hidden_states = transformer_outputs[0] + logits = self.score(hidden_states) + + if input_ids is not None: + batch_size = input_ids.shape[0] + else: + batch_size = inputs_embeds.shape[0] + + if self.config.pad_token_id is None and batch_size != 1: + raise ValueError("Cannot handle batch sizes > 1 if no padding token is defined.") + if self.config.pad_token_id is None: + last_non_pad_token = -1 + elif input_ids is not None: + # To handle both left- and right- padding, we take the rightmost token that is not equal to pad_token_id + non_pad_mask = (input_ids != self.config.pad_token_id).to(logits.device, torch.int32) + token_indices = torch.arange(input_ids.shape[-1], device=logits.device, dtype=torch.int32) + last_non_pad_token = (token_indices * non_pad_mask).argmax(-1) + else: + last_non_pad_token = -1 + logger.warning_once( + f"{self.__class__.__name__} will not detect padding tokens in `inputs_embeds`. Results may be " + "unexpected if using padding tokens in conjunction with `inputs_embeds.`" + ) + + pooled_logits = logits[torch.arange(batch_size, device=logits.device), last_non_pad_token] + + loss = None + if labels is not None: + if self.config.problem_type is None: + if self.num_labels == 1: + self.config.problem_type = "regression" + elif self.num_labels > 1 and (labels.dtype == torch.long or labels.dtype == torch.int): + self.config.problem_type = "single_label_classification" + else: + self.config.problem_type = "multi_label_classification" + + if self.config.problem_type == "regression": + loss_fct = MSELoss() + if self.num_labels == 1: + loss = loss_fct(pooled_logits.squeeze(), labels.squeeze()) + else: + loss = loss_fct(pooled_logits, labels) + elif self.config.problem_type == "single_label_classification": + loss_fct = CrossEntropyLoss() + loss = loss_fct(pooled_logits, labels) + elif self.config.problem_type == "multi_label_classification": + loss_fct = BCEWithLogitsLoss() + loss = loss_fct(pooled_logits, labels) + if not return_dict: + output = (pooled_logits,) + transformer_outputs[1:] + return ((loss,) + output) if loss is not None else output + + return SequenceClassifierOutputWithPast( + loss=loss, + logits=pooled_logits, + past_key_values=transformer_outputs.past_key_values, + hidden_states=transformer_outputs.hidden_states, + attentions=transformer_outputs.attentions, + ) + + +@add_start_docstrings( + """ + Bloom Model with a token classification head on top (a linear layer on top of the hidden-states output) e.g. for + Named-Entity-Recognition (NER) tasks. + """, + BLOOM_START_DOCSTRING, +) +class BloomForTokenClassification(BloomPreTrainedModel): + def __init__(self, config: BloomConfig): + super().__init__(config) + self.num_labels = config.num_labels + + self.transformer = BloomModel(config) + if hasattr(config, "classifier_dropout") and config.classifier_dropout is not None: + classifier_dropout = config.classifier_dropout + elif hasattr(config, "hidden_dropout") and config.hidden_dropout is not None: + classifier_dropout = config.hidden_dropout + else: + classifier_dropout = 0.1 + self.dropout = nn.Dropout(classifier_dropout) + self.classifier = nn.Linear(config.hidden_size, config.num_labels) + + # Initialize weights and apply final processing + self.post_init() + + @add_start_docstrings_to_model_forward(BLOOM_INPUTS_DOCSTRING) + @add_code_sample_docstrings( + checkpoint=_CHECKPOINT_FOR_DOC, + output_type=TokenClassifierOutput, + config_class=_CONFIG_FOR_DOC, + ) + def forward( + self, + input_ids: Optional[torch.LongTensor] = None, + past_key_values: Optional[Union[Cache, Tuple[Tuple[torch.Tensor, torch.Tensor], ...]]] = None, + attention_mask: Optional[torch.Tensor] = None, + head_mask: Optional[torch.Tensor] = None, + inputs_embeds: Optional[torch.Tensor] = None, + labels: Optional[torch.Tensor] = None, + use_cache: Optional[bool] = None, + output_attentions: Optional[bool] = None, + output_hidden_states: Optional[bool] = None, + return_dict: Optional[bool] = None, + **deprecated_arguments, + ) -> Union[Tuple[torch.Tensor], TokenClassifierOutput]: + r""" + labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*): + Labels for computing the sequence classification/regression loss. Indices should be in `[0, ..., + config.num_labels - 1]`. If `config.num_labels == 1` a regression loss is computed (Mean-Square loss), If + `config.num_labels > 1` a classification loss is computed (Cross-Entropy). + """ + if deprecated_arguments.pop("position_ids", False) is not False: + # `position_ids` could have been `torch.Tensor` or `None` so defaulting pop to `False` allows to detect if users were passing explicitly `None` + warnings.warn( + "`position_ids` have no functionality in BLOOM and will be removed in v5.0.0. You can safely ignore" + " passing `position_ids`.", + FutureWarning, + ) + if len(deprecated_arguments) > 0: + raise ValueError(f"Got unexpected arguments: {deprecated_arguments}") + + return_dict = return_dict if return_dict is not None else self.config.use_return_dict + + transformer_outputs = self.transformer( + input_ids, + past_key_values=past_key_values, + attention_mask=attention_mask, + head_mask=head_mask, + inputs_embeds=inputs_embeds, + use_cache=use_cache, + output_attentions=output_attentions, + output_hidden_states=output_hidden_states, + return_dict=return_dict, + ) + + hidden_states = transformer_outputs[0] + hidden_states = self.dropout(hidden_states) + logits = self.classifier(hidden_states) + + loss = None + if labels is not None: + # move labels to correct device to enable model parallelism + labels = labels.to(logits.device) + batch_size, seq_length = labels.shape + loss_fct = CrossEntropyLoss() + loss = loss_fct( + logits.view(batch_size * seq_length, self.num_labels), labels.view(batch_size * seq_length) + ) + + if not return_dict: + output = (logits,) + transformer_outputs[2:] + return ((loss,) + output) if loss is not None else output + + return TokenClassifierOutput( + loss=loss, + logits=logits, + hidden_states=transformer_outputs.hidden_states, + attentions=transformer_outputs.attentions, + ) + + +@add_start_docstrings( + """ + The BLOOM Model transformer with a span classification head on top for extractive question-answering tasks like + SQuAD (a linear layers on top of the hidden-states output to compute `span start logits` and `span end logits`). + """, + BLOOM_START_DOCSTRING, +) +class BloomForQuestionAnswering(BloomPreTrainedModel): + def __init__(self, config): + super().__init__(config) + self.transformer = BloomModel(config) + self.qa_outputs = nn.Linear(config.hidden_size, 2) + + # Initialize weights and apply final processing + self.post_init() + + @add_start_docstrings_to_model_forward(BLOOM_INPUTS_DOCSTRING.format("batch_size, sequence_length")) + def forward( + self, + input_ids: Optional[torch.LongTensor] = None, + attention_mask: Optional[torch.FloatTensor] = None, + position_ids: Optional[torch.LongTensor] = None, + head_mask: Optional[torch.FloatTensor] = None, + inputs_embeds: Optional[torch.FloatTensor] = None, + start_positions: Optional[torch.LongTensor] = None, + end_positions: Optional[torch.LongTensor] = None, + output_attentions: Optional[bool] = None, + output_hidden_states: Optional[bool] = None, + return_dict: Optional[bool] = None, + ) -> Union[Tuple, QuestionAnsweringModelOutput]: + r""" + start_positions (`torch.LongTensor` of shape `(batch_size,)`, *optional*): + Labels for position (index) of the start of the labelled span for computing the token classification loss. + Positions are clamped to the length of the sequence (`sequence_length`). Position outside of the sequence + are not taken into account for computing the loss. + end_positions (`torch.LongTensor` of shape `(batch_size,)`, *optional*): + Labels for position (index) of the end of the labelled span for computing the token classification loss. + Positions are clamped to the length of the sequence (`sequence_length`). Position outside of the sequence + are not taken into account for computing the loss. + """ + return_dict = return_dict if return_dict is not None else self.config.use_return_dict + + outputs = self.transformer( + input_ids, + attention_mask=attention_mask, + position_ids=position_ids, + head_mask=head_mask, + inputs_embeds=inputs_embeds, + output_attentions=output_attentions, + output_hidden_states=output_hidden_states, + return_dict=return_dict, + ) + + sequence_output = outputs[0] + + logits = self.qa_outputs(sequence_output) + start_logits, end_logits = logits.split(1, dim=-1) + start_logits = start_logits.squeeze(-1).contiguous() + end_logits = end_logits.squeeze(-1).contiguous() + + total_loss = None + if start_positions is not None and end_positions is not None: + # If we are on multi-GPU, split add a dimension + if len(start_positions.size()) > 1: + start_positions = start_positions.squeeze(-1) + if len(end_positions.size()) > 1: + end_positions = end_positions.squeeze(-1) + # sometimes the start/end positions are outside our model inputs, we ignore these terms + ignored_index = start_logits.size(1) + start_positions = start_positions.clamp(0, ignored_index) + end_positions = end_positions.clamp(0, ignored_index) + + loss_fct = CrossEntropyLoss(ignore_index=ignored_index) + start_loss = loss_fct(start_logits, start_positions) + end_loss = loss_fct(end_logits, end_positions) + total_loss = (start_loss + end_loss) / 2 + + if not return_dict: + output = (start_logits, end_logits) + outputs[2:] + return ((total_loss,) + output) if total_loss is not None else output + + return QuestionAnsweringModelOutput( + loss=total_loss, + start_logits=start_logits, + end_logits=end_logits, + hidden_states=outputs.hidden_states, + attentions=outputs.attentions, + ) + + +__all__ = [ + "BloomForCausalLM", + "BloomModel", + "BloomPreTrainedModel", + "BloomForSequenceClassification", + "BloomForTokenClassification", + "BloomForQuestionAnswering", +] diff --git a/vllm/lib/python3.10/site-packages/transformers/models/bloom/tokenization_bloom_fast.py b/vllm/lib/python3.10/site-packages/transformers/models/bloom/tokenization_bloom_fast.py new file mode 100644 index 0000000000000000000000000000000000000000..c84322637cb7e879b2cf50fdbd81fdc79b08ed98 --- /dev/null +++ b/vllm/lib/python3.10/site-packages/transformers/models/bloom/tokenization_bloom_fast.py @@ -0,0 +1,152 @@ +# coding=utf-8 +# Copyright 2022 The HuggingFace Inc. team. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +"""Tokenization classes for Bloom.""" + +import pickle +from typing import Optional, Tuple + +from ...tokenization_utils_base import BatchEncoding +from ...tokenization_utils_fast import PreTrainedTokenizerFast +from ...utils import logging + + +logger = logging.get_logger(__name__) + +VOCAB_FILES_NAMES = {"tokenizer_file": "tokenizer.json"} + + +class BloomTokenizerFast(PreTrainedTokenizerFast): + """ + Construct a "fast" Bloom tokenizer (backed by HuggingFace's *tokenizers* library). Based on byte-level + Byte-Pair-Encoding. + + This tokenizer has been trained to treat spaces like parts of the tokens (a bit like sentencepiece) so a word will + be encoded differently whether it is at the beginning of the sentence (without space) or not: + + ```python + >>> from transformers import BloomTokenizerFast + + >>> tokenizer = BloomTokenizerFast.from_pretrained("bigscience/bloom") + >>> tokenizer("Hello world")["input_ids"] + [59414, 8876] + + >>> tokenizer(" Hello world")["input_ids"] + [86153, 8876] + ``` + + You can get around that behavior by passing `add_prefix_space=True` when instantiating this tokenizer, but since + the model was not pretrained this way, it might yield a decrease in performance. + + + + When used with `is_split_into_words=True`, this tokenizer needs to be instantiated with `add_prefix_space=True`. + + + + This tokenizer inherits from [`PreTrainedTokenizerFast`] which contains most of the main methods. Users should + refer to this superclass for more information regarding those methods. + + Args: + vocab_file (`str`): + Path to the vocabulary file. + merges_file (`str`): + Path to the merges file. + errors (`str`, *optional*, defaults to `"replace"`): + Paradigm to follow when decoding bytes to UTF-8. See + [bytes.decode](https://docs.python.org/3/library/stdtypes.html#bytes.decode) for more information. + unk_token (`str`, *optional*, defaults to `<|endoftext|>`): + The unknown token. A token that is not in the vocabulary cannot be converted to an ID and is set to be this + token instead. + bos_token (`str`, *optional*, defaults to `<|endoftext|>`): + The beginning of sequence token. + eos_token (`str`, *optional*, defaults to `<|endoftext|>`): + The end of sequence token. + add_prefix_space (`bool`, *optional*, defaults to `False`): + Whether or not to add an initial space to the input. This allows to treat the leading word just as any + other word. (Bloom tokenizer detect beginning of words by the preceding space). + trim_offsets (`bool`, *optional*, defaults to `True`): + Whether or not the post-processing step should trim offsets to avoid including whitespaces. + """ + + vocab_files_names = VOCAB_FILES_NAMES + model_input_names = ["input_ids", "attention_mask"] + slow_tokenizer_class = None + # No `max_model_input_sizes` as BLOOM uses ALiBi positional embeddings + + def __init__( + self, + vocab_file=None, + merges_file=None, + tokenizer_file=None, + unk_token="", + bos_token="", + eos_token="", + pad_token="", + add_prefix_space=False, + clean_up_tokenization_spaces=False, + **kwargs, + ): + super().__init__( + vocab_file=vocab_file, + merges_file=merges_file, + tokenizer_file=tokenizer_file, + unk_token=unk_token, + bos_token=bos_token, + eos_token=eos_token, + pad_token=pad_token, + add_prefix_space=add_prefix_space, + clean_up_tokenization_spaces=clean_up_tokenization_spaces, + **kwargs, + ) + # TODO @ArthurZucker this can only work one way for now, to update later-on. Tests should also properly + # check this as they were green before. + pre_tok_state = pickle.dumps(self.backend_tokenizer.pre_tokenizer) + decoder_state = pickle.dumps(self.backend_tokenizer.decoder) + + if add_prefix_space: + pre_tok_state = pre_tok_state.replace(b'"add_prefix_space":false', b'"add_prefix_space": true') + decoder_state = decoder_state.replace(b'"add_prefix_space":false', b'"add_prefix_space": true') + self.backend_tokenizer.pre_tokenizer = pickle.loads(pre_tok_state) + self.backend_tokenizer.decoder = pickle.loads(decoder_state) + + self.add_prefix_space = add_prefix_space + + def _batch_encode_plus(self, *args, **kwargs) -> BatchEncoding: + is_split_into_words = kwargs.get("is_split_into_words", False) + if not (self.add_prefix_space or not is_split_into_words): + raise Exception( + f"You need to instantiate {self.__class__.__name__} with add_prefix_space=True to use it with" + " pretokenized inputs." + ) + + return super()._batch_encode_plus(*args, **kwargs) + + def _encode_plus(self, *args, **kwargs) -> BatchEncoding: + is_split_into_words = kwargs.get("is_split_into_words", False) + + if not (self.add_prefix_space or not is_split_into_words): + raise Exception( + f"You need to instantiate {self.__class__.__name__} with add_prefix_space=True to use it with" + " pretokenized inputs." + ) + + return super()._encode_plus(*args, **kwargs) + + def save_vocabulary(self, save_directory: str, filename_prefix: Optional[str] = None) -> Tuple[str]: + files = self._tokenizer.model.save(save_directory, name=filename_prefix) + return tuple(files) + + +__all__ = ["BloomTokenizerFast"] diff --git a/vllm/lib/python3.10/site-packages/transformers/models/cpmant/__init__.py b/vllm/lib/python3.10/site-packages/transformers/models/cpmant/__init__.py new file mode 100644 index 0000000000000000000000000000000000000000..d92eea75693e728f4c1e9afc199d40751ef6af7a --- /dev/null +++ b/vllm/lib/python3.10/site-packages/transformers/models/cpmant/__init__.py @@ -0,0 +1,28 @@ +# Copyright 2024 The HuggingFace Team. All rights reserved. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +from typing import TYPE_CHECKING + +from ...utils import _LazyModule +from ...utils.import_utils import define_import_structure + + +if TYPE_CHECKING: + from .configuration_cpmant import * + from .modeling_cpmant import * + from .tokenization_cpmant import * +else: + import sys + + _file = globals()["__file__"] + sys.modules[__name__] = _LazyModule(__name__, _file, define_import_structure(_file), module_spec=__spec__) diff --git a/vllm/lib/python3.10/site-packages/transformers/models/cpmant/__pycache__/__init__.cpython-310.pyc b/vllm/lib/python3.10/site-packages/transformers/models/cpmant/__pycache__/__init__.cpython-310.pyc new file mode 100644 index 0000000000000000000000000000000000000000..0421f403a4bd44f4d7f8abfff2d84861fc327501 Binary files /dev/null and b/vllm/lib/python3.10/site-packages/transformers/models/cpmant/__pycache__/__init__.cpython-310.pyc differ diff --git a/vllm/lib/python3.10/site-packages/transformers/models/cpmant/__pycache__/configuration_cpmant.cpython-310.pyc b/vllm/lib/python3.10/site-packages/transformers/models/cpmant/__pycache__/configuration_cpmant.cpython-310.pyc new file mode 100644 index 0000000000000000000000000000000000000000..5bb66e6f991381172ceb0c4b7e7cbfaa9447c6bc Binary files /dev/null and b/vllm/lib/python3.10/site-packages/transformers/models/cpmant/__pycache__/configuration_cpmant.cpython-310.pyc differ diff --git a/vllm/lib/python3.10/site-packages/transformers/models/cpmant/__pycache__/modeling_cpmant.cpython-310.pyc b/vllm/lib/python3.10/site-packages/transformers/models/cpmant/__pycache__/modeling_cpmant.cpython-310.pyc new file mode 100644 index 0000000000000000000000000000000000000000..de60a066f3972667140ac0c5f04e9293a9739532 Binary files /dev/null and b/vllm/lib/python3.10/site-packages/transformers/models/cpmant/__pycache__/modeling_cpmant.cpython-310.pyc differ diff --git a/vllm/lib/python3.10/site-packages/transformers/models/cpmant/__pycache__/tokenization_cpmant.cpython-310.pyc b/vllm/lib/python3.10/site-packages/transformers/models/cpmant/__pycache__/tokenization_cpmant.cpython-310.pyc new file mode 100644 index 0000000000000000000000000000000000000000..8356598f7bb61a36cbdb1475aaa0e1c22657a3ae Binary files /dev/null and b/vllm/lib/python3.10/site-packages/transformers/models/cpmant/__pycache__/tokenization_cpmant.cpython-310.pyc differ diff --git a/vllm/lib/python3.10/site-packages/transformers/models/cpmant/modeling_cpmant.py b/vllm/lib/python3.10/site-packages/transformers/models/cpmant/modeling_cpmant.py new file mode 100644 index 0000000000000000000000000000000000000000..df0aebe3cbf850d5ce022352c1f52e9b2d1f8d0e --- /dev/null +++ b/vllm/lib/python3.10/site-packages/transformers/models/cpmant/modeling_cpmant.py @@ -0,0 +1,860 @@ +# coding=utf-8 +# Copyright 2022 The OpenBMB Team and The HuggingFace Inc. team. All rights reserved. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +"""PyTorch CPMAnt""" + +import math +from typing import List, Optional, Tuple, Union + +import torch +import torch.nn.functional as F +import torch.utils.checkpoint +from torch import nn +from torch.nn import CrossEntropyLoss + +from ...activations import ACT2FN +from ...generation import GenerationMixin +from ...modeling_outputs import BaseModelOutputWithPast, CausalLMOutputWithPast +from ...modeling_utils import PreTrainedModel +from ...utils import add_code_sample_docstrings, add_start_docstrings, add_start_docstrings_to_model_forward, logging +from .configuration_cpmant import CpmAntConfig + + +logger = logging.get_logger(__name__) + +_CHECKPOINT_FOR_DOC = "openbmb/cpm-ant-10b" +_CONFIG_FOR_DOC = "CpmAntConfig" + + +class CpmAntLayerNorm(nn.Module): + """ + We use Root Mean Square (RMS) Layer Normalization, please see https://arxiv.org/abs/1910.07467 for details." + """ + + def __init__(self, config: CpmAntConfig): + super().__init__() + + self.eps = config.eps + self.dim_norm = config.hidden_size + self.weight = nn.Parameter(torch.empty(config.hidden_size)) + + def forward(self, hidden_states: torch.Tensor): + """ + Args: + hidden_states (`torch.Tensor` of shape `(batch, seq_len, dim_in)`) + """ + if hidden_states.size(-1) != self.dim_norm: + raise AssertionError("hidden_states.size(-1) != self.dim_norm") + old_dtype = hidden_states.dtype + variance = hidden_states.to(torch.float32).pow(2).mean(dim=-1, keepdim=True) + hidden_states = (hidden_states * torch.rsqrt(variance + self.eps)).to(old_dtype) * self.weight + return hidden_states + + +class CpmAntAttention(nn.Module): + def __init__(self, config: CpmAntConfig): + super().__init__() + self.dim_model = config.hidden_size + self.num_heads = config.num_attention_heads + self.dim_head = config.dim_head + + self.project_q = nn.Linear(self.dim_model, self.num_heads * self.dim_head, bias=False) + self.project_k = nn.Linear(self.dim_model, self.num_heads * self.dim_head, bias=False) + self.project_v = nn.Linear(self.dim_model, self.num_heads * self.dim_head, bias=False) + + self.attention_out = nn.Linear(self.num_heads * self.dim_head, self.dim_model, bias=False) + + self.softmax = torch.nn.Softmax(dim=-1) + + if config.dropout_p is not None: + self.dropout = torch.nn.Dropout(p=config.dropout_p) + else: + self.dropout = None + + def forward( + self, + hidden_q: torch.Tensor, + hidden_kv: torch.Tensor, + attention_mask: torch.BoolTensor, + position_bias: torch.Tensor, + output_attentions: Optional[bool] = False, + past_key_values: Optional[Tuple[torch.Tensor, torch.Tensor]] = None, + use_cache: Optional[bool] = None, + ): + """ + Args: + hidden_q (`torch.Tensor`): + Input of transformer block(self-attention block). It can be the raw embedding of a batch of sequences. + hidden_kv (`torch.Tensor` of shape `(batch, len_k, dim_model)`)): + Tensor *key_value* and *query* of shape `(batch, len_k, dim_model)` + attention_mask (`torch.Tensor` of shape `(batch, len_seq, len_seq)`): + Avoid invalid areas to participate in the calculation of self-attention. + position_bias (`torch.Tensor` of shape `(batch, len_seq, len_seq)`): + Provide positional information to self-attention block. + output_attentions (`bool`, *optional*): + Whether or not to return the attentions tensors of all attention layers. + past_key_values (`Tuple[torch.Tensor, torch.Tensor]`, *optional*): + Cached past key and value projection states. + use_cache (`bool`, *optional*): + If set to `True`, `past_key_values` key value states are returned and can be used to speed up decoding + (see `past_key_values`). + """ + batch_size = hidden_q.size(0) + len_q = hidden_q.size(1) + len_k = hidden_kv.size(1) + + query = self.project_q(hidden_q) + key = self.project_k(hidden_kv) + value = self.project_v(hidden_kv) + + query = query.view(batch_size, len_q, self.num_heads, self.dim_head).permute(0, 2, 1, 3) + key = key.view(batch_size, len_k, self.num_heads, self.dim_head).permute(0, 2, 1, 3) + value = value.view(batch_size, len_k, self.num_heads, self.dim_head).permute(0, 2, 1, 3) + + if past_key_values is not None: + key = torch.cat([past_key_values[0], key], dim=-2) + value = torch.cat([past_key_values[1], value], dim=-2) + len_k = key.size(-2) + + # (batch_size, num_heads, len_q, dim_head) @ (batch_size, num_heads, dim_head, len_k) -> (batch_size, num_heads, len_q, len_k) + score = torch.matmul(query, key.transpose(-1, -2)) / math.sqrt(self.dim_head) + score = score + position_bias + + score = torch.masked_fill( + score, + attention_mask.view(batch_size, 1, len_q, len_k) == torch.tensor(False), + torch.scalar_tensor(float("-inf"), device=score.device, dtype=score.dtype), + ) + score = self.softmax(score) + + score = torch.masked_fill( + score, + attention_mask.view(batch_size, 1, len_q, len_k) == torch.tensor(False), + torch.scalar_tensor(0, device=score.device, dtype=score.dtype), + ) + if output_attentions: + attn_weights = score + else: + attn_weights = None + + if self.dropout is not None: + score = self.dropout(score) + + # (batch_size, num_heads, len_q, len_k) @ (batch_size, num_heads, len_k, dim_head) -> (batch_size, num_heads, len_q, dim_head) + score = torch.matmul(score, value) + + score = score.view(batch_size, self.num_heads, len_q, self.dim_head).permute(0, 2, 1, 3) + score = score.contiguous().view(batch_size, len_q, self.num_heads * self.dim_head) + + score = self.attention_out(score) + + past_key_values = None + if use_cache: + past_key_values = (key, value) + + return score, attn_weights, past_key_values + + +class CpmAntSelfAttentionBlock(nn.Module): + def __init__(self, config: CpmAntConfig): + super().__init__() + self.layernorm_before_attention = CpmAntLayerNorm(config) + self.self_attention = CpmAntAttention(config) + if config.dropout_p: + self.dropout = torch.nn.Dropout(config.dropout_p) + else: + self.dropout = None + + def forward( + self, + hidden_states: torch.Tensor, + attention_mask: torch.Tensor, + position_bias: Optional[torch.Tensor] = None, + output_attentions: Optional[bool] = False, + past_key_values: Optional[Tuple[torch.Tensor, torch.Tensor]] = None, + use_cache: Optional[bool] = None, + ): + """ + Args: + hidden_states (`torch.Tensor` of shape `(batch, len_seq, dim_model)`): + Input of transformer block(self-attention block). It can be the raw embedding of a batch of sequences. + attention_mask (`torch.Tensor` of shape `(batch, len_seq, len_seq)`): + Avoid invalid areas to participate in the calculation of self-attention. + position_bias (`torch.Tensor` of shape `(batch, len_seq, len_seq)`): + Provide positional information to self-attention block. + output_attentions (`bool`, *optional*): + Whether or not to return the attentions tensors of all attention layers. + past_key_values (`Tuple(torch.FloatTensor)`, *optional*): + Cached past key and value projection states. + use_cache (`bool`, *optional*): + If set to `True`, `past_key_values` key value states are returned and can be used to speed up decoding + (see `past_key_values`). + """ + outputs = self.layernorm_before_attention(hidden_states) + outputs = self.self_attention( + outputs, outputs, attention_mask, position_bias, output_attentions, past_key_values, use_cache + ) + + outputs, attn_weights, current_key_value = outputs + + if self.dropout is not None: + outputs = self.dropout(outputs) + hidden_states = hidden_states + outputs + + return hidden_states, attn_weights, current_key_value + + +class CpmAntDenseGatedACT(nn.Module): + def __init__(self, config: CpmAntConfig): + super().__init__() + self.w_0 = nn.Linear(config.hidden_size, config.dim_ff, bias=False) + self.w_1 = nn.Linear(config.hidden_size, config.dim_ff, bias=False) + self.act = torch.nn.GELU() + + def forward(self, hidden_states: torch.Tensor): + """Transform an input tensor from one feature space to another via a nonlinear operation + + Args: + hidden_states (`torch.Tensor` of shape `(batch, seq_len, dim_in)`) + """ + gate_score = self.act(self.w_0(hidden_states)) + hidden_states = self.w_1(hidden_states) + + hidden_states = gate_score * hidden_states + return hidden_states + + +class CpmAntFeedForward(nn.Module): + def __init__(self, config: CpmAntConfig): + super().__init__() + self.w_in = CpmAntDenseGatedACT(config) + if config.dropout_p is not None: + self.dropout = torch.nn.Dropout(config.dropout_p) + else: + self.dropout = None + + self.w_out = nn.Linear(config.dim_ff, config.hidden_size, bias=False) + + def forward(self, hidden_states: torch.Tensor): + """ + Args: + hidden_states (`torch.Tensor` of shape `(batch, seq_len, dim_in)`) + """ + hidden_states = self.w_in(hidden_states) + + if self.dropout is not None: + hidden_states = self.dropout(hidden_states) + + hidden_states = self.w_out(hidden_states) + + return hidden_states + + +class CpmAntFFNBlock(nn.Module): + def __init__(self, config: CpmAntConfig): + super().__init__() + self.layernorm_before_ffn = CpmAntLayerNorm(config) + self.ffn = CpmAntFeedForward(config) + if config.dropout_p: + self.dropout = torch.nn.Dropout(config.dropout_p) + else: + self.dropout = None + + def forward( + self, + hidden_states: torch.Tensor, + ): + """ + Args: + hidden_states (`torch.Tensor` of shape `(batch, len_seq, dim_model)`): + Hidden states before feed forward layer. + """ + ln_outputs = self.layernorm_before_ffn(hidden_states) + outputs = self.ffn(ln_outputs) + if self.dropout is not None: + outputs = self.dropout(outputs) + hidden_states = hidden_states + outputs + return hidden_states + + +class CpmAntTransformerBlock(nn.Module): + def __init__(self, config: CpmAntConfig): + super().__init__() + self.self_att = CpmAntSelfAttentionBlock(config) + self.ffn = CpmAntFFNBlock(config) + + def forward( + self, + hidden_states: torch.Tensor, + attention_mask: torch.Tensor, + position_bias: Optional[torch.Tensor] = None, + output_attentions: Optional[bool] = False, + past_key_values: Optional[Tuple[torch.Tensor, torch.Tensor]] = None, + use_cache: Optional[bool] = None, + ): + """ + Args: + hidden_states (`torch.Tensor`): + Input to the layer of shape `(batch, seq_len, dim_model)` + attention_mask (`torch.Tensor`): + Avoid invalid areas to participate in the calculation of shape `(batch, seq_len, seq_len)` + position_bias (`torch.Tensor`): + Provides position information to attention mechanism of shape `(num_heads, seq_len, seq_len)` + output_attentions (`bool`, *optional*): + Whether or not to return the attentions tensors of all attention layers. + past_key_values (`Tuple[torch.Tensor, torch.Tensor])`, *optional*): + Cached past key and value projection states + use_cache (`bool`, *optional*): + If set to `True`, `past_key_values` key value states are returned and can be used to speed up decoding + (see `past_key_values`). + """ + hidden_states = self.self_att( + hidden_states, + attention_mask=attention_mask, + position_bias=position_bias, + output_attentions=output_attentions, + past_key_values=past_key_values, + use_cache=use_cache, + ) + + hidden_states, attn_weights, current_key_value = hidden_states + + hidden_states = self.ffn(hidden_states) + + return hidden_states, attn_weights, current_key_value + + +class CpmAntEncoder(nn.Module): + def __init__(self, config: CpmAntConfig): + super().__init__() + self.num_layers = config.num_hidden_layers + self.layers = nn.ModuleList([CpmAntTransformerBlock(config) for ith in range(self.num_layers)]) + + self.output_layernorm = CpmAntLayerNorm(config) + + def forward( + self, + hidden_states: torch.Tensor, + attention_mask: torch.Tensor, + position_bias: torch.Tensor, + output_attentions: Optional[bool] = None, + output_hidden_states: Optional[bool] = None, + past_key_values: Optional[Tuple[torch.Tensor, torch.Tensor]] = None, + use_cache: Optional[bool] = None, + ): + """ + Args: + hidden_states (`torch.Tensor`): + Input to the layer of shape `(batch, seq_len, dim_model)` + attention_mask (`torch.Tensor`): + Avoid invalid areas to participate in the calculation of shape `(batch, seq_len, seq_len)` + position_bias (`torch.Tensor`): + Provides position information to attention mechanism of shape `(num_heads, seq_len, seq_len)` + output_attentions (`bool`, *optional*): + Whether or not to return the attentions tensors of all attention layers. + output_hidden_states (`bool`, *optional*): + Whether or not to return the hidden states of all layers. + past_key_values (`Tuple[torch.Tensor, torch.Tensor])`, *optional*): + Cached past key and value projection states + use_cache (`bool`, *optional*): + If set to `True`, `past_key_values` key value states are returned and can be used to speed up decoding + (see `past_key_values`). + """ + all_hidden_states = () if output_hidden_states else None + all_self_attns = () if output_attentions else None + current_key_values = () if use_cache else None + + for i, layer in enumerate(self.layers): + if output_hidden_states: + all_hidden_states += (hidden_states,) + layer_outputs = layer( + hidden_states, + attention_mask, + position_bias, + output_attentions=output_attentions, + past_key_values=past_key_values[i] if past_key_values else None, + use_cache=use_cache, + ) + hidden_states, attn_weights, current_key_value = layer_outputs + if output_attentions: + all_self_attns += (attn_weights,) + if current_key_value is not None: + current_key_values = current_key_values + (current_key_value,) + + hidden_states = self.output_layernorm(hidden_states) + + if output_hidden_states: + all_hidden_states += (hidden_states,) + + return hidden_states, current_key_values, all_hidden_states, all_self_attns + + +# Copied from transformers.models.bert.modeling_bert.BertIntermediate with Bert->CPMAnt +class CpmAntIntermediate(nn.Module): + def __init__(self, config): + super().__init__() + self.dense = nn.Linear(config.hidden_size, config.intermediate_size) + if isinstance(config.hidden_act, str): + self.intermediate_act_fn = ACT2FN[config.hidden_act] + else: + self.intermediate_act_fn = config.hidden_act + + def forward(self, hidden_states: torch.Tensor) -> torch.Tensor: + hidden_states = self.dense(hidden_states) + hidden_states = self.intermediate_act_fn(hidden_states) + return hidden_states + + +class CpmAntSegmentPositionEmbedding(nn.Module): + def __init__(self, config: CpmAntConfig): + super().__init__() + + self.num_heads = config.num_attention_heads + self.num_buckets = config.position_bias_num_buckets + self.max_distance = config.position_bias_max_distance + self.num_segments = config.segment_types + + self.relative_attention_bias = nn.Parameter( + torch.empty( + config.segment_types * config.segment_types + config.position_bias_num_buckets, + config.num_attention_heads, + ) + ) + + def forward( + self, + key_pos: torch.Tensor, + query_pos: torch.Tensor, + key_segment: torch.Tensor, + query_segment: torch.Tensor, + ): + with torch.no_grad(): + batch = key_pos.size(0) + keylen = key_pos.size(1) + querylen = query_pos.size(1) + + if key_pos.size(0) != query_pos.size(0): + raise AssertionError( + f"key_pos.size(0) should be equal to query_pos.size(0), but got {key_pos.size(0)} and {query_pos.size(0)}!" + ) + if keylen != key_segment.size(1) or querylen != query_segment.size(1): + raise AssertionError( + f"keylen should be equal to key_segment.size(1), but got {keylen} and {key_segment.size(1)}!" + ) + if querylen != query_segment.size(1): + raise AssertionError( + f"querylen should be equal to query_segment.size(1), but got {querylen} and {query_segment.szie(1)}!" + ) + + key_pos = key_pos.view(batch, -1, keylen) + query_pos = query_pos.view(batch, querylen, -1) + key_segment = key_segment.view(batch, -1, keylen) + query_segment = query_segment.view(batch, querylen, -1) + + relative_position_bucket = self._segment_relative_position_bucket(query_segment, key_segment) + relative_position_bucket = relative_position_bucket + self.num_buckets + + # (batch, len_q, len_k) + absolute_position_bucket = self._position_bucket( + torch.arange(keylen, dtype=torch.int32, device=relative_position_bucket.device)[None, :] + - torch.arange(querylen, dtype=torch.int32, device=relative_position_bucket.device)[:, None], + num_buckets=self.num_buckets, + max_distance=self.max_distance, + ) + relative_position_bucket = torch.where( + (key_segment == query_segment), + absolute_position_bucket[None, :, :], + relative_position_bucket, + ) + + # (batch, len_q, len_k, num_heads) + embeds = F.embedding(relative_position_bucket, self.relative_attention_bias) + # (batch, num_heads, len_q, len_k) + embeds = embeds.permute(0, 3, 1, 2).contiguous() + return embeds + + def _segment_relative_position_bucket(self, query_segment, key_segment): + return query_segment * self.num_segments + key_segment + + def _position_bucket(self, relative_position, num_buckets=32, max_distance=128): + relative_buckets = 0 + # always bidirectional in CPMAnt + num_buckets //= 2 + relative_buckets = (relative_position > 0).to(torch.int32) * num_buckets + relative_position = torch.abs(relative_position) + max_exact = num_buckets // 2 + is_small = relative_position < max_exact + relative_postion_if_large = max_exact + ( + torch.log(relative_position.float() / max_exact) + / math.log(max_distance / max_exact) + * (num_buckets - max_exact) + ).to(torch.int32) + relative_postion_if_large = torch.min( + relative_postion_if_large, + torch.full_like(relative_postion_if_large, num_buckets - 1), + ) + relative_buckets += torch.where(is_small, relative_position.to(torch.int32), relative_postion_if_large) + return relative_buckets + + +# Copied from transformers.models.bert.modeling_bert.BertOutput with Bert->CPMAnt +class CpmAntOutput(nn.Module): + def __init__(self, config): + super().__init__() + self.dense = nn.Linear(config.intermediate_size, config.hidden_size) + self.LayerNorm = nn.LayerNorm(config.hidden_size, eps=config.layer_norm_eps) + self.dropout = nn.Dropout(config.hidden_dropout_prob) + + def forward(self, hidden_states: torch.Tensor, input_tensor: torch.Tensor) -> torch.Tensor: + hidden_states = self.dense(hidden_states) + hidden_states = self.dropout(hidden_states) + hidden_states = self.LayerNorm(hidden_states + input_tensor) + return hidden_states + + +class CpmAntPreTrainedModel(PreTrainedModel): + """ + An abstract class to handle weights initialization and a simple interface for downloading and loading pretrained + models. + """ + + config_class = CpmAntConfig + base_model_prefix = "cpmant" + + def _init_weights(self, module): + """Initialize the weights""" + if isinstance(module, nn.Linear): + module.weight.data.normal_(mean=0.0, std=self.config.init_std) + if module.bias is not None: + module.bias.data.zero_() + elif isinstance(module, nn.Embedding): + module.weight.data.normal_(mean=0.0, std=self.config.init_std) + if module.padding_idx is not None: + module.weight.data[module.padding_idx].zero_() + elif isinstance(module, nn.LayerNorm): + module.bias.data.zero_() + module.weight.data.fill_(1.0) + elif isinstance(module, CpmAntLayerNorm): + module.weight.data.fill_(1.0) + elif isinstance(module, CpmAntSegmentPositionEmbedding): + module.relative_attention_bias.data.normal_(mean=0.0, std=self.config.init_std) + + +CPMANT_START_DOCSTRING = r""" + This model is a PyTorch [torch.nn.Module](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) sub-class. Use + it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and + behavior. + + Parameters + config ([`~CpmAntConfig`]): Model configuration class with all the parameters of the + Initializing with a config file does not load the weights associated with the model, only the + configuration. Check out the [`~PreTrainedModel.from_pretrained`] method to load the model weights. +""" + +CPMANT_INPUTS_DOCSTRING = r""" + Args: + input_ids (`torch.Tensor` of shape `(batch_size, seq_len)`): + Indices of input sequence tokens in the vocabulary. + + Indices can be obtained using [`CPMAntTokenizer`]. See [`PreTrainedTokenizer.encode`] and + [`PreTrainedTokenizer.__call__`] for details. + + [What are input IDs?](../glossary#input-ids) + past_key_values (`tuple(tuple(torch.FloatTensor))`, *optional*, returned when `use_cache=True` is passed or when `config.use_cache=True`): + Contains pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention + blocks) that can be used (see `past_key_values` input) to speed up sequential decoding. + use_cache (`bool`, *optional*): + If set to `True`, `past_key_values` key value states are returned and can be used to speed up decoding (see + `past_key_values`). + output_attentions (`bool`, *optional*): + Whether or not to return the attentions tensors of all attention layers. + output_hidden_states (`bool`, *optional*): + Whether or not to return the hidden states of all layers. + return_dict (`bool`, *optional*): + Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple. +""" + + +@add_start_docstrings( + "The bare CPMAnt Model outputting raw hidden-states without any specific head on top.", + CPMANT_START_DOCSTRING, +) +class CpmAntModel(CpmAntPreTrainedModel): + def __init__(self, config: CpmAntConfig): + super().__init__(config) + self.encoder = CpmAntEncoder(config) + self.segment_embedding = nn.Embedding(config.segment_types, config.hidden_size) + self.input_embedding = nn.Embedding( + config.vocab_size + config.prompt_types * config.prompt_length, config.hidden_size + ) + self.position_bias = CpmAntSegmentPositionEmbedding(config) + self.prompt_length = config.prompt_length + self.vocab_size = config.vocab_size + + self.post_init() + + def get_input_embeddings(self): + return self.input_embedding + + def set_input_embeddings(self, embeddings, **kwargs): + self.input_embedding = embeddings + + def _prepare_attention_mask(self, input_ids, span, context, length): + batch = input_ids.size(0) + seqlen = input_ids.size(1) + device = input_ids.device + directional_mask_2d = torch.arange(seqlen, device=device) <= torch.arange(seqlen, device=device).view(-1, 1) + attention_mask = context[:, None, :] | ( + context[:, :, None].logical_not() & directional_mask_2d.view(1, seqlen, seqlen) + ) + attention_mask = attention_mask & (span[:, None, :] == span[:, :, None]) + # mask for left padding + mask_1d = ( + torch.tensor(list(range(seqlen - self.prompt_length))[::-1], device=device)[None, :].repeat(batch, 1) + < length[:, None] + ) + mask_1d = torch.cat((torch.ones(batch, self.prompt_length, device=device).bool(), mask_1d), dim=1) + attention_mask = mask_1d.view(batch, seqlen, 1) & mask_1d.view(batch, 1, seqlen) & attention_mask + return attention_mask + + @add_start_docstrings_to_model_forward(CPMANT_INPUTS_DOCSTRING) + @add_code_sample_docstrings( + checkpoint=_CHECKPOINT_FOR_DOC, + output_type=BaseModelOutputWithPast, + config_class=_CONFIG_FOR_DOC, + ) + def forward( + self, + input_ids: Optional[torch.Tensor] = None, + output_attentions: Optional[bool] = None, + output_hidden_states: Optional[bool] = None, + past_key_values: Optional[Tuple[Tuple[torch.Tensor]]] = None, + use_cache: Optional[bool] = None, + return_dict: Optional[bool] = None, + **kwargs, + ) -> Union[Tuple[torch.Tensor], BaseModelOutputWithPast]: + output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions + output_hidden_states = ( + output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states + ) + return_dict = return_dict if return_dict is not None else self.config.use_return_dict + use_cache = use_cache if use_cache is not None else self.config.use_cache + + # add prompts ahead + if input_ids.dtype != torch.int32: + input_ids = input_ids.to(torch.int32) + dtype, device = input_ids.dtype, input_ids.device + segment = torch.where(input_ids != 0, 2, 0).to(dtype=dtype, device=device) + length = (segment != 0).sum(-1).to(dtype=dtype, device=device) + input_ids = torch.cat( + ( + torch.arange( + self.prompt_length * 2 + self.vocab_size, + self.prompt_length * 3 + self.vocab_size, + dtype=dtype, + device=device, + ).repeat(input_ids.size(0), 1), + input_ids, + ), + dim=1, + ) + batch, seq_length = input_ids.size() + segment = torch.cat((torch.zeros(batch, self.prompt_length, dtype=dtype, device=device), segment), dim=1) + context = torch.full((batch, seq_length), 1, dtype=dtype, device=device) + position = torch.arange(seq_length, dtype=dtype, device=device).repeat(batch, 1) + span = torch.full((batch, seq_length), 0, dtype=dtype, device=device) + + if past_key_values is None: + past_length = 0 + past_key_values = tuple([None] * self.encoder.num_layers) + input_ids = input_ids.contiguous() + hidden_states = self.input_embedding(input_ids) + segment_states = self.segment_embedding(segment) + hidden_states = hidden_states + segment_states + else: + past_length = past_key_values[0][0].size(-2) + segment_states = self.segment_embedding(segment) + hidden_states = self.input_embedding(input_ids) + segment_states[:, -1:, :] + + attention_mask = self._prepare_attention_mask(input_ids, span, context, length) + position_bias = self.position_bias(position, position, segment, segment) + + attention_mask = attention_mask[:, past_length:, :] + position_bias = position_bias[:, :, past_length:, :] + hidden_states = hidden_states[:, past_length:, :] + + hidden_states, present_key_values, all_hidden_states, all_attentions = self.encoder( + hidden_states, + attention_mask, + position_bias, + output_attentions, + output_hidden_states, + past_key_values, + use_cache, + ) + + if past_length == 0: + hidden_states = hidden_states[:, self.prompt_length :, :] + # drop the prompt + if all_attentions is not None: + new_attentions = () + for attention in all_attentions: + new_attentions += (attention[:, :, self.prompt_length :, self.prompt_length :],) + all_attentions = new_attentions + if all_hidden_states is not None: + new_hidden_states = () + for hidden_state in all_hidden_states: + new_hidden_states += (hidden_state[:, self.prompt_length :, :],) + all_hidden_states = new_hidden_states + + if not return_dict: + return tuple( + v for v in [hidden_states, present_key_values, all_hidden_states, all_attentions] if v is not None + ) + + return BaseModelOutputWithPast( + last_hidden_state=hidden_states, + past_key_values=present_key_values, + hidden_states=all_hidden_states, + attentions=all_attentions, + ) + + +@add_start_docstrings( + """ + The CPMAnt Model with a language modeling head on top (linear layer with weights tied to the input embeddings). + """, + CPMANT_START_DOCSTRING, +) +class CpmAntForCausalLM(CpmAntPreTrainedModel, GenerationMixin): + _tied_weights_keys = ["lm_head.weight"] + + def __init__(self, config: CpmAntConfig): + super().__init__(config) + self.cpmant = CpmAntModel(config) + + # lm_head.weight is tied to cpmant.input_embedding.weight + self.lm_head = nn.Linear( + config.hidden_size, config.vocab_size + config.prompt_types * config.prompt_length, bias=False + ) + self.post_init() + + @add_start_docstrings_to_model_forward(CPMANT_INPUTS_DOCSTRING) + @add_code_sample_docstrings( + checkpoint=_CHECKPOINT_FOR_DOC, + output_type=CausalLMOutputWithPast, + config_class=_CONFIG_FOR_DOC, + ) + def forward( + self, + input_ids: Optional[torch.Tensor] = None, + past_key_values: Optional[List[Tuple[torch.Tensor, torch.Tensor]]] = None, + use_cache: Optional[bool] = None, + output_attentions: Optional[bool] = None, + output_hidden_states: Optional[bool] = None, + labels: Optional[torch.Tensor] = None, + return_dict: Optional[bool] = None, + attention_mask: Optional[torch.Tensor] = None, # dummy parameter for text-generation pipeline + **kwargs, + ) -> Union[Tuple, CausalLMOutputWithPast]: + r""" + Args: + input_ids (`torch.Tensor` of shape `(batch_size, seq_len)`): + Indices of input sequence tokens in the vocabulary. + + Indices can be obtained using [`CPMAntTokenizer`]. See [`PreTrainedTokenizer.encode`] and + [`PreTrainedTokenizer.__call__`] for details. + + [What are input IDs?](../glossary#input-ids) + past_key_values (`tuple(tuple(torch.FloatTensor))`, *optional*, returned when `use_cache=True` is passed or when `config.use_cache=True`): + Contains pre-computed hidden-states (key and values in the self-attention blocks and in the + cross-attention blocks) that can be used (see `past_key_values` input) to speed up sequential decoding. + use_cache (`bool`, *optional*): + If set to `True`, `past_key_values` key value states are returned and can be used to speed up decoding + (see `past_key_values`). + output_attentions (`bool`, *optional*): + Whether or not to return the attentions tensors of all attention layers. + output_hidden_states (`bool`, *optional*): + Whether or not to return the hidden states of all layers. + labels (`torch.Tensor` of shape `(batch_size, sequence_length)`, *optional*): + Labels for computing the masked language modeling loss. + return_dict (`bool`, *optional*): + Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple. + attention_mask (`torch.Tensor` of shape `(batch_size, sequence_length)`, *optional*): + CPMAnt will process attention mask automatically, this parameter is a dummy parameter for + text-generation pipeline. + + Example: + + Text Generation with CpmAntForCausalLM. + ```python + >>> from transformers import CPMAntTokenizer, CpmAntForCausalLM + + >>> texts = "今天天气不错," + >>> model = CpmAntForCausalLM.from_pretrained("openbmb/cpm-ant-10b") + >>> tokenizer = CPMAntTokenizer.from_pretrained("openbmb/cpm-ant-10b") + >>> input_ids = tokenizer(texts, return_tensors="pt") + >>> outputs = model.generate(**input_ids) + >>> output_texts = tokenizer.batch_decode(outputs) + >>> print(output_texts) + ['今天天气不错,阳光明媚,我和妈妈一起去超市买东西。\n在超市里,我看到了一个很好玩的玩具,它的名字叫“机器人”。它有一个圆圆的脑袋,两只圆圆的眼睛,还有一个圆圆的'] + ``` + """ + return_dict = return_dict if return_dict is not None else self.config.use_return_dict + + model_output = self.cpmant( + input_ids, output_attentions, output_hidden_states, past_key_values, use_cache, return_dict + ) + hidden_states = model_output.last_hidden_state if return_dict else model_output[0] + + logits = self.lm_head(hidden_states) + + loss = None + if labels is not None: + loss_func = CrossEntropyLoss() + loss = loss_func(logits.view(-1, logits.size(-1)), labels.view(-1)) + + if not return_dict: + output = (logits,) + model_output[1:] + return ((loss,) + output) if loss is not None else output + + return CausalLMOutputWithPast( + loss=loss, + logits=logits, + past_key_values=model_output.past_key_values, + hidden_states=model_output.hidden_states, + attentions=model_output.attentions, + ) + + def get_input_embeddings(self): + return self.cpmant.input_embedding + + def set_input_embeddings(self, embeddings): + self.cpmant.input_embedding = embeddings + + def get_output_embeddings(self): + return self.lm_head + + def set_output_embeddings(self, new_embeddings): + self.lm_head = new_embeddings + + def _reorder_cache(self, past_key_values, beam_idx): + past_key_values = [list(each) if each is not None else each for each in past_key_values] + for key_value_layer in past_key_values: + key_value_layer[0] = key_value_layer[0][beam_idx] + key_value_layer[1] = key_value_layer[1][beam_idx] + return past_key_values + + +__all__ = ["CpmAntForCausalLM", "CpmAntModel", "CpmAntPreTrainedModel"] diff --git a/vllm/lib/python3.10/site-packages/transformers/models/cpmant/tokenization_cpmant.py b/vllm/lib/python3.10/site-packages/transformers/models/cpmant/tokenization_cpmant.py new file mode 100644 index 0000000000000000000000000000000000000000..2da1d6286c5e8b3a39750ae705cc1486ddd63416 --- /dev/null +++ b/vllm/lib/python3.10/site-packages/transformers/models/cpmant/tokenization_cpmant.py @@ -0,0 +1,270 @@ +# coding=utf-8 +# Copyright 2022 The OpenBMB Team and The HuggingFace Inc. team. All rights reserved. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +"""Tokenization classes for CPMAnt.""" + +import collections +import os +from typing import List, Optional, Tuple + +from transformers.utils import is_jieba_available, requires_backends + + +if is_jieba_available(): + import jieba + +from ...tokenization_utils import PreTrainedTokenizer +from ...utils import logging + + +logger = logging.get_logger(__name__) + +VOCAB_FILES_NAMES = {"vocab_file": "vocab.txt"} + + +def load_vocab(vocab_file): + """Loads a vocabulary file into a dictionary.""" + vocab = collections.OrderedDict() + with open(vocab_file, "r", encoding="utf-8") as reader: + tokens = reader.readlines() + for index, token in enumerate(tokens): + token = token.rstrip("\n") + vocab[token] = index + return vocab + + +class WordpieceTokenizer: + def __init__(self, vocab, unk_token="", max_input_chars_per_word=200): + self.vocab = vocab + self.unk_token = unk_token + self.max_input_chars_per_word = max_input_chars_per_word + + def tokenize(self, token): + chars = list(token) + if len(chars) > self.max_input_chars_per_word: + return [self.unk_token] + + start = 0 + sub_tokens = [] + while start < len(chars): + end = len(chars) + cur_substr = None + while start < end: + substr = "".join(chars[start:end]) + if substr in self.vocab: + cur_substr = substr + break + end -= 1 + if cur_substr is None: + sub_tokens.append(self.unk_token) + start += 1 + else: + sub_tokens.append(cur_substr) + start = end + + return sub_tokens + + +class CpmAntTokenizer(PreTrainedTokenizer): + """ + Construct a CPMAnt tokenizer. Based on byte-level Byte-Pair-Encoding. + + Args: + vocab_file (`str`): + Path to the vocabulary file. + bod_token (`str`, *optional*, defaults to `""`): + The beginning of document token. + eod_token (`str`, *optional*, defaults to `""`): + The end of document token. + bos_token (`str`, *optional*, defaults to `""`): + The beginning of sequence token. + eos_token (`str`, *optional*, defaults to `""`): + The end of sequence token. + pad_token (`str`, *optional*, defaults to `""`): + The token used for padding. + unk_token (`str`, *optional*, defaults to `""`): + The unknown token. + line_token (`str`, *optional*, defaults to `""`): + The line token. + space_token (`str`, *optional*, defaults to `""`): + The space token. + """ + + vocab_files_names = VOCAB_FILES_NAMES + model_input_names = ["input_ids", "attention_mask"] + add_prefix_space = False + + def __init__( + self, + vocab_file, + bod_token="", + eod_token="", + bos_token="", + eos_token="", + pad_token="", + unk_token="", + line_token="", + space_token="", + padding_side="left", + **kwargs, + ): + requires_backends(self, ["jieba"]) + self.bod_token = bod_token + self.eod_token = eod_token + self.encoder = load_vocab(vocab_file) + self.encoder[" "] = self.encoder[space_token] + self.encoder["\n"] = self.encoder[line_token] + + del self.encoder[space_token] + del self.encoder[line_token] + + self.encoder = collections.OrderedDict(sorted(self.encoder.items(), key=lambda x: x[1])) + self.decoder = {v: k for k, v in self.encoder.items()} + + self.wordpiece_tokenizer = WordpieceTokenizer(vocab=self.encoder, unk_token=unk_token) + + super().__init__( + bod_token=bod_token, + eod_token=eod_token, + bos_token=bos_token, + eos_token=eos_token, + pad_token=pad_token, + unk_token=unk_token, + line_token=line_token, + space_token=space_token, + padding_side=padding_side, + **kwargs, + ) + + @property + def bod_token_id(self): + return self.encoder[self.bod_token] + + @property + def eod_token_id(self): + return self.encoder[self.eod_token] + + @property + def newline_id(self): + return self.encoder["\n"] + + @property + def vocab_size(self) -> int: + return len(self.encoder) + + def get_vocab(self): + return dict(self.encoder, **self.added_tokens_encoder) + + def _tokenize(self, text): + """Tokenize a string.""" + output_tokens = [] + for x in jieba.cut(text, cut_all=False): + output_tokens.extend(self.wordpiece_tokenizer.tokenize(x)) + return output_tokens + + def _decode(self, token_ids, **kwargs): + """Decode ids into a string.""" + token_ids = [i for i in token_ids if i >= 0] + token_ids = [ + x for x in token_ids if x != self.pad_token_id and x != self.eos_token_id and x != self.bos_token_id + ] + return super()._decode(token_ids, **kwargs) + + def check(self, token): + return token in self.encoder + + def convert_tokens_to_string(self, tokens: List[str]) -> str: + return "".join(tokens) + + def _convert_token_to_id(self, token): + """Converts a token (str) in an id using the vocab.""" + return self.encoder.get(token, self.encoder.get(self.unk_token)) + + def _convert_id_to_token(self, index): + """Converts an index (integer) in a token (str) using the vocab.""" + return self.decoder.get(index, self.unk_token) + + def save_vocabulary(self, save_directory: str, filename_prefix: Optional[str] = None) -> Tuple[str]: + if os.path.isdir(save_directory): + vocab_file = os.path.join( + save_directory, (filename_prefix + "-" if filename_prefix else "") + VOCAB_FILES_NAMES["vocab_file"] + ) + else: + vocab_file = (filename_prefix + "-" if filename_prefix else "") + save_directory + index = 0 + if " " in self.encoder: + self.encoder[""] = self.encoder[" "] + del self.encoder[" "] + if "\n" in self.encoder: + self.encoder[""] = self.encoder["\n"] + del self.encoder["\n"] + self.encoder = collections.OrderedDict(sorted(self.encoder.items(), key=lambda x: x[1])) + with open(vocab_file, "w", encoding="utf-8") as writer: + for token, token_index in self.encoder.items(): + if index != token_index: + logger.warning( + f"Saving vocabulary to {vocab_file}: vocabulary indices are not consecutive." + " Please check that the vocabulary is not corrupted!" + ) + index = token_index + writer.write(token + "\n") + index += 1 + return (vocab_file,) + + def build_inputs_with_special_tokens(self, token_ids_0: List[int], token_ids_1: List[int] = None) -> List[int]: + """ + Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and + adding special tokens. A CPMAnt sequence has the following format: + + - single sequence: `[BOS] Sequence`. + + Args: + token_ids_0 (`List[int]`): The first tokenized sequence that special tokens will be added. + token_ids_1 (`List[int]`): The optional second tokenized sequence that special tokens will be added. + + Returns: + `List[int]`: The model input with special tokens. + """ + if token_ids_1 is None: + return [self.bos_token_id] + token_ids_0 + return [self.bos_token_id] + token_ids_0 + [self.bos_token_id] + token_ids_1 + + def get_special_tokens_mask( + self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None, already_has_special_tokens: bool = False + ) -> List[int]: + """ + Retrieve sequence ids from a token list that has no special tokens added. This method is called when adding + special tokens using the tokenizer `prepare_for_model` method. + + Args: + token_ids_0 (`List[int]`): List of IDs. + token_ids_1 (`List[int]`, *optional*): Optional second list of IDs for sequence pairs. + already_has_special_tokens (`bool`, *optional*, defaults to `False`): + Whether or not the token list is already formatted with special tokens for the model. + + Returns: + `List[int]`: A list of integers in the range [0, 1]: 1 for a special token, 0 for a sequence token. + """ + + if already_has_special_tokens: + return super().get_special_tokens_mask( + token_ids_0=token_ids_0, token_ids_1=token_ids_1, already_has_special_tokens=True + ) + + if token_ids_1 is not None: + return [1] + ([0] * len(token_ids_0)) + [1] + ([0] * len(token_ids_1)) + return [1] + ([0] * len(token_ids_0)) + + +__all__ = ["CpmAntTokenizer"] diff --git a/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/__init__.py b/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/__init__.py new file mode 100644 index 0000000000000000000000000000000000000000..16a1959c30ff5474ed82a6b0a2e22896514d6a44 --- /dev/null +++ b/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/__init__.py @@ -0,0 +1,32 @@ +# Copyright 2022 The HuggingFace Team. All rights reserved. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + + +from typing import TYPE_CHECKING + +from ...utils import _LazyModule +from ...utils.import_utils import define_import_structure + + +if TYPE_CHECKING: + from .configuration_deformable_detr import * + from .feature_extraction_deformable_detr import * + from .image_processing_deformable_detr import * + from .image_processing_deformable_detr_fast import * + from .modeling_deformable_detr import * +else: + import sys + + _file = globals()["__file__"] + sys.modules[__name__] = _LazyModule(__name__, _file, define_import_structure(_file), module_spec=__spec__) diff --git a/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/__pycache__/__init__.cpython-310.pyc b/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/__pycache__/__init__.cpython-310.pyc new file mode 100644 index 0000000000000000000000000000000000000000..2e575ea7dfc0f6b91e9b6213fd8733a5b4dbb05e Binary files /dev/null and b/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/__pycache__/__init__.cpython-310.pyc differ diff --git a/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/__pycache__/configuration_deformable_detr.cpython-310.pyc b/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/__pycache__/configuration_deformable_detr.cpython-310.pyc new file mode 100644 index 0000000000000000000000000000000000000000..b2347fea5490cadf82cdac67a0160b6469ab06fc Binary files /dev/null and b/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/__pycache__/configuration_deformable_detr.cpython-310.pyc differ diff --git a/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/__pycache__/feature_extraction_deformable_detr.cpython-310.pyc b/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/__pycache__/feature_extraction_deformable_detr.cpython-310.pyc new file mode 100644 index 0000000000000000000000000000000000000000..9c6c5f4623ed15fa61a0062aa218dec16b2a590b Binary files /dev/null and b/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/__pycache__/feature_extraction_deformable_detr.cpython-310.pyc differ diff --git a/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/__pycache__/image_processing_deformable_detr.cpython-310.pyc b/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/__pycache__/image_processing_deformable_detr.cpython-310.pyc new file mode 100644 index 0000000000000000000000000000000000000000..73b90cbcfaa72cc5076811eca565c1a461f6b4f4 Binary files /dev/null and b/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/__pycache__/image_processing_deformable_detr.cpython-310.pyc differ diff --git a/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/__pycache__/image_processing_deformable_detr_fast.cpython-310.pyc b/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/__pycache__/image_processing_deformable_detr_fast.cpython-310.pyc new file mode 100644 index 0000000000000000000000000000000000000000..d636ad94e27369893e25486cbf0199f9fdb3259e Binary files /dev/null and b/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/__pycache__/image_processing_deformable_detr_fast.cpython-310.pyc differ diff --git a/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/__pycache__/modeling_deformable_detr.cpython-310.pyc b/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/__pycache__/modeling_deformable_detr.cpython-310.pyc new file mode 100644 index 0000000000000000000000000000000000000000..11729f3e13ba2f189e019e383902f1c7a7f86aee Binary files /dev/null and b/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/__pycache__/modeling_deformable_detr.cpython-310.pyc differ diff --git a/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/__pycache__/modular_deformable_detr.cpython-310.pyc b/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/__pycache__/modular_deformable_detr.cpython-310.pyc new file mode 100644 index 0000000000000000000000000000000000000000..572e72ceb91c85d77f3b30421e667f8dacd8c766 Binary files /dev/null and b/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/__pycache__/modular_deformable_detr.cpython-310.pyc differ diff --git a/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/configuration_deformable_detr.py b/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/configuration_deformable_detr.py new file mode 100644 index 0000000000000000000000000000000000000000..05bd0f906a110820c36f12fb75b73c1fd0927006 --- /dev/null +++ b/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/configuration_deformable_detr.py @@ -0,0 +1,282 @@ +# coding=utf-8 +# Copyright 2022 SenseTime and The HuggingFace Inc. team. All rights reserved. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +"""Deformable DETR model configuration""" + +from ...configuration_utils import PretrainedConfig +from ...utils import logging +from ...utils.backbone_utils import verify_backbone_config_arguments +from ..auto import CONFIG_MAPPING + + +logger = logging.get_logger(__name__) + + +class DeformableDetrConfig(PretrainedConfig): + r""" + This is the configuration class to store the configuration of a [`DeformableDetrModel`]. It is used to instantiate + a Deformable DETR model according to the specified arguments, defining the model architecture. Instantiating a + configuration with the defaults will yield a similar configuration to that of the Deformable DETR + [SenseTime/deformable-detr](https://huggingface.co/SenseTime/deformable-detr) architecture. + + Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the + documentation from [`PretrainedConfig`] for more information. + + Args: + use_timm_backbone (`bool`, *optional*, defaults to `True`): + Whether or not to use the `timm` library for the backbone. If set to `False`, will use the [`AutoBackbone`] + API. + backbone_config (`PretrainedConfig` or `dict`, *optional*): + The configuration of the backbone model. Only used in case `use_timm_backbone` is set to `False` in which + case it will default to `ResNetConfig()`. + num_channels (`int`, *optional*, defaults to 3): + The number of input channels. + num_queries (`int`, *optional*, defaults to 300): + Number of object queries, i.e. detection slots. This is the maximal number of objects + [`DeformableDetrModel`] can detect in a single image. In case `two_stage` is set to `True`, we use + `two_stage_num_proposals` instead. + d_model (`int`, *optional*, defaults to 256): + Dimension of the layers. + encoder_layers (`int`, *optional*, defaults to 6): + Number of encoder layers. + decoder_layers (`int`, *optional*, defaults to 6): + Number of decoder layers. + encoder_attention_heads (`int`, *optional*, defaults to 8): + Number of attention heads for each attention layer in the Transformer encoder. + decoder_attention_heads (`int`, *optional*, defaults to 8): + Number of attention heads for each attention layer in the Transformer decoder. + decoder_ffn_dim (`int`, *optional*, defaults to 1024): + Dimension of the "intermediate" (often named feed-forward) layer in decoder. + encoder_ffn_dim (`int`, *optional*, defaults to 1024): + Dimension of the "intermediate" (often named feed-forward) layer in decoder. + activation_function (`str` or `function`, *optional*, defaults to `"relu"`): + The non-linear activation function (function or string) in the encoder and pooler. If string, `"gelu"`, + `"relu"`, `"silu"` and `"gelu_new"` are supported. + dropout (`float`, *optional*, defaults to 0.1): + The dropout probability for all fully connected layers in the embeddings, encoder, and pooler. + attention_dropout (`float`, *optional*, defaults to 0.0): + The dropout ratio for the attention probabilities. + activation_dropout (`float`, *optional*, defaults to 0.0): + The dropout ratio for activations inside the fully connected layer. + init_std (`float`, *optional*, defaults to 0.02): + The standard deviation of the truncated_normal_initializer for initializing all weight matrices. + init_xavier_std (`float`, *optional*, defaults to 1): + The scaling factor used for the Xavier initialization gain in the HM Attention map module. + encoder_layerdrop (`float`, *optional*, defaults to 0.0): + The LayerDrop probability for the encoder. See the [LayerDrop paper](see https://arxiv.org/abs/1909.11556) + for more details. + auxiliary_loss (`bool`, *optional*, defaults to `False`): + Whether auxiliary decoding losses (loss at each decoder layer) are to be used. + position_embedding_type (`str`, *optional*, defaults to `"sine"`): + Type of position embeddings to be used on top of the image features. One of `"sine"` or `"learned"`. + backbone (`str`, *optional*, defaults to `"resnet50"`): + Name of backbone to use when `backbone_config` is `None`. If `use_pretrained_backbone` is `True`, this + will load the corresponding pretrained weights from the timm or transformers library. If `use_pretrained_backbone` + is `False`, this loads the backbone's config and uses that to initialize the backbone with random weights. + use_pretrained_backbone (`bool`, *optional*, defaults to `True`): + Whether to use pretrained weights for the backbone. + backbone_kwargs (`dict`, *optional*): + Keyword arguments to be passed to AutoBackbone when loading from a checkpoint + e.g. `{'out_indices': (0, 1, 2, 3)}`. Cannot be specified if `backbone_config` is set. + dilation (`bool`, *optional*, defaults to `False`): + Whether to replace stride with dilation in the last convolutional block (DC5). Only supported when + `use_timm_backbone` = `True`. + class_cost (`float`, *optional*, defaults to 1): + Relative weight of the classification error in the Hungarian matching cost. + bbox_cost (`float`, *optional*, defaults to 5): + Relative weight of the L1 error of the bounding box coordinates in the Hungarian matching cost. + giou_cost (`float`, *optional*, defaults to 2): + Relative weight of the generalized IoU loss of the bounding box in the Hungarian matching cost. + mask_loss_coefficient (`float`, *optional*, defaults to 1): + Relative weight of the Focal loss in the panoptic segmentation loss. + dice_loss_coefficient (`float`, *optional*, defaults to 1): + Relative weight of the DICE/F-1 loss in the panoptic segmentation loss. + bbox_loss_coefficient (`float`, *optional*, defaults to 5): + Relative weight of the L1 bounding box loss in the object detection loss. + giou_loss_coefficient (`float`, *optional*, defaults to 2): + Relative weight of the generalized IoU loss in the object detection loss. + eos_coefficient (`float`, *optional*, defaults to 0.1): + Relative classification weight of the 'no-object' class in the object detection loss. + num_feature_levels (`int`, *optional*, defaults to 4): + The number of input feature levels. + encoder_n_points (`int`, *optional*, defaults to 4): + The number of sampled keys in each feature level for each attention head in the encoder. + decoder_n_points (`int`, *optional*, defaults to 4): + The number of sampled keys in each feature level for each attention head in the decoder. + two_stage (`bool`, *optional*, defaults to `False`): + Whether to apply a two-stage deformable DETR, where the region proposals are also generated by a variant of + Deformable DETR, which are further fed into the decoder for iterative bounding box refinement. + two_stage_num_proposals (`int`, *optional*, defaults to 300): + The number of region proposals to be generated, in case `two_stage` is set to `True`. + with_box_refine (`bool`, *optional*, defaults to `False`): + Whether to apply iterative bounding box refinement, where each decoder layer refines the bounding boxes + based on the predictions from the previous layer. + focal_alpha (`float`, *optional*, defaults to 0.25): + Alpha parameter in the focal loss. + disable_custom_kernels (`bool`, *optional*, defaults to `False`): + Disable the use of custom CUDA and CPU kernels. This option is necessary for the ONNX export, as custom + kernels are not supported by PyTorch ONNX export. + + Examples: + + ```python + >>> from transformers import DeformableDetrConfig, DeformableDetrModel + + >>> # Initializing a Deformable DETR SenseTime/deformable-detr style configuration + >>> configuration = DeformableDetrConfig() + + >>> # Initializing a model (with random weights) from the SenseTime/deformable-detr style configuration + >>> model = DeformableDetrModel(configuration) + + >>> # Accessing the model configuration + >>> configuration = model.config + ```""" + + model_type = "deformable_detr" + attribute_map = { + "hidden_size": "d_model", + "num_attention_heads": "encoder_attention_heads", + } + + def __init__( + self, + use_timm_backbone=True, + backbone_config=None, + num_channels=3, + num_queries=300, + max_position_embeddings=1024, + encoder_layers=6, + encoder_ffn_dim=1024, + encoder_attention_heads=8, + decoder_layers=6, + decoder_ffn_dim=1024, + decoder_attention_heads=8, + encoder_layerdrop=0.0, + is_encoder_decoder=True, + activation_function="relu", + d_model=256, + dropout=0.1, + attention_dropout=0.0, + activation_dropout=0.0, + init_std=0.02, + init_xavier_std=1.0, + return_intermediate=True, + auxiliary_loss=False, + position_embedding_type="sine", + backbone="resnet50", + use_pretrained_backbone=True, + backbone_kwargs=None, + dilation=False, + num_feature_levels=4, + encoder_n_points=4, + decoder_n_points=4, + two_stage=False, + two_stage_num_proposals=300, + with_box_refine=False, + class_cost=1, + bbox_cost=5, + giou_cost=2, + mask_loss_coefficient=1, + dice_loss_coefficient=1, + bbox_loss_coefficient=5, + giou_loss_coefficient=2, + eos_coefficient=0.1, + focal_alpha=0.25, + disable_custom_kernels=False, + **kwargs, + ): + # We default to values which were previously hard-coded in the model. This enables configurability of the config + # while keeping the default behavior the same. + if use_timm_backbone and backbone_kwargs is None: + backbone_kwargs = {} + if dilation: + backbone_kwargs["output_stride"] = 16 + backbone_kwargs["out_indices"] = [2, 3, 4] if num_feature_levels > 1 else [4] + backbone_kwargs["in_chans"] = num_channels + # Backwards compatibility + elif not use_timm_backbone and backbone in (None, "resnet50"): + if backbone_config is None: + logger.info("`backbone_config` is `None`. Initializing the config with the default `ResNet` backbone.") + backbone_config = CONFIG_MAPPING["resnet"](out_features=["stage4"]) + elif isinstance(backbone_config, dict): + backbone_model_type = backbone_config.get("model_type") + config_class = CONFIG_MAPPING[backbone_model_type] + backbone_config = config_class.from_dict(backbone_config) + + verify_backbone_config_arguments( + use_timm_backbone=use_timm_backbone, + use_pretrained_backbone=use_pretrained_backbone, + backbone=backbone, + backbone_config=backbone_config, + backbone_kwargs=backbone_kwargs, + ) + + self.use_timm_backbone = use_timm_backbone + self.backbone_config = backbone_config + self.num_channels = num_channels + self.num_queries = num_queries + self.max_position_embeddings = max_position_embeddings + self.d_model = d_model + self.encoder_ffn_dim = encoder_ffn_dim + self.encoder_layers = encoder_layers + self.encoder_attention_heads = encoder_attention_heads + self.decoder_ffn_dim = decoder_ffn_dim + self.decoder_layers = decoder_layers + self.decoder_attention_heads = decoder_attention_heads + self.dropout = dropout + self.attention_dropout = attention_dropout + self.activation_dropout = activation_dropout + self.activation_function = activation_function + self.init_std = init_std + self.init_xavier_std = init_xavier_std + self.encoder_layerdrop = encoder_layerdrop + self.auxiliary_loss = auxiliary_loss + self.position_embedding_type = position_embedding_type + self.backbone = backbone + self.use_pretrained_backbone = use_pretrained_backbone + self.backbone_kwargs = backbone_kwargs + self.dilation = dilation + # deformable attributes + self.num_feature_levels = num_feature_levels + self.encoder_n_points = encoder_n_points + self.decoder_n_points = decoder_n_points + self.two_stage = two_stage + self.two_stage_num_proposals = two_stage_num_proposals + self.with_box_refine = with_box_refine + if two_stage is True and with_box_refine is False: + raise ValueError("If two_stage is True, with_box_refine must be True.") + # Hungarian matcher + self.class_cost = class_cost + self.bbox_cost = bbox_cost + self.giou_cost = giou_cost + # Loss coefficients + self.mask_loss_coefficient = mask_loss_coefficient + self.dice_loss_coefficient = dice_loss_coefficient + self.bbox_loss_coefficient = bbox_loss_coefficient + self.giou_loss_coefficient = giou_loss_coefficient + self.eos_coefficient = eos_coefficient + self.focal_alpha = focal_alpha + self.disable_custom_kernels = disable_custom_kernels + super().__init__(is_encoder_decoder=is_encoder_decoder, **kwargs) + + @property + def num_attention_heads(self) -> int: + return self.encoder_attention_heads + + @property + def hidden_size(self) -> int: + return self.d_model + + +__all__ = ["DeformableDetrConfig"] diff --git a/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/feature_extraction_deformable_detr.py b/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/feature_extraction_deformable_detr.py new file mode 100644 index 0000000000000000000000000000000000000000..de1676740647f65dddee160ea88b2026bb6551a9 --- /dev/null +++ b/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/feature_extraction_deformable_detr.py @@ -0,0 +1,46 @@ +# coding=utf-8 +# Copyright 2022 The HuggingFace Inc. team. All rights reserved. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +"""Feature extractor class for Deformable DETR.""" + +import warnings + +from ...image_transforms import rgb_to_id as _rgb_to_id +from ...utils import logging +from .image_processing_deformable_detr import DeformableDetrImageProcessor + + +logger = logging.get_logger(__name__) + + +def rgb_to_id(x): + warnings.warn( + "rgb_to_id has moved and will not be importable from this module from v5. " + "Please import from transformers.image_transforms instead.", + FutureWarning, + ) + return _rgb_to_id(x) + + +class DeformableDetrFeatureExtractor(DeformableDetrImageProcessor): + def __init__(self, *args, **kwargs) -> None: + warnings.warn( + "The class DeformableDetrFeatureExtractor is deprecated and will be removed in version 5 of Transformers." + " Please use DeformableDetrImageProcessor instead.", + FutureWarning, + ) + super().__init__(*args, **kwargs) + + +__all__ = ["DeformableDetrFeatureExtractor"] diff --git a/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/image_processing_deformable_detr.py b/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/image_processing_deformable_detr.py new file mode 100644 index 0000000000000000000000000000000000000000..e44733be812e5d7409c8fba5039912ebd977bcff --- /dev/null +++ b/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/image_processing_deformable_detr.py @@ -0,0 +1,1631 @@ +# coding=utf-8 +# Copyright 2022 The HuggingFace Inc. team. All rights reserved. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +"""Image processor class for Deformable DETR.""" + +import io +import pathlib +from collections import defaultdict +from typing import Any, Callable, Dict, Iterable, List, Optional, Set, Tuple, Union + +import numpy as np + +from ...feature_extraction_utils import BatchFeature +from ...image_processing_utils import BaseImageProcessor, get_size_dict +from ...image_transforms import ( + PaddingMode, + center_to_corners_format, + corners_to_center_format, + id_to_rgb, + pad, + rescale, + resize, + rgb_to_id, + to_channel_dimension_format, +) +from ...image_utils import ( + IMAGENET_DEFAULT_MEAN, + IMAGENET_DEFAULT_STD, + AnnotationFormat, + AnnotationType, + ChannelDimension, + ImageInput, + PILImageResampling, + get_image_size, + infer_channel_dimension_format, + is_scaled_image, + make_list_of_images, + to_numpy_array, + valid_images, + validate_annotations, + validate_kwargs, + validate_preprocess_arguments, +) +from ...utils import ( + TensorType, + is_flax_available, + is_jax_tensor, + is_scipy_available, + is_tf_available, + is_tf_tensor, + is_torch_available, + is_torch_tensor, + is_vision_available, + logging, +) + + +if is_torch_available(): + import torch + from torch import nn + + +if is_vision_available(): + import PIL + +if is_scipy_available(): + import scipy.special + import scipy.stats + + +logger = logging.get_logger(__name__) # pylint: disable=invalid-name + +SUPPORTED_ANNOTATION_FORMATS = (AnnotationFormat.COCO_DETECTION, AnnotationFormat.COCO_PANOPTIC) + + +# Copied from transformers.models.detr.image_processing_detr.get_size_with_aspect_ratio +def get_size_with_aspect_ratio(image_size, size, max_size=None) -> Tuple[int, int]: + """ + Computes the output image size given the input image size and the desired output size. + + Args: + image_size (`Tuple[int, int]`): + The input image size. + size (`int`): + The desired output size. + max_size (`int`, *optional*): + The maximum allowed output size. + """ + height, width = image_size + raw_size = None + if max_size is not None: + min_original_size = float(min((height, width))) + max_original_size = float(max((height, width))) + if max_original_size / min_original_size * size > max_size: + raw_size = max_size * min_original_size / max_original_size + size = int(round(raw_size)) + + if (height <= width and height == size) or (width <= height and width == size): + oh, ow = height, width + elif width < height: + ow = size + if max_size is not None and raw_size is not None: + oh = int(raw_size * height / width) + else: + oh = int(size * height / width) + else: + oh = size + if max_size is not None and raw_size is not None: + ow = int(raw_size * width / height) + else: + ow = int(size * width / height) + + return (oh, ow) + + +# Copied from transformers.models.detr.image_processing_detr.get_resize_output_image_size +def get_resize_output_image_size( + input_image: np.ndarray, + size: Union[int, Tuple[int, int], List[int]], + max_size: Optional[int] = None, + input_data_format: Optional[Union[str, ChannelDimension]] = None, +) -> Tuple[int, int]: + """ + Computes the output image size given the input image size and the desired output size. If the desired output size + is a tuple or list, the output image size is returned as is. If the desired output size is an integer, the output + image size is computed by keeping the aspect ratio of the input image size. + + Args: + input_image (`np.ndarray`): + The image to resize. + size (`int` or `Tuple[int, int]` or `List[int]`): + The desired output size. + max_size (`int`, *optional*): + The maximum allowed output size. + input_data_format (`ChannelDimension` or `str`, *optional*): + The channel dimension format of the input image. If not provided, it will be inferred from the input image. + """ + image_size = get_image_size(input_image, input_data_format) + if isinstance(size, (list, tuple)): + return size + + return get_size_with_aspect_ratio(image_size, size, max_size) + + +# Copied from transformers.models.detr.image_processing_detr.get_image_size_for_max_height_width +def get_image_size_for_max_height_width( + input_image: np.ndarray, + max_height: int, + max_width: int, + input_data_format: Optional[Union[str, ChannelDimension]] = None, +) -> Tuple[int, int]: + """ + Computes the output image size given the input image and the maximum allowed height and width. Keep aspect ratio. + Important, even if image_height < max_height and image_width < max_width, the image will be resized + to at least one of the edges be equal to max_height or max_width. + + For example: + - input_size: (100, 200), max_height: 50, max_width: 50 -> output_size: (25, 50) + - input_size: (100, 200), max_height: 200, max_width: 500 -> output_size: (200, 400) + + Args: + input_image (`np.ndarray`): + The image to resize. + max_height (`int`): + The maximum allowed height. + max_width (`int`): + The maximum allowed width. + input_data_format (`ChannelDimension` or `str`, *optional*): + The channel dimension format of the input image. If not provided, it will be inferred from the input image. + """ + image_size = get_image_size(input_image, input_data_format) + height, width = image_size + height_scale = max_height / height + width_scale = max_width / width + min_scale = min(height_scale, width_scale) + new_height = int(height * min_scale) + new_width = int(width * min_scale) + return new_height, new_width + + +# Copied from transformers.models.detr.image_processing_detr.get_numpy_to_framework_fn +def get_numpy_to_framework_fn(arr) -> Callable: + """ + Returns a function that converts a numpy array to the framework of the input array. + + Args: + arr (`np.ndarray`): The array to convert. + """ + if isinstance(arr, np.ndarray): + return np.array + if is_tf_available() and is_tf_tensor(arr): + import tensorflow as tf + + return tf.convert_to_tensor + if is_torch_available() and is_torch_tensor(arr): + import torch + + return torch.tensor + if is_flax_available() and is_jax_tensor(arr): + import jax.numpy as jnp + + return jnp.array + raise ValueError(f"Cannot convert arrays of type {type(arr)}") + + +# Copied from transformers.models.detr.image_processing_detr.safe_squeeze +def safe_squeeze(arr: np.ndarray, axis: Optional[int] = None) -> np.ndarray: + """ + Squeezes an array, but only if the axis specified has dim 1. + """ + if axis is None: + return arr.squeeze() + + try: + return arr.squeeze(axis=axis) + except ValueError: + return arr + + +# Copied from transformers.models.detr.image_processing_detr.normalize_annotation +def normalize_annotation(annotation: Dict, image_size: Tuple[int, int]) -> Dict: + image_height, image_width = image_size + norm_annotation = {} + for key, value in annotation.items(): + if key == "boxes": + boxes = value + boxes = corners_to_center_format(boxes) + boxes /= np.asarray([image_width, image_height, image_width, image_height], dtype=np.float32) + norm_annotation[key] = boxes + else: + norm_annotation[key] = value + return norm_annotation + + +# Copied from transformers.models.detr.image_processing_detr.max_across_indices +def max_across_indices(values: Iterable[Any]) -> List[Any]: + """ + Return the maximum value across all indices of an iterable of values. + """ + return [max(values_i) for values_i in zip(*values)] + + +# Copied from transformers.models.detr.image_processing_detr.get_max_height_width +def get_max_height_width( + images: List[np.ndarray], input_data_format: Optional[Union[str, ChannelDimension]] = None +) -> List[int]: + """ + Get the maximum height and width across all images in a batch. + """ + if input_data_format is None: + input_data_format = infer_channel_dimension_format(images[0]) + + if input_data_format == ChannelDimension.FIRST: + _, max_height, max_width = max_across_indices([img.shape for img in images]) + elif input_data_format == ChannelDimension.LAST: + max_height, max_width, _ = max_across_indices([img.shape for img in images]) + else: + raise ValueError(f"Invalid channel dimension format: {input_data_format}") + return (max_height, max_width) + + +# Copied from transformers.models.detr.image_processing_detr.make_pixel_mask +def make_pixel_mask( + image: np.ndarray, output_size: Tuple[int, int], input_data_format: Optional[Union[str, ChannelDimension]] = None +) -> np.ndarray: + """ + Make a pixel mask for the image, where 1 indicates a valid pixel and 0 indicates padding. + + Args: + image (`np.ndarray`): + Image to make the pixel mask for. + output_size (`Tuple[int, int]`): + Output size of the mask. + """ + input_height, input_width = get_image_size(image, channel_dim=input_data_format) + mask = np.zeros(output_size, dtype=np.int64) + mask[:input_height, :input_width] = 1 + return mask + + +# Copied from transformers.models.detr.image_processing_detr.convert_coco_poly_to_mask +def convert_coco_poly_to_mask(segmentations, height: int, width: int) -> np.ndarray: + """ + Convert a COCO polygon annotation to a mask. + + Args: + segmentations (`List[List[float]]`): + List of polygons, each polygon represented by a list of x-y coordinates. + height (`int`): + Height of the mask. + width (`int`): + Width of the mask. + """ + try: + from pycocotools import mask as coco_mask + except ImportError: + raise ImportError("Pycocotools is not installed in your environment.") + + masks = [] + for polygons in segmentations: + rles = coco_mask.frPyObjects(polygons, height, width) + mask = coco_mask.decode(rles) + if len(mask.shape) < 3: + mask = mask[..., None] + mask = np.asarray(mask, dtype=np.uint8) + mask = np.any(mask, axis=2) + masks.append(mask) + if masks: + masks = np.stack(masks, axis=0) + else: + masks = np.zeros((0, height, width), dtype=np.uint8) + + return masks + + +# Copied from transformers.models.detr.image_processing_detr.prepare_coco_detection_annotation with DETR->DeformableDetr +def prepare_coco_detection_annotation( + image, + target, + return_segmentation_masks: bool = False, + input_data_format: Optional[Union[ChannelDimension, str]] = None, +): + """ + Convert the target in COCO format into the format expected by DeformableDetr. + """ + image_height, image_width = get_image_size(image, channel_dim=input_data_format) + + image_id = target["image_id"] + image_id = np.asarray([image_id], dtype=np.int64) + + # Get all COCO annotations for the given image. + annotations = target["annotations"] + annotations = [obj for obj in annotations if "iscrowd" not in obj or obj["iscrowd"] == 0] + + classes = [obj["category_id"] for obj in annotations] + classes = np.asarray(classes, dtype=np.int64) + + # for conversion to coco api + area = np.asarray([obj["area"] for obj in annotations], dtype=np.float32) + iscrowd = np.asarray([obj["iscrowd"] if "iscrowd" in obj else 0 for obj in annotations], dtype=np.int64) + + boxes = [obj["bbox"] for obj in annotations] + # guard against no boxes via resizing + boxes = np.asarray(boxes, dtype=np.float32).reshape(-1, 4) + boxes[:, 2:] += boxes[:, :2] + boxes[:, 0::2] = boxes[:, 0::2].clip(min=0, max=image_width) + boxes[:, 1::2] = boxes[:, 1::2].clip(min=0, max=image_height) + + keep = (boxes[:, 3] > boxes[:, 1]) & (boxes[:, 2] > boxes[:, 0]) + + new_target = {} + new_target["image_id"] = image_id + new_target["class_labels"] = classes[keep] + new_target["boxes"] = boxes[keep] + new_target["area"] = area[keep] + new_target["iscrowd"] = iscrowd[keep] + new_target["orig_size"] = np.asarray([int(image_height), int(image_width)], dtype=np.int64) + + if annotations and "keypoints" in annotations[0]: + keypoints = [obj["keypoints"] for obj in annotations] + # Converting the filtered keypoints list to a numpy array + keypoints = np.asarray(keypoints, dtype=np.float32) + # Apply the keep mask here to filter the relevant annotations + keypoints = keypoints[keep] + num_keypoints = keypoints.shape[0] + keypoints = keypoints.reshape((-1, 3)) if num_keypoints else keypoints + new_target["keypoints"] = keypoints + + if return_segmentation_masks: + segmentation_masks = [obj["segmentation"] for obj in annotations] + masks = convert_coco_poly_to_mask(segmentation_masks, image_height, image_width) + new_target["masks"] = masks[keep] + + return new_target + + +# Copied from transformers.models.detr.image_processing_detr.masks_to_boxes +def masks_to_boxes(masks: np.ndarray) -> np.ndarray: + """ + Compute the bounding boxes around the provided panoptic segmentation masks. + + Args: + masks: masks in format `[number_masks, height, width]` where N is the number of masks + + Returns: + boxes: bounding boxes in format `[number_masks, 4]` in xyxy format + """ + if masks.size == 0: + return np.zeros((0, 4)) + + h, w = masks.shape[-2:] + y = np.arange(0, h, dtype=np.float32) + x = np.arange(0, w, dtype=np.float32) + # see https://github.com/pytorch/pytorch/issues/50276 + y, x = np.meshgrid(y, x, indexing="ij") + + x_mask = masks * np.expand_dims(x, axis=0) + x_max = x_mask.reshape(x_mask.shape[0], -1).max(-1) + x = np.ma.array(x_mask, mask=~(np.array(masks, dtype=bool))) + x_min = x.filled(fill_value=1e8) + x_min = x_min.reshape(x_min.shape[0], -1).min(-1) + + y_mask = masks * np.expand_dims(y, axis=0) + y_max = y_mask.reshape(x_mask.shape[0], -1).max(-1) + y = np.ma.array(y_mask, mask=~(np.array(masks, dtype=bool))) + y_min = y.filled(fill_value=1e8) + y_min = y_min.reshape(y_min.shape[0], -1).min(-1) + + return np.stack([x_min, y_min, x_max, y_max], 1) + + +# Copied from transformers.models.detr.image_processing_detr.prepare_coco_panoptic_annotation with DETR->DeformableDetr +def prepare_coco_panoptic_annotation( + image: np.ndarray, + target: Dict, + masks_path: Union[str, pathlib.Path], + return_masks: bool = True, + input_data_format: Union[ChannelDimension, str] = None, +) -> Dict: + """ + Prepare a coco panoptic annotation for DeformableDetr. + """ + image_height, image_width = get_image_size(image, channel_dim=input_data_format) + annotation_path = pathlib.Path(masks_path) / target["file_name"] + + new_target = {} + new_target["image_id"] = np.asarray([target["image_id"] if "image_id" in target else target["id"]], dtype=np.int64) + new_target["size"] = np.asarray([image_height, image_width], dtype=np.int64) + new_target["orig_size"] = np.asarray([image_height, image_width], dtype=np.int64) + + if "segments_info" in target: + masks = np.asarray(PIL.Image.open(annotation_path), dtype=np.uint32) + masks = rgb_to_id(masks) + + ids = np.array([segment_info["id"] for segment_info in target["segments_info"]]) + masks = masks == ids[:, None, None] + masks = masks.astype(np.uint8) + if return_masks: + new_target["masks"] = masks + new_target["boxes"] = masks_to_boxes(masks) + new_target["class_labels"] = np.array( + [segment_info["category_id"] for segment_info in target["segments_info"]], dtype=np.int64 + ) + new_target["iscrowd"] = np.asarray( + [segment_info["iscrowd"] for segment_info in target["segments_info"]], dtype=np.int64 + ) + new_target["area"] = np.asarray( + [segment_info["area"] for segment_info in target["segments_info"]], dtype=np.float32 + ) + + return new_target + + +# Copied from transformers.models.detr.image_processing_detr.get_segmentation_image +def get_segmentation_image( + masks: np.ndarray, input_size: Tuple, target_size: Tuple, stuff_equiv_classes, deduplicate=False +): + h, w = input_size + final_h, final_w = target_size + + m_id = scipy.special.softmax(masks.transpose(0, 1), -1) + + if m_id.shape[-1] == 0: + # We didn't detect any mask :( + m_id = np.zeros((h, w), dtype=np.int64) + else: + m_id = m_id.argmax(-1).reshape(h, w) + + if deduplicate: + # Merge the masks corresponding to the same stuff class + for equiv in stuff_equiv_classes.values(): + for eq_id in equiv: + m_id[m_id == eq_id] = equiv[0] + + seg_img = id_to_rgb(m_id) + seg_img = resize(seg_img, (final_w, final_h), resample=PILImageResampling.NEAREST) + return seg_img + + +# Copied from transformers.models.detr.image_processing_detr.get_mask_area +def get_mask_area(seg_img: np.ndarray, target_size: Tuple[int, int], n_classes: int) -> np.ndarray: + final_h, final_w = target_size + np_seg_img = seg_img.astype(np.uint8) + np_seg_img = np_seg_img.reshape(final_h, final_w, 3) + m_id = rgb_to_id(np_seg_img) + area = [(m_id == i).sum() for i in range(n_classes)] + return area + + +# Copied from transformers.models.detr.image_processing_detr.score_labels_from_class_probabilities +def score_labels_from_class_probabilities(logits: np.ndarray) -> Tuple[np.ndarray, np.ndarray]: + probs = scipy.special.softmax(logits, axis=-1) + labels = probs.argmax(-1, keepdims=True) + scores = np.take_along_axis(probs, labels, axis=-1) + scores, labels = scores.squeeze(-1), labels.squeeze(-1) + return scores, labels + + +# Copied from transformers.models.detr.image_processing_detr.post_process_panoptic_sample +def post_process_panoptic_sample( + out_logits: np.ndarray, + masks: np.ndarray, + boxes: np.ndarray, + processed_size: Tuple[int, int], + target_size: Tuple[int, int], + is_thing_map: Dict, + threshold=0.85, +) -> Dict: + """ + Converts the output of [`DetrForSegmentation`] into panoptic segmentation predictions for a single sample. + + Args: + out_logits (`torch.Tensor`): + The logits for this sample. + masks (`torch.Tensor`): + The predicted segmentation masks for this sample. + boxes (`torch.Tensor`): + The prediced bounding boxes for this sample. The boxes are in the normalized format `(center_x, center_y, + width, height)` and values between `[0, 1]`, relative to the size the image (disregarding padding). + processed_size (`Tuple[int, int]`): + The processed size of the image `(height, width)`, as returned by the preprocessing step i.e. the size + after data augmentation but before batching. + target_size (`Tuple[int, int]`): + The target size of the image, `(height, width)` corresponding to the requested final size of the + prediction. + is_thing_map (`Dict`): + A dictionary mapping class indices to a boolean value indicating whether the class is a thing or not. + threshold (`float`, *optional*, defaults to 0.85): + The threshold used to binarize the segmentation masks. + """ + # we filter empty queries and detection below threshold + scores, labels = score_labels_from_class_probabilities(out_logits) + keep = (labels != out_logits.shape[-1] - 1) & (scores > threshold) + + cur_scores = scores[keep] + cur_classes = labels[keep] + cur_boxes = center_to_corners_format(boxes[keep]) + + if len(cur_boxes) != len(cur_classes): + raise ValueError("Not as many boxes as there are classes") + + cur_masks = masks[keep] + cur_masks = resize(cur_masks[:, None], processed_size, resample=PILImageResampling.BILINEAR) + cur_masks = safe_squeeze(cur_masks, 1) + b, h, w = cur_masks.shape + + # It may be that we have several predicted masks for the same stuff class. + # In the following, we track the list of masks ids for each stuff class (they are merged later on) + cur_masks = cur_masks.reshape(b, -1) + stuff_equiv_classes = defaultdict(list) + for k, label in enumerate(cur_classes): + if not is_thing_map[label]: + stuff_equiv_classes[label].append(k) + + seg_img = get_segmentation_image(cur_masks, processed_size, target_size, stuff_equiv_classes, deduplicate=True) + area = get_mask_area(cur_masks, processed_size, n_classes=len(cur_scores)) + + # We filter out any mask that is too small + if cur_classes.size() > 0: + # We know filter empty masks as long as we find some + filtered_small = np.array([a <= 4 for a in area], dtype=bool) + while filtered_small.any(): + cur_masks = cur_masks[~filtered_small] + cur_scores = cur_scores[~filtered_small] + cur_classes = cur_classes[~filtered_small] + seg_img = get_segmentation_image(cur_masks, (h, w), target_size, stuff_equiv_classes, deduplicate=True) + area = get_mask_area(seg_img, target_size, n_classes=len(cur_scores)) + filtered_small = np.array([a <= 4 for a in area], dtype=bool) + else: + cur_classes = np.ones((1, 1), dtype=np.int64) + + segments_info = [ + {"id": i, "isthing": is_thing_map[cat], "category_id": int(cat), "area": a} + for i, (cat, a) in enumerate(zip(cur_classes, area)) + ] + del cur_classes + + with io.BytesIO() as out: + PIL.Image.fromarray(seg_img).save(out, format="PNG") + predictions = {"png_string": out.getvalue(), "segments_info": segments_info} + + return predictions + + +# Copied from transformers.models.detr.image_processing_detr.resize_annotation +def resize_annotation( + annotation: Dict[str, Any], + orig_size: Tuple[int, int], + target_size: Tuple[int, int], + threshold: float = 0.5, + resample: PILImageResampling = PILImageResampling.NEAREST, +): + """ + Resizes an annotation to a target size. + + Args: + annotation (`Dict[str, Any]`): + The annotation dictionary. + orig_size (`Tuple[int, int]`): + The original size of the input image. + target_size (`Tuple[int, int]`): + The target size of the image, as returned by the preprocessing `resize` step. + threshold (`float`, *optional*, defaults to 0.5): + The threshold used to binarize the segmentation masks. + resample (`PILImageResampling`, defaults to `PILImageResampling.NEAREST`): + The resampling filter to use when resizing the masks. + """ + ratios = tuple(float(s) / float(s_orig) for s, s_orig in zip(target_size, orig_size)) + ratio_height, ratio_width = ratios + + new_annotation = {} + new_annotation["size"] = target_size + + for key, value in annotation.items(): + if key == "boxes": + boxes = value + scaled_boxes = boxes * np.asarray([ratio_width, ratio_height, ratio_width, ratio_height], dtype=np.float32) + new_annotation["boxes"] = scaled_boxes + elif key == "area": + area = value + scaled_area = area * (ratio_width * ratio_height) + new_annotation["area"] = scaled_area + elif key == "masks": + masks = value[:, None] + masks = np.array([resize(mask, target_size, resample=resample) for mask in masks]) + masks = masks.astype(np.float32) + masks = masks[:, 0] > threshold + new_annotation["masks"] = masks + elif key == "size": + new_annotation["size"] = target_size + else: + new_annotation[key] = value + + return new_annotation + + +# Copied from transformers.models.detr.image_processing_detr.binary_mask_to_rle +def binary_mask_to_rle(mask): + """ + Converts given binary mask of shape `(height, width)` to the run-length encoding (RLE) format. + + Args: + mask (`torch.Tensor` or `numpy.array`): + A binary mask tensor of shape `(height, width)` where 0 denotes background and 1 denotes the target + segment_id or class_id. + Returns: + `List`: Run-length encoded list of the binary mask. Refer to COCO API for more information about the RLE + format. + """ + if is_torch_tensor(mask): + mask = mask.numpy() + + pixels = mask.flatten() + pixels = np.concatenate([[0], pixels, [0]]) + runs = np.where(pixels[1:] != pixels[:-1])[0] + 1 + runs[1::2] -= runs[::2] + return list(runs) + + +# Copied from transformers.models.detr.image_processing_detr.convert_segmentation_to_rle +def convert_segmentation_to_rle(segmentation): + """ + Converts given segmentation map of shape `(height, width)` to the run-length encoding (RLE) format. + + Args: + segmentation (`torch.Tensor` or `numpy.array`): + A segmentation map of shape `(height, width)` where each value denotes a segment or class id. + Returns: + `List[List]`: A list of lists, where each list is the run-length encoding of a segment / class id. + """ + segment_ids = torch.unique(segmentation) + + run_length_encodings = [] + for idx in segment_ids: + mask = torch.where(segmentation == idx, 1, 0) + rle = binary_mask_to_rle(mask) + run_length_encodings.append(rle) + + return run_length_encodings + + +# Copied from transformers.models.detr.image_processing_detr.remove_low_and_no_objects +def remove_low_and_no_objects(masks, scores, labels, object_mask_threshold, num_labels): + """ + Binarize the given masks using `object_mask_threshold`, it returns the associated values of `masks`, `scores` and + `labels`. + + Args: + masks (`torch.Tensor`): + A tensor of shape `(num_queries, height, width)`. + scores (`torch.Tensor`): + A tensor of shape `(num_queries)`. + labels (`torch.Tensor`): + A tensor of shape `(num_queries)`. + object_mask_threshold (`float`): + A number between 0 and 1 used to binarize the masks. + Raises: + `ValueError`: Raised when the first dimension doesn't match in all input tensors. + Returns: + `Tuple[`torch.Tensor`, `torch.Tensor`, `torch.Tensor`]`: The `masks`, `scores` and `labels` without the region + < `object_mask_threshold`. + """ + if not (masks.shape[0] == scores.shape[0] == labels.shape[0]): + raise ValueError("mask, scores and labels must have the same shape!") + + to_keep = labels.ne(num_labels) & (scores > object_mask_threshold) + + return masks[to_keep], scores[to_keep], labels[to_keep] + + +# Copied from transformers.models.detr.image_processing_detr.check_segment_validity +def check_segment_validity(mask_labels, mask_probs, k, mask_threshold=0.5, overlap_mask_area_threshold=0.8): + # Get the mask associated with the k class + mask_k = mask_labels == k + mask_k_area = mask_k.sum() + + # Compute the area of all the stuff in query k + original_area = (mask_probs[k] >= mask_threshold).sum() + mask_exists = mask_k_area > 0 and original_area > 0 + + # Eliminate disconnected tiny segments + if mask_exists: + area_ratio = mask_k_area / original_area + if not area_ratio.item() > overlap_mask_area_threshold: + mask_exists = False + + return mask_exists, mask_k + + +# Copied from transformers.models.detr.image_processing_detr.compute_segments +def compute_segments( + mask_probs, + pred_scores, + pred_labels, + mask_threshold: float = 0.5, + overlap_mask_area_threshold: float = 0.8, + label_ids_to_fuse: Optional[Set[int]] = None, + target_size: Tuple[int, int] = None, +): + height = mask_probs.shape[1] if target_size is None else target_size[0] + width = mask_probs.shape[2] if target_size is None else target_size[1] + + segmentation = torch.zeros((height, width), dtype=torch.int32, device=mask_probs.device) + segments: List[Dict] = [] + + if target_size is not None: + mask_probs = nn.functional.interpolate( + mask_probs.unsqueeze(0), size=target_size, mode="bilinear", align_corners=False + )[0] + + current_segment_id = 0 + + # Weigh each mask by its prediction score + mask_probs *= pred_scores.view(-1, 1, 1) + mask_labels = mask_probs.argmax(0) # [height, width] + + # Keep track of instances of each class + stuff_memory_list: Dict[str, int] = {} + for k in range(pred_labels.shape[0]): + pred_class = pred_labels[k].item() + should_fuse = pred_class in label_ids_to_fuse + + # Check if mask exists and large enough to be a segment + mask_exists, mask_k = check_segment_validity( + mask_labels, mask_probs, k, mask_threshold, overlap_mask_area_threshold + ) + + if mask_exists: + if pred_class in stuff_memory_list: + current_segment_id = stuff_memory_list[pred_class] + else: + current_segment_id += 1 + + # Add current object segment to final segmentation map + segmentation[mask_k] = current_segment_id + segment_score = round(pred_scores[k].item(), 6) + segments.append( + { + "id": current_segment_id, + "label_id": pred_class, + "was_fused": should_fuse, + "score": segment_score, + } + ) + if should_fuse: + stuff_memory_list[pred_class] = current_segment_id + + return segmentation, segments + + +class DeformableDetrImageProcessor(BaseImageProcessor): + r""" + Constructs a Deformable DETR image processor. + + Args: + format (`str`, *optional*, defaults to `"coco_detection"`): + Data format of the annotations. One of "coco_detection" or "coco_panoptic". + do_resize (`bool`, *optional*, defaults to `True`): + Controls whether to resize the image's (height, width) dimensions to the specified `size`. Can be + overridden by the `do_resize` parameter in the `preprocess` method. + size (`Dict[str, int]` *optional*, defaults to `{"shortest_edge": 800, "longest_edge": 1333}`): + Size of the image's `(height, width)` dimensions after resizing. Can be overridden by the `size` parameter + in the `preprocess` method. Available options are: + - `{"height": int, "width": int}`: The image will be resized to the exact size `(height, width)`. + Do NOT keep the aspect ratio. + - `{"shortest_edge": int, "longest_edge": int}`: The image will be resized to a maximum size respecting + the aspect ratio and keeping the shortest edge less or equal to `shortest_edge` and the longest edge + less or equal to `longest_edge`. + - `{"max_height": int, "max_width": int}`: The image will be resized to the maximum size respecting the + aspect ratio and keeping the height less or equal to `max_height` and the width less or equal to + `max_width`. + resample (`PILImageResampling`, *optional*, defaults to `PILImageResampling.BILINEAR`): + Resampling filter to use if resizing the image. + do_rescale (`bool`, *optional*, defaults to `True`): + Controls whether to rescale the image by the specified scale `rescale_factor`. Can be overridden by the + `do_rescale` parameter in the `preprocess` method. + rescale_factor (`int` or `float`, *optional*, defaults to `1/255`): + Scale factor to use if rescaling the image. Can be overridden by the `rescale_factor` parameter in the + `preprocess` method. + do_normalize: + Controls whether to normalize the image. Can be overridden by the `do_normalize` parameter in the + `preprocess` method. + image_mean (`float` or `List[float]`, *optional*, defaults to `IMAGENET_DEFAULT_MEAN`): + Mean values to use when normalizing the image. Can be a single value or a list of values, one for each + channel. Can be overridden by the `image_mean` parameter in the `preprocess` method. + image_std (`float` or `List[float]`, *optional*, defaults to `IMAGENET_DEFAULT_STD`): + Standard deviation values to use when normalizing the image. Can be a single value or a list of values, one + for each channel. Can be overridden by the `image_std` parameter in the `preprocess` method. + do_convert_annotations (`bool`, *optional*, defaults to `True`): + Controls whether to convert the annotations to the format expected by the DETR model. Converts the + bounding boxes to the format `(center_x, center_y, width, height)` and in the range `[0, 1]`. + Can be overridden by the `do_convert_annotations` parameter in the `preprocess` method. + do_pad (`bool`, *optional*, defaults to `True`): + Controls whether to pad the image. Can be overridden by the `do_pad` parameter in the `preprocess` + method. If `True`, padding will be applied to the bottom and right of the image with zeros. + If `pad_size` is provided, the image will be padded to the specified dimensions. + Otherwise, the image will be padded to the maximum height and width of the batch. + pad_size (`Dict[str, int]`, *optional*): + The size `{"height": int, "width" int}` to pad the images to. Must be larger than any image size + provided for preprocessing. If `pad_size` is not provided, images will be padded to the largest + height and width in the batch. + """ + + model_input_names = ["pixel_values", "pixel_mask"] + + # Copied from transformers.models.detr.image_processing_detr.DetrImageProcessor.__init__ + def __init__( + self, + format: Union[str, AnnotationFormat] = AnnotationFormat.COCO_DETECTION, + do_resize: bool = True, + size: Dict[str, int] = None, + resample: PILImageResampling = PILImageResampling.BILINEAR, + do_rescale: bool = True, + rescale_factor: Union[int, float] = 1 / 255, + do_normalize: bool = True, + image_mean: Union[float, List[float]] = None, + image_std: Union[float, List[float]] = None, + do_convert_annotations: Optional[bool] = None, + do_pad: bool = True, + pad_size: Optional[Dict[str, int]] = None, + **kwargs, + ) -> None: + if "pad_and_return_pixel_mask" in kwargs: + do_pad = kwargs.pop("pad_and_return_pixel_mask") + + if "max_size" in kwargs: + logger.warning_once( + "The `max_size` parameter is deprecated and will be removed in v4.26. " + "Please specify in `size['longest_edge'] instead`.", + ) + max_size = kwargs.pop("max_size") + else: + max_size = None if size is None else 1333 + + size = size if size is not None else {"shortest_edge": 800, "longest_edge": 1333} + size = get_size_dict(size, max_size=max_size, default_to_square=False) + + # Backwards compatibility + if do_convert_annotations is None: + do_convert_annotations = do_normalize + + super().__init__(**kwargs) + self.format = format + self.do_resize = do_resize + self.size = size + self.resample = resample + self.do_rescale = do_rescale + self.rescale_factor = rescale_factor + self.do_normalize = do_normalize + self.do_convert_annotations = do_convert_annotations + self.image_mean = image_mean if image_mean is not None else IMAGENET_DEFAULT_MEAN + self.image_std = image_std if image_std is not None else IMAGENET_DEFAULT_STD + self.do_pad = do_pad + self.pad_size = pad_size + self._valid_processor_keys = [ + "images", + "annotations", + "return_segmentation_masks", + "masks_path", + "do_resize", + "size", + "resample", + "do_rescale", + "rescale_factor", + "do_normalize", + "do_convert_annotations", + "image_mean", + "image_std", + "do_pad", + "pad_size", + "format", + "return_tensors", + "data_format", + "input_data_format", + ] + + @classmethod + # Copied from transformers.models.detr.image_processing_detr.DetrImageProcessor.from_dict with Detr->DeformableDetr + def from_dict(cls, image_processor_dict: Dict[str, Any], **kwargs): + """ + Overrides the `from_dict` method from the base class to make sure parameters are updated if image processor is + created using from_dict and kwargs e.g. `DeformableDetrImageProcessor.from_pretrained(checkpoint, size=600, + max_size=800)` + """ + image_processor_dict = image_processor_dict.copy() + if "max_size" in kwargs: + image_processor_dict["max_size"] = kwargs.pop("max_size") + if "pad_and_return_pixel_mask" in kwargs: + image_processor_dict["pad_and_return_pixel_mask"] = kwargs.pop("pad_and_return_pixel_mask") + return super().from_dict(image_processor_dict, **kwargs) + + # Copied from transformers.models.detr.image_processing_detr.DetrImageProcessor.prepare_annotation with DETR->DeformableDetr + def prepare_annotation( + self, + image: np.ndarray, + target: Dict, + format: Optional[AnnotationFormat] = None, + return_segmentation_masks: bool = None, + masks_path: Optional[Union[str, pathlib.Path]] = None, + input_data_format: Optional[Union[str, ChannelDimension]] = None, + ) -> Dict: + """ + Prepare an annotation for feeding into DeformableDetr model. + """ + format = format if format is not None else self.format + + if format == AnnotationFormat.COCO_DETECTION: + return_segmentation_masks = False if return_segmentation_masks is None else return_segmentation_masks + target = prepare_coco_detection_annotation( + image, target, return_segmentation_masks, input_data_format=input_data_format + ) + elif format == AnnotationFormat.COCO_PANOPTIC: + return_segmentation_masks = True if return_segmentation_masks is None else return_segmentation_masks + target = prepare_coco_panoptic_annotation( + image, + target, + masks_path=masks_path, + return_masks=return_segmentation_masks, + input_data_format=input_data_format, + ) + else: + raise ValueError(f"Format {format} is not supported.") + return target + + # Copied from transformers.models.detr.image_processing_detr.DetrImageProcessor.resize + def resize( + self, + image: np.ndarray, + size: Dict[str, int], + resample: PILImageResampling = PILImageResampling.BILINEAR, + data_format: Optional[ChannelDimension] = None, + input_data_format: Optional[Union[str, ChannelDimension]] = None, + **kwargs, + ) -> np.ndarray: + """ + Resize the image to the given size. Size can be `min_size` (scalar) or `(height, width)` tuple. If size is an + int, smaller edge of the image will be matched to this number. + + Args: + image (`np.ndarray`): + Image to resize. + size (`Dict[str, int]`): + Size of the image's `(height, width)` dimensions after resizing. Available options are: + - `{"height": int, "width": int}`: The image will be resized to the exact size `(height, width)`. + Do NOT keep the aspect ratio. + - `{"shortest_edge": int, "longest_edge": int}`: The image will be resized to a maximum size respecting + the aspect ratio and keeping the shortest edge less or equal to `shortest_edge` and the longest edge + less or equal to `longest_edge`. + - `{"max_height": int, "max_width": int}`: The image will be resized to the maximum size respecting the + aspect ratio and keeping the height less or equal to `max_height` and the width less or equal to + `max_width`. + resample (`PILImageResampling`, *optional*, defaults to `PILImageResampling.BILINEAR`): + Resampling filter to use if resizing the image. + data_format (`str` or `ChannelDimension`, *optional*): + The channel dimension format for the output image. If unset, the channel dimension format of the input + image is used. + input_data_format (`ChannelDimension` or `str`, *optional*): + The channel dimension format of the input image. If not provided, it will be inferred. + """ + if "max_size" in kwargs: + logger.warning_once( + "The `max_size` parameter is deprecated and will be removed in v4.26. " + "Please specify in `size['longest_edge'] instead`.", + ) + max_size = kwargs.pop("max_size") + else: + max_size = None + size = get_size_dict(size, max_size=max_size, default_to_square=False) + if "shortest_edge" in size and "longest_edge" in size: + new_size = get_resize_output_image_size( + image, size["shortest_edge"], size["longest_edge"], input_data_format=input_data_format + ) + elif "max_height" in size and "max_width" in size: + new_size = get_image_size_for_max_height_width( + image, size["max_height"], size["max_width"], input_data_format=input_data_format + ) + elif "height" in size and "width" in size: + new_size = (size["height"], size["width"]) + else: + raise ValueError( + "Size must contain 'height' and 'width' keys or 'shortest_edge' and 'longest_edge' keys. Got" + f" {size.keys()}." + ) + image = resize( + image, + size=new_size, + resample=resample, + data_format=data_format, + input_data_format=input_data_format, + **kwargs, + ) + return image + + # Copied from transformers.models.detr.image_processing_detr.DetrImageProcessor.resize_annotation + def resize_annotation( + self, + annotation, + orig_size, + size, + resample: PILImageResampling = PILImageResampling.NEAREST, + ) -> Dict: + """ + Resize the annotation to match the resized image. If size is an int, smaller edge of the mask will be matched + to this number. + """ + return resize_annotation(annotation, orig_size=orig_size, target_size=size, resample=resample) + + # Copied from transformers.models.detr.image_processing_detr.DetrImageProcessor.rescale + def rescale( + self, + image: np.ndarray, + rescale_factor: float, + data_format: Optional[Union[str, ChannelDimension]] = None, + input_data_format: Optional[Union[str, ChannelDimension]] = None, + ) -> np.ndarray: + """ + Rescale the image by the given factor. image = image * rescale_factor. + + Args: + image (`np.ndarray`): + Image to rescale. + rescale_factor (`float`): + The value to use for rescaling. + data_format (`str` or `ChannelDimension`, *optional*): + The channel dimension format for the output image. If unset, the channel dimension format of the input + image is used. Can be one of: + - `"channels_first"` or `ChannelDimension.FIRST`: image in (num_channels, height, width) format. + - `"channels_last"` or `ChannelDimension.LAST`: image in (height, width, num_channels) format. + input_data_format (`str` or `ChannelDimension`, *optional*): + The channel dimension format for the input image. If unset, is inferred from the input image. Can be + one of: + - `"channels_first"` or `ChannelDimension.FIRST`: image in (num_channels, height, width) format. + - `"channels_last"` or `ChannelDimension.LAST`: image in (height, width, num_channels) format. + """ + return rescale(image, rescale_factor, data_format=data_format, input_data_format=input_data_format) + + # Copied from transformers.models.detr.image_processing_detr.DetrImageProcessor.normalize_annotation + def normalize_annotation(self, annotation: Dict, image_size: Tuple[int, int]) -> Dict: + """ + Normalize the boxes in the annotation from `[top_left_x, top_left_y, bottom_right_x, bottom_right_y]` to + `[center_x, center_y, width, height]` format and from absolute to relative pixel values. + """ + return normalize_annotation(annotation, image_size=image_size) + + # Copied from transformers.models.detr.image_processing_detr.DetrImageProcessor._update_annotation_for_padded_image + def _update_annotation_for_padded_image( + self, + annotation: Dict, + input_image_size: Tuple[int, int], + output_image_size: Tuple[int, int], + padding, + update_bboxes, + ) -> Dict: + """ + Update the annotation for a padded image. + """ + new_annotation = {} + new_annotation["size"] = output_image_size + + for key, value in annotation.items(): + if key == "masks": + masks = value + masks = pad( + masks, + padding, + mode=PaddingMode.CONSTANT, + constant_values=0, + input_data_format=ChannelDimension.FIRST, + ) + masks = safe_squeeze(masks, 1) + new_annotation["masks"] = masks + elif key == "boxes" and update_bboxes: + boxes = value + boxes *= np.asarray( + [ + input_image_size[1] / output_image_size[1], + input_image_size[0] / output_image_size[0], + input_image_size[1] / output_image_size[1], + input_image_size[0] / output_image_size[0], + ] + ) + new_annotation["boxes"] = boxes + elif key == "size": + new_annotation["size"] = output_image_size + else: + new_annotation[key] = value + return new_annotation + + # Copied from transformers.models.detr.image_processing_detr.DetrImageProcessor._pad_image + def _pad_image( + self, + image: np.ndarray, + output_size: Tuple[int, int], + annotation: Optional[Dict[str, Any]] = None, + constant_values: Union[float, Iterable[float]] = 0, + data_format: Optional[ChannelDimension] = None, + input_data_format: Optional[Union[str, ChannelDimension]] = None, + update_bboxes: bool = True, + ) -> np.ndarray: + """ + Pad an image with zeros to the given size. + """ + input_height, input_width = get_image_size(image, channel_dim=input_data_format) + output_height, output_width = output_size + + pad_bottom = output_height - input_height + pad_right = output_width - input_width + padding = ((0, pad_bottom), (0, pad_right)) + padded_image = pad( + image, + padding, + mode=PaddingMode.CONSTANT, + constant_values=constant_values, + data_format=data_format, + input_data_format=input_data_format, + ) + if annotation is not None: + annotation = self._update_annotation_for_padded_image( + annotation, (input_height, input_width), (output_height, output_width), padding, update_bboxes + ) + return padded_image, annotation + + # Copied from transformers.models.detr.image_processing_detr.DetrImageProcessor.pad + def pad( + self, + images: List[np.ndarray], + annotations: Optional[Union[AnnotationType, List[AnnotationType]]] = None, + constant_values: Union[float, Iterable[float]] = 0, + return_pixel_mask: bool = True, + return_tensors: Optional[Union[str, TensorType]] = None, + data_format: Optional[ChannelDimension] = None, + input_data_format: Optional[Union[str, ChannelDimension]] = None, + update_bboxes: bool = True, + pad_size: Optional[Dict[str, int]] = None, + ) -> BatchFeature: + """ + Pads a batch of images to the bottom and right of the image with zeros to the size of largest height and width + in the batch and optionally returns their corresponding pixel mask. + + Args: + images (List[`np.ndarray`]): + Images to pad. + annotations (`AnnotationType` or `List[AnnotationType]`, *optional*): + Annotations to transform according to the padding that is applied to the images. + constant_values (`float` or `Iterable[float]`, *optional*): + The value to use for the padding if `mode` is `"constant"`. + return_pixel_mask (`bool`, *optional*, defaults to `True`): + Whether to return a pixel mask. + return_tensors (`str` or `TensorType`, *optional*): + The type of tensors to return. Can be one of: + - Unset: Return a list of `np.ndarray`. + - `TensorType.TENSORFLOW` or `'tf'`: Return a batch of type `tf.Tensor`. + - `TensorType.PYTORCH` or `'pt'`: Return a batch of type `torch.Tensor`. + - `TensorType.NUMPY` or `'np'`: Return a batch of type `np.ndarray`. + - `TensorType.JAX` or `'jax'`: Return a batch of type `jax.numpy.ndarray`. + data_format (`str` or `ChannelDimension`, *optional*): + The channel dimension format of the image. If not provided, it will be the same as the input image. + input_data_format (`ChannelDimension` or `str`, *optional*): + The channel dimension format of the input image. If not provided, it will be inferred. + update_bboxes (`bool`, *optional*, defaults to `True`): + Whether to update the bounding boxes in the annotations to match the padded images. If the + bounding boxes have not been converted to relative coordinates and `(centre_x, centre_y, width, height)` + format, the bounding boxes will not be updated. + pad_size (`Dict[str, int]`, *optional*): + The size `{"height": int, "width" int}` to pad the images to. Must be larger than any image size + provided for preprocessing. If `pad_size` is not provided, images will be padded to the largest + height and width in the batch. + """ + pad_size = pad_size if pad_size is not None else self.pad_size + if pad_size is not None: + padded_size = (pad_size["height"], pad_size["width"]) + else: + padded_size = get_max_height_width(images, input_data_format=input_data_format) + + annotation_list = annotations if annotations is not None else [None] * len(images) + padded_images = [] + padded_annotations = [] + for image, annotation in zip(images, annotation_list): + padded_image, padded_annotation = self._pad_image( + image, + padded_size, + annotation, + constant_values=constant_values, + data_format=data_format, + input_data_format=input_data_format, + update_bboxes=update_bboxes, + ) + padded_images.append(padded_image) + padded_annotations.append(padded_annotation) + + data = {"pixel_values": padded_images} + + if return_pixel_mask: + masks = [ + make_pixel_mask(image=image, output_size=padded_size, input_data_format=input_data_format) + for image in images + ] + data["pixel_mask"] = masks + + encoded_inputs = BatchFeature(data=data, tensor_type=return_tensors) + + if annotations is not None: + encoded_inputs["labels"] = [ + BatchFeature(annotation, tensor_type=return_tensors) for annotation in padded_annotations + ] + + return encoded_inputs + + # Copied from transformers.models.detr.image_processing_detr.DetrImageProcessor.preprocess + def preprocess( + self, + images: ImageInput, + annotations: Optional[Union[AnnotationType, List[AnnotationType]]] = None, + return_segmentation_masks: bool = None, + masks_path: Optional[Union[str, pathlib.Path]] = None, + do_resize: Optional[bool] = None, + size: Optional[Dict[str, int]] = None, + resample=None, # PILImageResampling + do_rescale: Optional[bool] = None, + rescale_factor: Optional[Union[int, float]] = None, + do_normalize: Optional[bool] = None, + do_convert_annotations: Optional[bool] = None, + image_mean: Optional[Union[float, List[float]]] = None, + image_std: Optional[Union[float, List[float]]] = None, + do_pad: Optional[bool] = None, + format: Optional[Union[str, AnnotationFormat]] = None, + return_tensors: Optional[Union[TensorType, str]] = None, + data_format: Union[str, ChannelDimension] = ChannelDimension.FIRST, + input_data_format: Optional[Union[str, ChannelDimension]] = None, + pad_size: Optional[Dict[str, int]] = None, + **kwargs, + ) -> BatchFeature: + """ + Preprocess an image or a batch of images so that it can be used by the model. + + Args: + images (`ImageInput`): + Image or batch of images to preprocess. Expects a single or batch of images with pixel values ranging + from 0 to 255. If passing in images with pixel values between 0 and 1, set `do_rescale=False`. + annotations (`AnnotationType` or `List[AnnotationType]`, *optional*): + List of annotations associated with the image or batch of images. If annotation is for object + detection, the annotations should be a dictionary with the following keys: + - "image_id" (`int`): The image id. + - "annotations" (`List[Dict]`): List of annotations for an image. Each annotation should be a + dictionary. An image can have no annotations, in which case the list should be empty. + If annotation is for segmentation, the annotations should be a dictionary with the following keys: + - "image_id" (`int`): The image id. + - "segments_info" (`List[Dict]`): List of segments for an image. Each segment should be a dictionary. + An image can have no segments, in which case the list should be empty. + - "file_name" (`str`): The file name of the image. + return_segmentation_masks (`bool`, *optional*, defaults to self.return_segmentation_masks): + Whether to return segmentation masks. + masks_path (`str` or `pathlib.Path`, *optional*): + Path to the directory containing the segmentation masks. + do_resize (`bool`, *optional*, defaults to self.do_resize): + Whether to resize the image. + size (`Dict[str, int]`, *optional*, defaults to self.size): + Size of the image's `(height, width)` dimensions after resizing. Available options are: + - `{"height": int, "width": int}`: The image will be resized to the exact size `(height, width)`. + Do NOT keep the aspect ratio. + - `{"shortest_edge": int, "longest_edge": int}`: The image will be resized to a maximum size respecting + the aspect ratio and keeping the shortest edge less or equal to `shortest_edge` and the longest edge + less or equal to `longest_edge`. + - `{"max_height": int, "max_width": int}`: The image will be resized to the maximum size respecting the + aspect ratio and keeping the height less or equal to `max_height` and the width less or equal to + `max_width`. + resample (`PILImageResampling`, *optional*, defaults to self.resample): + Resampling filter to use when resizing the image. + do_rescale (`bool`, *optional*, defaults to self.do_rescale): + Whether to rescale the image. + rescale_factor (`float`, *optional*, defaults to self.rescale_factor): + Rescale factor to use when rescaling the image. + do_normalize (`bool`, *optional*, defaults to self.do_normalize): + Whether to normalize the image. + do_convert_annotations (`bool`, *optional*, defaults to self.do_convert_annotations): + Whether to convert the annotations to the format expected by the model. Converts the bounding + boxes from the format `(top_left_x, top_left_y, width, height)` to `(center_x, center_y, width, height)` + and in relative coordinates. + image_mean (`float` or `List[float]`, *optional*, defaults to self.image_mean): + Mean to use when normalizing the image. + image_std (`float` or `List[float]`, *optional*, defaults to self.image_std): + Standard deviation to use when normalizing the image. + do_pad (`bool`, *optional*, defaults to self.do_pad): + Whether to pad the image. If `True`, padding will be applied to the bottom and right of + the image with zeros. If `pad_size` is provided, the image will be padded to the specified + dimensions. Otherwise, the image will be padded to the maximum height and width of the batch. + format (`str` or `AnnotationFormat`, *optional*, defaults to self.format): + Format of the annotations. + return_tensors (`str` or `TensorType`, *optional*, defaults to self.return_tensors): + Type of tensors to return. If `None`, will return the list of images. + data_format (`ChannelDimension` or `str`, *optional*, defaults to `ChannelDimension.FIRST`): + The channel dimension format for the output image. Can be one of: + - `"channels_first"` or `ChannelDimension.FIRST`: image in (num_channels, height, width) format. + - `"channels_last"` or `ChannelDimension.LAST`: image in (height, width, num_channels) format. + - Unset: Use the channel dimension format of the input image. + input_data_format (`ChannelDimension` or `str`, *optional*): + The channel dimension format for the input image. If unset, the channel dimension format is inferred + from the input image. Can be one of: + - `"channels_first"` or `ChannelDimension.FIRST`: image in (num_channels, height, width) format. + - `"channels_last"` or `ChannelDimension.LAST`: image in (height, width, num_channels) format. + - `"none"` or `ChannelDimension.NONE`: image in (height, width) format. + pad_size (`Dict[str, int]`, *optional*): + The size `{"height": int, "width" int}` to pad the images to. Must be larger than any image size + provided for preprocessing. If `pad_size` is not provided, images will be padded to the largest + height and width in the batch. + """ + if "pad_and_return_pixel_mask" in kwargs: + logger.warning_once( + "The `pad_and_return_pixel_mask` argument is deprecated and will be removed in a future version, " + "use `do_pad` instead." + ) + do_pad = kwargs.pop("pad_and_return_pixel_mask") + + if "max_size" in kwargs: + logger.warning_once( + "The `max_size` argument is deprecated and will be removed in a future version, use" + " `size['longest_edge']` instead." + ) + size = kwargs.pop("max_size") + + do_resize = self.do_resize if do_resize is None else do_resize + size = self.size if size is None else size + size = get_size_dict(size=size, default_to_square=False) + resample = self.resample if resample is None else resample + do_rescale = self.do_rescale if do_rescale is None else do_rescale + rescale_factor = self.rescale_factor if rescale_factor is None else rescale_factor + do_normalize = self.do_normalize if do_normalize is None else do_normalize + image_mean = self.image_mean if image_mean is None else image_mean + image_std = self.image_std if image_std is None else image_std + do_convert_annotations = ( + self.do_convert_annotations if do_convert_annotations is None else do_convert_annotations + ) + do_pad = self.do_pad if do_pad is None else do_pad + pad_size = self.pad_size if pad_size is None else pad_size + format = self.format if format is None else format + + images = make_list_of_images(images) + + if not valid_images(images): + raise ValueError( + "Invalid image type. Must be of type PIL.Image.Image, numpy.ndarray, " + "torch.Tensor, tf.Tensor or jax.ndarray." + ) + validate_kwargs(captured_kwargs=kwargs.keys(), valid_processor_keys=self._valid_processor_keys) + + # Here, the pad() method pads to the maximum of (width, height). It does not need to be validated. + validate_preprocess_arguments( + do_rescale=do_rescale, + rescale_factor=rescale_factor, + do_normalize=do_normalize, + image_mean=image_mean, + image_std=image_std, + do_resize=do_resize, + size=size, + resample=resample, + ) + + if annotations is not None and isinstance(annotations, dict): + annotations = [annotations] + + if annotations is not None and len(images) != len(annotations): + raise ValueError( + f"The number of images ({len(images)}) and annotations ({len(annotations)}) do not match." + ) + + format = AnnotationFormat(format) + if annotations is not None: + validate_annotations(format, SUPPORTED_ANNOTATION_FORMATS, annotations) + + if ( + masks_path is not None + and format == AnnotationFormat.COCO_PANOPTIC + and not isinstance(masks_path, (pathlib.Path, str)) + ): + raise ValueError( + "The path to the directory containing the mask PNG files should be provided as a" + f" `pathlib.Path` or string object, but is {type(masks_path)} instead." + ) + + # All transformations expect numpy arrays + images = [to_numpy_array(image) for image in images] + + if do_rescale and is_scaled_image(images[0]): + logger.warning_once( + "It looks like you are trying to rescale already rescaled images. If the input" + " images have pixel values between 0 and 1, set `do_rescale=False` to avoid rescaling them again." + ) + + if input_data_format is None: + # We assume that all images have the same channel dimension format. + input_data_format = infer_channel_dimension_format(images[0]) + + # prepare (COCO annotations as a list of Dict -> DETR target as a single Dict per image) + if annotations is not None: + prepared_images = [] + prepared_annotations = [] + for image, target in zip(images, annotations): + target = self.prepare_annotation( + image, + target, + format, + return_segmentation_masks=return_segmentation_masks, + masks_path=masks_path, + input_data_format=input_data_format, + ) + prepared_images.append(image) + prepared_annotations.append(target) + images = prepared_images + annotations = prepared_annotations + del prepared_images, prepared_annotations + + # transformations + if do_resize: + if annotations is not None: + resized_images, resized_annotations = [], [] + for image, target in zip(images, annotations): + orig_size = get_image_size(image, input_data_format) + resized_image = self.resize( + image, size=size, resample=resample, input_data_format=input_data_format + ) + resized_annotation = self.resize_annotation( + target, orig_size, get_image_size(resized_image, input_data_format) + ) + resized_images.append(resized_image) + resized_annotations.append(resized_annotation) + images = resized_images + annotations = resized_annotations + del resized_images, resized_annotations + else: + images = [ + self.resize(image, size=size, resample=resample, input_data_format=input_data_format) + for image in images + ] + + if do_rescale: + images = [self.rescale(image, rescale_factor, input_data_format=input_data_format) for image in images] + + if do_normalize: + images = [ + self.normalize(image, image_mean, image_std, input_data_format=input_data_format) for image in images + ] + + if do_convert_annotations and annotations is not None: + annotations = [ + self.normalize_annotation(annotation, get_image_size(image, input_data_format)) + for annotation, image in zip(annotations, images) + ] + + if do_pad: + # Pads images and returns their mask: {'pixel_values': ..., 'pixel_mask': ...} + encoded_inputs = self.pad( + images, + annotations=annotations, + return_pixel_mask=True, + data_format=data_format, + input_data_format=input_data_format, + update_bboxes=do_convert_annotations, + return_tensors=return_tensors, + pad_size=pad_size, + ) + else: + images = [ + to_channel_dimension_format(image, data_format, input_channel_dim=input_data_format) + for image in images + ] + encoded_inputs = BatchFeature(data={"pixel_values": images}, tensor_type=return_tensors) + if annotations is not None: + encoded_inputs["labels"] = [ + BatchFeature(annotation, tensor_type=return_tensors) for annotation in annotations + ] + + return encoded_inputs + + # POSTPROCESSING METHODS - TODO: add support for other frameworks + def post_process(self, outputs, target_sizes): + """ + Converts the raw output of [`DeformableDetrForObjectDetection`] into final bounding boxes in (top_left_x, + top_left_y, bottom_right_x, bottom_right_y) format. Only supports PyTorch. + + Args: + outputs ([`DeformableDetrObjectDetectionOutput`]): + Raw outputs of the model. + target_sizes (`torch.Tensor` of shape `(batch_size, 2)`): + Tensor containing the size (height, width) of each image of the batch. For evaluation, this must be the + original image size (before any data augmentation). For visualization, this should be the image size + after data augment, but before padding. + Returns: + `List[Dict]`: A list of dictionaries, each dictionary containing the scores, labels and boxes for an image + in the batch as predicted by the model. + """ + logger.warning_once( + "`post_process` is deprecated and will be removed in v5 of Transformers, please use" + " `post_process_object_detection` instead, with `threshold=0.` for equivalent results.", + ) + + out_logits, out_bbox = outputs.logits, outputs.pred_boxes + + if len(out_logits) != len(target_sizes): + raise ValueError("Make sure that you pass in as many target sizes as the batch dimension of the logits") + if target_sizes.shape[1] != 2: + raise ValueError("Each element of target_sizes must contain the size (h, w) of each image of the batch") + + prob = out_logits.sigmoid() + topk_values, topk_indexes = torch.topk(prob.view(out_logits.shape[0], -1), 100, dim=1) + scores = topk_values + topk_boxes = torch.div(topk_indexes, out_logits.shape[2], rounding_mode="floor") + labels = topk_indexes % out_logits.shape[2] + boxes = center_to_corners_format(out_bbox) + boxes = torch.gather(boxes, 1, topk_boxes.unsqueeze(-1).repeat(1, 1, 4)) + + # and from relative [0, 1] to absolute [0, height] coordinates + img_h, img_w = target_sizes.unbind(1) + scale_fct = torch.stack([img_w, img_h, img_w, img_h], dim=1) + boxes = boxes * scale_fct[:, None, :] + + results = [{"scores": s, "labels": l, "boxes": b} for s, l, b in zip(scores, labels, boxes)] + + return results + + def post_process_object_detection( + self, outputs, threshold: float = 0.5, target_sizes: Union[TensorType, List[Tuple]] = None, top_k: int = 100 + ): + """ + Converts the raw output of [`DeformableDetrForObjectDetection`] into final bounding boxes in (top_left_x, + top_left_y, bottom_right_x, bottom_right_y) format. Only supports PyTorch. + + Args: + outputs ([`DetrObjectDetectionOutput`]): + Raw outputs of the model. + threshold (`float`, *optional*): + Score threshold to keep object detection predictions. + target_sizes (`torch.Tensor` or `List[Tuple[int, int]]`, *optional*): + Tensor of shape `(batch_size, 2)` or list of tuples (`Tuple[int, int]`) containing the target size + (height, width) of each image in the batch. If left to None, predictions will not be resized. + top_k (`int`, *optional*, defaults to 100): + Keep only top k bounding boxes before filtering by thresholding. + + Returns: + `List[Dict]`: A list of dictionaries, each dictionary containing the scores, labels and boxes for an image + in the batch as predicted by the model. + """ + out_logits, out_bbox = outputs.logits, outputs.pred_boxes + + if target_sizes is not None: + if len(out_logits) != len(target_sizes): + raise ValueError( + "Make sure that you pass in as many target sizes as the batch dimension of the logits" + ) + + prob = out_logits.sigmoid() + prob = prob.view(out_logits.shape[0], -1) + k_value = min(top_k, prob.size(1)) + topk_values, topk_indexes = torch.topk(prob, k_value, dim=1) + scores = topk_values + topk_boxes = torch.div(topk_indexes, out_logits.shape[2], rounding_mode="floor") + labels = topk_indexes % out_logits.shape[2] + boxes = center_to_corners_format(out_bbox) + boxes = torch.gather(boxes, 1, topk_boxes.unsqueeze(-1).repeat(1, 1, 4)) + + # and from relative [0, 1] to absolute [0, height] coordinates + if target_sizes is not None: + if isinstance(target_sizes, List): + img_h = torch.Tensor([i[0] for i in target_sizes]) + img_w = torch.Tensor([i[1] for i in target_sizes]) + else: + img_h, img_w = target_sizes.unbind(1) + scale_fct = torch.stack([img_w, img_h, img_w, img_h], dim=1).to(boxes.device) + boxes = boxes * scale_fct[:, None, :] + + results = [] + for s, l, b in zip(scores, labels, boxes): + score = s[s > threshold] + label = l[s > threshold] + box = b[s > threshold] + results.append({"scores": score, "labels": label, "boxes": box}) + + return results + + +__all__ = ["DeformableDetrImageProcessor"] diff --git a/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/image_processing_deformable_detr_fast.py b/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/image_processing_deformable_detr_fast.py new file mode 100644 index 0000000000000000000000000000000000000000..5d400f3d139c1b54fd21a9ce92937ed16b58d98e --- /dev/null +++ b/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/image_processing_deformable_detr_fast.py @@ -0,0 +1,843 @@ +# 🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨 +# This file was automatically generated from src/transformers/models/deformable_detr/modular_deformable_detr.py. +# Do NOT edit this file manually as any edits will be overwritten by the generation of +# the file from the modular. If any change should be done, please apply the change to the +# modular_deformable_detr.py file directly. One of our CI enforces this. +# 🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨 +import pathlib +from typing import Any, Dict, List, Optional, Tuple, Union + +from ...image_processing_utils import BatchFeature, get_size_dict +from ...image_processing_utils_fast import ( + BASE_IMAGE_PROCESSOR_FAST_DOCSTRING, + BASE_IMAGE_PROCESSOR_FAST_DOCSTRING_PREPROCESS, + BaseImageProcessorFast, + DefaultFastImageProcessorKwargs, + SizeDict, + get_image_size_for_max_height_width, + get_max_height_width, + safe_squeeze, +) +from ...image_transforms import center_to_corners_format, corners_to_center_format +from ...image_utils import ( + IMAGENET_DEFAULT_MEAN, + IMAGENET_DEFAULT_STD, + AnnotationFormat, + AnnotationType, + ChannelDimension, + ImageInput, + PILImageResampling, + get_image_size, + validate_annotations, +) +from ...processing_utils import Unpack +from ...utils import ( + TensorType, + add_start_docstrings, + is_torch_available, + is_torchvision_available, + is_torchvision_v2_available, + logging, +) +from .image_processing_deformable_detr import get_size_with_aspect_ratio + + +if is_torch_available(): + import torch + + +if is_torchvision_v2_available(): + from torchvision.io import read_image + from torchvision.transforms.v2 import functional as F +elif is_torchvision_available(): + from torchvision.io import read_image + from torchvision.transforms import functional as F + + +logger = logging.get_logger(__name__) + + +class DeformableDetrFastImageProcessorKwargs(DefaultFastImageProcessorKwargs): + format: Optional[Union[str, AnnotationFormat]] + do_convert_annotations: Optional[bool] + do_pad: Optional[bool] + pad_size: Optional[Dict[str, int]] + return_segmentation_masks: Optional[bool] + + +SUPPORTED_ANNOTATION_FORMATS = (AnnotationFormat.COCO_DETECTION, AnnotationFormat.COCO_PANOPTIC) + + +# inspired by https://github.com/facebookresearch/deformable_detr/blob/master/datasets/coco.py#L33 +def convert_coco_poly_to_mask(segmentations, height: int, width: int, device: torch.device) -> torch.Tensor: + """ + Convert a COCO polygon annotation to a mask. + + Args: + segmentations (`List[List[float]]`): + List of polygons, each polygon represented by a list of x-y coordinates. + height (`int`): + Height of the mask. + width (`int`): + Width of the mask. + """ + try: + from pycocotools import mask as coco_mask + except ImportError: + raise ImportError("Pycocotools is not installed in your environment.") + + masks = [] + for polygons in segmentations: + rles = coco_mask.frPyObjects(polygons, height, width) + mask = coco_mask.decode(rles) + if len(mask.shape) < 3: + mask = mask[..., None] + mask = torch.as_tensor(mask, dtype=torch.uint8, device=device) + mask = torch.any(mask, axis=2) + masks.append(mask) + if masks: + masks = torch.stack(masks, axis=0) + else: + masks = torch.zeros((0, height, width), dtype=torch.uint8, device=device) + + return masks + + +# inspired by https://github.com/facebookresearch/deformable_detr/blob/master/datasets/coco.py#L50 +def prepare_coco_detection_annotation( + image, + target, + return_segmentation_masks: bool = False, + input_data_format: Optional[Union[ChannelDimension, str]] = None, +): + """ + Convert the target in COCO format into the format expected by DEFORMABLE_DETR. + """ + image_height, image_width = image.size()[-2:] + + image_id = target["image_id"] + image_id = torch.as_tensor([image_id], dtype=torch.int64, device=image.device) + + # Get all COCO annotations for the given image. + annotations = target["annotations"] + classes = [] + area = [] + boxes = [] + keypoints = [] + for obj in annotations: + if "iscrowd" not in obj or obj["iscrowd"] == 0: + classes.append(obj["category_id"]) + area.append(obj["area"]) + boxes.append(obj["bbox"]) + if "keypoints" in obj: + keypoints.append(obj["keypoints"]) + + classes = torch.as_tensor(classes, dtype=torch.int64, device=image.device) + area = torch.as_tensor(area, dtype=torch.float32, device=image.device) + iscrowd = torch.zeros_like(classes, dtype=torch.int64, device=image.device) + # guard against no boxes via resizing + boxes = torch.as_tensor(boxes, dtype=torch.float32, device=image.device).reshape(-1, 4) + boxes[:, 2:] += boxes[:, :2] + boxes[:, 0::2] = boxes[:, 0::2].clip(min=0, max=image_width) + boxes[:, 1::2] = boxes[:, 1::2].clip(min=0, max=image_height) + + keep = (boxes[:, 3] > boxes[:, 1]) & (boxes[:, 2] > boxes[:, 0]) + + new_target = { + "image_id": image_id, + "class_labels": classes[keep], + "boxes": boxes[keep], + "area": area[keep], + "iscrowd": iscrowd[keep], + "orig_size": torch.as_tensor([int(image_height), int(image_width)], dtype=torch.int64, device=image.device), + } + + if keypoints: + keypoints = torch.as_tensor(keypoints, dtype=torch.float32, device=image.device) + # Apply the keep mask here to filter the relevant annotations + keypoints = keypoints[keep] + num_keypoints = keypoints.shape[0] + keypoints = keypoints.reshape((-1, 3)) if num_keypoints else keypoints + new_target["keypoints"] = keypoints + + if return_segmentation_masks: + segmentation_masks = [obj["segmentation"] for obj in annotations] + masks = convert_coco_poly_to_mask(segmentation_masks, image_height, image_width, device=image.device) + new_target["masks"] = masks[keep] + + return new_target + + +def masks_to_boxes(masks: torch.Tensor) -> torch.Tensor: + """ + Compute the bounding boxes around the provided panoptic segmentation masks. + + Args: + masks: masks in format `[number_masks, height, width]` where N is the number of masks + + Returns: + boxes: bounding boxes in format `[number_masks, 4]` in xyxy format + """ + if masks.numel() == 0: + return torch.zeros((0, 4), device=masks.device) + + h, w = masks.shape[-2:] + y = torch.arange(0, h, dtype=torch.float32, device=masks.device) + x = torch.arange(0, w, dtype=torch.float32, device=masks.device) + # see https://github.com/pytorch/pytorch/issues/50276 + y, x = torch.meshgrid(y, x, indexing="ij") + + x_mask = masks * torch.unsqueeze(x, 0) + x_max = x_mask.view(x_mask.shape[0], -1).max(-1)[0] + x_min = ( + torch.where(masks, x.unsqueeze(0), torch.tensor(1e8, device=masks.device)).view(masks.shape[0], -1).min(-1)[0] + ) + + y_mask = masks * torch.unsqueeze(y, 0) + y_max = y_mask.view(y_mask.shape[0], -1).max(-1)[0] + y_min = ( + torch.where(masks, y.unsqueeze(0), torch.tensor(1e8, device=masks.device)).view(masks.shape[0], -1).min(-1)[0] + ) + + return torch.stack([x_min, y_min, x_max, y_max], 1) + + +# 2 functions below adapted from https://github.com/cocodataset/panopticapi/blob/master/panopticapi/utils.py +# Copyright (c) 2018, Alexander Kirillov +# All rights reserved. +def rgb_to_id(color): + """ + Converts RGB color to unique ID. + """ + if isinstance(color, torch.Tensor) and len(color.shape) == 3: + if color.dtype == torch.uint8: + color = color.to(torch.int32) + return color[:, :, 0] + 256 * color[:, :, 1] + 256 * 256 * color[:, :, 2] + return int(color[0] + 256 * color[1] + 256 * 256 * color[2]) + + +def prepare_coco_panoptic_annotation( + image: torch.Tensor, + target: Dict, + masks_path: Union[str, pathlib.Path], + return_masks: bool = True, + input_data_format: Union[ChannelDimension, str] = None, +) -> Dict: + """ + Prepare a coco panoptic annotation for DEFORMABLE_DETR. + """ + image_height, image_width = get_image_size(image, channel_dim=input_data_format) + annotation_path = pathlib.Path(masks_path) / target["file_name"] + + new_target = {} + new_target["image_id"] = torch.as_tensor( + [target["image_id"] if "image_id" in target else target["id"]], dtype=torch.int64, device=image.device + ) + new_target["size"] = torch.as_tensor([image_height, image_width], dtype=torch.int64, device=image.device) + new_target["orig_size"] = torch.as_tensor([image_height, image_width], dtype=torch.int64, device=image.device) + + if "segments_info" in target: + masks = read_image(annotation_path).permute(1, 2, 0).to(torch.int32).to(image.device) + masks = rgb_to_id(masks) + + ids = torch.as_tensor([segment_info["id"] for segment_info in target["segments_info"]], device=image.device) + masks = masks == ids[:, None, None] + masks = masks.to(torch.bool) + if return_masks: + new_target["masks"] = masks + new_target["boxes"] = masks_to_boxes(masks) + new_target["class_labels"] = torch.as_tensor( + [segment_info["category_id"] for segment_info in target["segments_info"]], + dtype=torch.int64, + device=image.device, + ) + new_target["iscrowd"] = torch.as_tensor( + [segment_info["iscrowd"] for segment_info in target["segments_info"]], + dtype=torch.int64, + device=image.device, + ) + new_target["area"] = torch.as_tensor( + [segment_info["area"] for segment_info in target["segments_info"]], + dtype=torch.float32, + device=image.device, + ) + + return new_target + + +@add_start_docstrings( + "Constructs a fast DeformableDetr image processor.", + BASE_IMAGE_PROCESSOR_FAST_DOCSTRING, + """ + format (`str`, *optional*, defaults to `AnnotationFormat.COCO_DETECTION`): + Data format of the annotations. One of "coco_detection" or "coco_panoptic". + do_convert_annotations (`bool`, *optional*, defaults to `True`): + Controls whether to convert the annotations to the format expected by the DEFORMABLE_DETR model. Converts the + bounding boxes to the format `(center_x, center_y, width, height)` and in the range `[0, 1]`. + Can be overridden by the `do_convert_annotations` parameter in the `preprocess` method. + do_pad (`bool`, *optional*, defaults to `True`): + Controls whether to pad the image. Can be overridden by the `do_pad` parameter in the `preprocess` + method. If `True`, padding will be applied to the bottom and right of the image with zeros. + If `pad_size` is provided, the image will be padded to the specified dimensions. + Otherwise, the image will be padded to the maximum height and width of the batch. + pad_size (`Dict[str, int]`, *optional*): + The size `{"height": int, "width" int}` to pad the images to. Must be larger than any image size + provided for preprocessing. If `pad_size` is not provided, images will be padded to the largest + height and width in the batch. + return_segmentation_masks (`bool`, *optional*, defaults to `False`): + Whether to return segmentation masks. + """, +) +class DeformableDetrImageProcessorFast(BaseImageProcessorFast): + resample = PILImageResampling.BILINEAR + image_mean = IMAGENET_DEFAULT_MEAN + image_std = IMAGENET_DEFAULT_STD + format = AnnotationFormat.COCO_DETECTION + do_resize = True + do_rescale = True + do_normalize = True + do_pad = True + size = {"shortest_edge": 800, "longest_edge": 1333} + default_to_square = False + model_input_names = ["pixel_values", "pixel_mask"] + valid_kwargs = DeformableDetrFastImageProcessorKwargs + + def __init__(self, **kwargs: Unpack[DeformableDetrFastImageProcessorKwargs]) -> None: + if "pad_and_return_pixel_mask" in kwargs: + kwargs["do_pad"] = kwargs.pop("pad_and_return_pixel_mask") + + size = kwargs.pop("size", None) + if "max_size" in kwargs: + logger.warning_once( + "The `max_size` parameter is deprecated and will be removed in v4.26. " + "Please specify in `size['longest_edge'] instead`.", + ) + max_size = kwargs.pop("max_size") + else: + max_size = None if size is None else 1333 + + size = size if size is not None else {"shortest_edge": 800, "longest_edge": 1333} + self.size = get_size_dict(size, max_size=max_size, default_to_square=False) + + # Backwards compatibility + do_convert_annotations = kwargs.get("do_convert_annotations", None) + do_normalize = kwargs.get("do_normalize", None) + if do_convert_annotations is None and getattr(self, "do_convert_annotations", None) is None: + self.do_convert_annotations = do_normalize if do_normalize is not None else self.do_normalize + + super().__init__(**kwargs) + + @classmethod + def from_dict(cls, image_processor_dict: Dict[str, Any], **kwargs): + """ + Overrides the `from_dict` method from the base class to make sure parameters are updated if image processor is + created using from_dict and kwargs e.g. `DeformableDetrImageProcessorFast.from_pretrained(checkpoint, size=600, + max_size=800)` + """ + image_processor_dict = image_processor_dict.copy() + if "max_size" in kwargs: + image_processor_dict["max_size"] = kwargs.pop("max_size") + if "pad_and_return_pixel_mask" in kwargs: + image_processor_dict["pad_and_return_pixel_mask"] = kwargs.pop("pad_and_return_pixel_mask") + return super().from_dict(image_processor_dict, **kwargs) + + def prepare_annotation( + self, + image: torch.Tensor, + target: Dict, + format: Optional[AnnotationFormat] = None, + return_segmentation_masks: bool = None, + masks_path: Optional[Union[str, pathlib.Path]] = None, + input_data_format: Optional[Union[str, ChannelDimension]] = None, + ) -> Dict: + """ + Prepare an annotation for feeding into DEFORMABLE_DETR model. + """ + format = format if format is not None else self.format + + if format == AnnotationFormat.COCO_DETECTION: + return_segmentation_masks = False if return_segmentation_masks is None else return_segmentation_masks + target = prepare_coco_detection_annotation( + image, target, return_segmentation_masks, input_data_format=input_data_format + ) + elif format == AnnotationFormat.COCO_PANOPTIC: + return_segmentation_masks = True if return_segmentation_masks is None else return_segmentation_masks + target = prepare_coco_panoptic_annotation( + image, + target, + masks_path=masks_path, + return_masks=return_segmentation_masks, + input_data_format=input_data_format, + ) + else: + raise ValueError(f"Format {format} is not supported.") + return target + + def resize( + self, + image: torch.Tensor, + size: SizeDict, + interpolation: "F.InterpolationMode" = None, + **kwargs, + ) -> torch.Tensor: + """ + Resize the image to the given size. Size can be `min_size` (scalar) or `(height, width)` tuple. If size is an + int, smaller edge of the image will be matched to this number. + + Args: + image (`torch.Tensor`): + Image to resize. + size (`SizeDict`): + Size of the image's `(height, width)` dimensions after resizing. Available options are: + - `{"height": int, "width": int}`: The image will be resized to the exact size `(height, width)`. + Do NOT keep the aspect ratio. + - `{"shortest_edge": int, "longest_edge": int}`: The image will be resized to a maximum size respecting + the aspect ratio and keeping the shortest edge less or equal to `shortest_edge` and the longest edge + less or equal to `longest_edge`. + - `{"max_height": int, "max_width": int}`: The image will be resized to the maximum size respecting the + aspect ratio and keeping the height less or equal to `max_height` and the width less or equal to + `max_width`. + interpolation (`InterpolationMode`, *optional*, defaults to `InterpolationMode.BILINEAR`): + Resampling filter to use if resizing the image. + """ + interpolation = interpolation if interpolation is not None else F.InterpolationMode.BILINEAR + if size.shortest_edge and size.longest_edge: + # Resize the image so that the shortest edge or the longest edge is of the given size + # while maintaining the aspect ratio of the original image. + new_size = get_size_with_aspect_ratio( + image.size()[-2:], + size["shortest_edge"], + size["longest_edge"], + ) + elif size.max_height and size.max_width: + new_size = get_image_size_for_max_height_width(image.size()[-2:], size["max_height"], size["max_width"]) + elif size.height and size.width: + new_size = (size["height"], size["width"]) + else: + raise ValueError( + "Size must contain 'height' and 'width' keys or 'shortest_edge' and 'longest_edge' keys. Got" + f" {size.keys()}." + ) + + image = F.resize( + image, + size=new_size, + interpolation=interpolation, + **kwargs, + ) + return image + + def resize_annotation( + self, + annotation: Dict[str, Any], + orig_size: Tuple[int, int], + target_size: Tuple[int, int], + threshold: float = 0.5, + interpolation: "F.InterpolationMode" = None, + ): + """ + Resizes an annotation to a target size. + + Args: + annotation (`Dict[str, Any]`): + The annotation dictionary. + orig_size (`Tuple[int, int]`): + The original size of the input image. + target_size (`Tuple[int, int]`): + The target size of the image, as returned by the preprocessing `resize` step. + threshold (`float`, *optional*, defaults to 0.5): + The threshold used to binarize the segmentation masks. + resample (`InterpolationMode`, defaults to `InterpolationMode.NEAREST`): + The resampling filter to use when resizing the masks. + """ + interpolation = interpolation if interpolation is not None else F.InterpolationMode.NEAREST + ratio_height, ratio_width = [target / orig for target, orig in zip(target_size, orig_size)] + + new_annotation = {} + new_annotation["size"] = target_size + + for key, value in annotation.items(): + if key == "boxes": + boxes = value + scaled_boxes = boxes * torch.as_tensor( + [ratio_width, ratio_height, ratio_width, ratio_height], dtype=torch.float32, device=boxes.device + ) + new_annotation["boxes"] = scaled_boxes + elif key == "area": + area = value + scaled_area = area * (ratio_width * ratio_height) + new_annotation["area"] = scaled_area + elif key == "masks": + masks = value[:, None] + masks = [F.resize(mask, target_size, interpolation=interpolation) for mask in masks] + masks = torch.stack(masks).to(torch.float32) + masks = masks[:, 0] > threshold + new_annotation["masks"] = masks + elif key == "size": + new_annotation["size"] = target_size + else: + new_annotation[key] = value + + return new_annotation + + def normalize_annotation(self, annotation: Dict, image_size: Tuple[int, int]) -> Dict: + image_height, image_width = image_size + norm_annotation = {} + for key, value in annotation.items(): + if key == "boxes": + boxes = value + boxes = corners_to_center_format(boxes) + boxes /= torch.as_tensor( + [image_width, image_height, image_width, image_height], dtype=torch.float32, device=boxes.device + ) + norm_annotation[key] = boxes + else: + norm_annotation[key] = value + return norm_annotation + + def _update_annotation_for_padded_image( + self, + annotation: Dict, + input_image_size: Tuple[int, int], + output_image_size: Tuple[int, int], + padding, + update_bboxes, + ) -> Dict: + """ + Update the annotation for a padded image. + """ + new_annotation = {} + new_annotation["size"] = output_image_size + ratio_height, ratio_width = (input / output for output, input in zip(output_image_size, input_image_size)) + + for key, value in annotation.items(): + if key == "masks": + masks = value + masks = F.pad( + masks, + padding, + fill=0, + ) + masks = safe_squeeze(masks, 1) + new_annotation["masks"] = masks + elif key == "boxes" and update_bboxes: + boxes = value + boxes *= torch.as_tensor([ratio_width, ratio_height, ratio_width, ratio_height], device=boxes.device) + new_annotation["boxes"] = boxes + elif key == "size": + new_annotation["size"] = output_image_size + else: + new_annotation[key] = value + return new_annotation + + def pad( + self, + image: torch.Tensor, + padded_size: Tuple[int, int], + annotation: Optional[Dict[str, Any]] = None, + update_bboxes: bool = True, + fill: int = 0, + ): + original_size = image.size()[-2:] + padding_bottom = padded_size[0] - original_size[0] + padding_right = padded_size[1] - original_size[1] + if padding_bottom < 0 or padding_right < 0: + raise ValueError( + f"Padding dimensions are negative. Please make sure that the padded size is larger than the " + f"original size. Got padded size: {padded_size}, original size: {original_size}." + ) + if original_size != padded_size: + padding = [0, 0, padding_right, padding_bottom] + image = F.pad(image, padding, fill=fill) + if annotation is not None: + annotation = self._update_annotation_for_padded_image( + annotation, original_size, padded_size, padding, update_bboxes + ) + + # Make a pixel mask for the image, where 1 indicates a valid pixel and 0 indicates padding. + pixel_mask = torch.zeros(padded_size, dtype=torch.int64, device=image.device) + pixel_mask[: original_size[0], : original_size[1]] = 1 + + return image, pixel_mask, annotation + + @add_start_docstrings( + BASE_IMAGE_PROCESSOR_FAST_DOCSTRING_PREPROCESS, + """ + annotations (`AnnotationType` or `List[AnnotationType]`, *optional*): + List of annotations associated with the image or batch of images. If annotation is for object + detection, the annotations should be a dictionary with the following keys: + - "image_id" (`int`): The image id. + - "annotations" (`List[Dict]`): List of annotations for an image. Each annotation should be a + dictionary. An image can have no annotations, in which case the list should be empty. + If annotation is for segmentation, the annotations should be a dictionary with the following keys: + - "image_id" (`int`): The image id. + - "segments_info" (`List[Dict]`): List of segments for an image. Each segment should be a dictionary. + An image can have no segments, in which case the list should be empty. + - "file_name" (`str`): The file name of the image. + format (`str`, *optional*, defaults to `AnnotationFormat.COCO_DETECTION`): + Data format of the annotations. One of "coco_detection" or "coco_panoptic". + do_convert_annotations (`bool`, *optional*, defaults to `True`): + Controls whether to convert the annotations to the format expected by the DEFORMABLE_DETR model. Converts the + bounding boxes to the format `(center_x, center_y, width, height)` and in the range `[0, 1]`. + Can be overridden by the `do_convert_annotations` parameter in the `preprocess` method. + do_pad (`bool`, *optional*, defaults to `True`): + Controls whether to pad the image. Can be overridden by the `do_pad` parameter in the `preprocess` + method. If `True`, padding will be applied to the bottom and right of the image with zeros. + If `pad_size` is provided, the image will be padded to the specified dimensions. + Otherwise, the image will be padded to the maximum height and width of the batch. + pad_size (`Dict[str, int]`, *optional*): + The size `{"height": int, "width" int}` to pad the images to. Must be larger than any image size + provided for preprocessing. If `pad_size` is not provided, images will be padded to the largest + height and width in the batch. + return_segmentation_masks (`bool`, *optional*, defaults to `False`): + Whether to return segmentation masks. + masks_path (`str` or `pathlib.Path`, *optional*): + Path to the directory containing the segmentation masks. + """, + ) + def preprocess( + self, + images: ImageInput, + annotations: Optional[Union[AnnotationType, List[AnnotationType]]] = None, + masks_path: Optional[Union[str, pathlib.Path]] = None, + **kwargs: Unpack[DeformableDetrFastImageProcessorKwargs], + ) -> BatchFeature: + if "pad_and_return_pixel_mask" in kwargs: + kwargs["do_pad"] = kwargs.pop("pad_and_return_pixel_mask") + logger.warning_once( + "The `pad_and_return_pixel_mask` argument is deprecated and will be removed in a future version, " + "use `do_pad` instead." + ) + + if "max_size" in kwargs: + logger.warning_once( + "The `max_size` argument is deprecated and will be removed in a future version, use" + " `size['longest_edge']` instead." + ) + kwargs["size"] = kwargs.pop("max_size") + + return super().preprocess(images, annotations=annotations, masks_path=masks_path, **kwargs) + + def _preprocess( + self, + images: List["torch.Tensor"], + annotations: Optional[Union[AnnotationType, List[AnnotationType]]], + return_segmentation_masks: bool, + masks_path: Optional[Union[str, pathlib.Path]], + do_resize: bool, + size: SizeDict, + interpolation: Optional["F.InterpolationMode"], + do_center_crop: bool, + crop_size: SizeDict, + do_rescale: bool, + rescale_factor: float, + do_normalize: bool, + do_convert_annotations: bool, + image_mean: Optional[Union[float, List[float]]], + image_std: Optional[Union[float, List[float]]], + do_pad: bool, + pad_size: Optional[Dict[str, int]], + format: Optional[Union[str, AnnotationFormat]], + return_tensors: Optional[Union[str, TensorType]], + ) -> BatchFeature: + """ + Preprocess an image or a batch of images so that it can be used by the model. + """ + if annotations is not None and isinstance(annotations, dict): + annotations = [annotations] + + if annotations is not None and len(images) != len(annotations): + raise ValueError( + f"The number of images ({len(images)}) and annotations ({len(annotations)}) do not match." + ) + + format = AnnotationFormat(format) + if annotations is not None: + validate_annotations(format, SUPPORTED_ANNOTATION_FORMATS, annotations) + + if ( + masks_path is not None + and format == AnnotationFormat.COCO_PANOPTIC + and not isinstance(masks_path, (pathlib.Path, str)) + ): + raise ValueError( + "The path to the directory containing the mask PNG files should be provided as a" + f" `pathlib.Path` or string object, but is {type(masks_path)} instead." + ) + + data = {} + + processed_images = [] + processed_annotations = [] + pixel_masks = [] # Initialize pixel_masks here + for image, annotation in zip(images, annotations if annotations is not None else [None] * len(images)): + # prepare (COCO annotations as a list of Dict -> DEFORMABLE_DETR target as a single Dict per image) + if annotations is not None: + annotation = self.prepare_annotation( + image, + annotation, + format, + return_segmentation_masks=return_segmentation_masks, + masks_path=masks_path, + input_data_format=ChannelDimension.FIRST, + ) + + if do_resize: + resized_image = self.resize(image, size=size, interpolation=interpolation) + if annotations is not None: + annotation = self.resize_annotation( + annotation, + orig_size=image.size()[-2:], + target_size=resized_image.size()[-2:], + ) + image = resized_image + # Fused rescale and normalize + image = self.rescale_and_normalize(image, do_rescale, rescale_factor, do_normalize, image_mean, image_std) + if do_convert_annotations and annotations is not None: + annotation = self.normalize_annotation(annotation, get_image_size(image, ChannelDimension.FIRST)) + + processed_images.append(image) + processed_annotations.append(annotation) + images = processed_images + annotations = processed_annotations if annotations is not None else None + + if do_pad: + # depends on all resized image shapes so we need another loop + if pad_size is not None: + padded_size = (pad_size["height"], pad_size["width"]) + else: + padded_size = get_max_height_width(images) + + padded_images = [] + padded_annotations = [] + for image, annotation in zip(images, annotations if annotations is not None else [None] * len(images)): + # Pads images and returns their mask: {'pixel_values': ..., 'pixel_mask': ...} + if padded_size == image.size()[-2:]: + padded_images.append(image) + pixel_masks.append(torch.ones(padded_size, dtype=torch.int64, device=image.device)) + padded_annotations.append(annotation) + continue + image, pixel_mask, annotation = self.pad( + image, padded_size, annotation=annotation, update_bboxes=do_convert_annotations + ) + padded_images.append(image) + padded_annotations.append(annotation) + pixel_masks.append(pixel_mask) + images = padded_images + annotations = padded_annotations if annotations is not None else None + data.update({"pixel_mask": torch.stack(pixel_masks, dim=0)}) + + data.update({"pixel_values": torch.stack(images, dim=0)}) + encoded_inputs = BatchFeature(data, tensor_type=return_tensors) + if annotations is not None: + encoded_inputs["labels"] = [ + BatchFeature(annotation, tensor_type=return_tensors) for annotation in annotations + ] + return encoded_inputs + + def post_process(self, outputs, target_sizes): + """ + Converts the raw output of [`DeformableDetrForObjectDetection`] into final bounding boxes in (top_left_x, + top_left_y, bottom_right_x, bottom_right_y) format. Only supports PyTorch. + + Args: + outputs ([`DeformableDetrObjectDetectionOutput`]): + Raw outputs of the model. + target_sizes (`torch.Tensor` of shape `(batch_size, 2)`): + Tensor containing the size (height, width) of each image of the batch. For evaluation, this must be the + original image size (before any data augmentation). For visualization, this should be the image size + after data augment, but before padding. + Returns: + `List[Dict]`: A list of dictionaries, each dictionary containing the scores, labels and boxes for an image + in the batch as predicted by the model. + """ + logger.warning_once( + "`post_process` is deprecated and will be removed in v5 of Transformers, please use" + " `post_process_object_detection` instead, with `threshold=0.` for equivalent results.", + ) + + out_logits, out_bbox = outputs.logits, outputs.pred_boxes + + if len(out_logits) != len(target_sizes): + raise ValueError("Make sure that you pass in as many target sizes as the batch dimension of the logits") + if target_sizes.shape[1] != 2: + raise ValueError("Each element of target_sizes must contain the size (h, w) of each image of the batch") + + prob = out_logits.sigmoid() + topk_values, topk_indexes = torch.topk(prob.view(out_logits.shape[0], -1), 100, dim=1) + scores = topk_values + topk_boxes = torch.div(topk_indexes, out_logits.shape[2], rounding_mode="floor") + labels = topk_indexes % out_logits.shape[2] + boxes = center_to_corners_format(out_bbox) + boxes = torch.gather(boxes, 1, topk_boxes.unsqueeze(-1).repeat(1, 1, 4)) + + # and from relative [0, 1] to absolute [0, height] coordinates + img_h, img_w = target_sizes.unbind(1) + scale_fct = torch.stack([img_w, img_h, img_w, img_h], dim=1) + boxes = boxes * scale_fct[:, None, :] + + results = [{"scores": s, "labels": l, "boxes": b} for s, l, b in zip(scores, labels, boxes)] + + return results + + def post_process_object_detection( + self, outputs, threshold: float = 0.5, target_sizes: Union[TensorType, List[Tuple]] = None, top_k: int = 100 + ): + """ + Converts the raw output of [`DeformableDetrForObjectDetection`] into final bounding boxes in (top_left_x, + top_left_y, bottom_right_x, bottom_right_y) format. Only supports PyTorch. + + Args: + outputs ([`DetrObjectDetectionOutput`]): + Raw outputs of the model. + threshold (`float`, *optional*): + Score threshold to keep object detection predictions. + target_sizes (`torch.Tensor` or `List[Tuple[int, int]]`, *optional*): + Tensor of shape `(batch_size, 2)` or list of tuples (`Tuple[int, int]`) containing the target size + (height, width) of each image in the batch. If left to None, predictions will not be resized. + top_k (`int`, *optional*, defaults to 100): + Keep only top k bounding boxes before filtering by thresholding. + + Returns: + `List[Dict]`: A list of dictionaries, each dictionary containing the scores, labels and boxes for an image + in the batch as predicted by the model. + """ + out_logits, out_bbox = outputs.logits, outputs.pred_boxes + + if target_sizes is not None: + if len(out_logits) != len(target_sizes): + raise ValueError( + "Make sure that you pass in as many target sizes as the batch dimension of the logits" + ) + + prob = out_logits.sigmoid() + prob = prob.view(out_logits.shape[0], -1) + k_value = min(top_k, prob.size(1)) + topk_values, topk_indexes = torch.topk(prob, k_value, dim=1) + scores = topk_values + topk_boxes = torch.div(topk_indexes, out_logits.shape[2], rounding_mode="floor") + labels = topk_indexes % out_logits.shape[2] + boxes = center_to_corners_format(out_bbox) + boxes = torch.gather(boxes, 1, topk_boxes.unsqueeze(-1).repeat(1, 1, 4)) + + # and from relative [0, 1] to absolute [0, height] coordinates + if target_sizes is not None: + if isinstance(target_sizes, List): + img_h = torch.Tensor([i[0] for i in target_sizes]) + img_w = torch.Tensor([i[1] for i in target_sizes]) + else: + img_h, img_w = target_sizes.unbind(1) + scale_fct = torch.stack([img_w, img_h, img_w, img_h], dim=1).to(boxes.device) + boxes = boxes * scale_fct[:, None, :] + + results = [] + for s, l, b in zip(scores, labels, boxes): + score = s[s > threshold] + label = l[s > threshold] + box = b[s > threshold] + results.append({"scores": score, "labels": label, "boxes": box}) + + return results + + +__all__ = ["DeformableDetrImageProcessorFast"] diff --git a/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/modeling_deformable_detr.py b/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/modeling_deformable_detr.py new file mode 100644 index 0000000000000000000000000000000000000000..8db75abc0a772c545f1fe3bca2a7917616f3f9e2 --- /dev/null +++ b/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/modeling_deformable_detr.py @@ -0,0 +1,2030 @@ +# coding=utf-8 +# Copyright 2022 SenseTime and The HuggingFace Inc. team. All rights reserved. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +"""PyTorch Deformable DETR model.""" + +import copy +import math +import warnings +from dataclasses import dataclass +from typing import Dict, List, Optional, Tuple, Union + +import torch +import torch.nn.functional as F +from torch import Tensor, nn + +from ...activations import ACT2FN +from ...integrations import use_kernel_forward_from_hub +from ...modeling_attn_mask_utils import _prepare_4d_attention_mask +from ...modeling_outputs import BaseModelOutput +from ...modeling_utils import PreTrainedModel +from ...pytorch_utils import meshgrid +from ...utils import ( + ModelOutput, + add_start_docstrings, + add_start_docstrings_to_model_forward, + is_timm_available, + logging, + replace_return_docstrings, + requires_backends, +) +from ...utils.backbone_utils import load_backbone +from .configuration_deformable_detr import DeformableDetrConfig + + +logger = logging.get_logger(__name__) + + +if is_timm_available(): + from timm import create_model + + +logger = logging.get_logger(__name__) + +_CONFIG_FOR_DOC = "DeformableDetrConfig" +_CHECKPOINT_FOR_DOC = "sensetime/deformable-detr" + + +@use_kernel_forward_from_hub("MultiScaleDeformableAttention") +class MultiScaleDeformableAttention(nn.Module): + def forward( + self, + value: Tensor, + value_spatial_shapes: Tensor, + value_spatial_shapes_list: List[Tuple], + level_start_index: Tensor, + sampling_locations: Tensor, + attention_weights: Tensor, + im2col_step: int, + ): + batch_size, _, num_heads, hidden_dim = value.shape + _, num_queries, num_heads, num_levels, num_points, _ = sampling_locations.shape + value_list = value.split([height * width for height, width in value_spatial_shapes_list], dim=1) + sampling_grids = 2 * sampling_locations - 1 + sampling_value_list = [] + for level_id, (height, width) in enumerate(value_spatial_shapes_list): + # batch_size, height*width, num_heads, hidden_dim + # -> batch_size, height*width, num_heads*hidden_dim + # -> batch_size, num_heads*hidden_dim, height*width + # -> batch_size*num_heads, hidden_dim, height, width + value_l_ = ( + value_list[level_id] + .flatten(2) + .transpose(1, 2) + .reshape(batch_size * num_heads, hidden_dim, height, width) + ) + # batch_size, num_queries, num_heads, num_points, 2 + # -> batch_size, num_heads, num_queries, num_points, 2 + # -> batch_size*num_heads, num_queries, num_points, 2 + sampling_grid_l_ = sampling_grids[:, :, :, level_id].transpose(1, 2).flatten(0, 1) + # batch_size*num_heads, hidden_dim, num_queries, num_points + sampling_value_l_ = nn.functional.grid_sample( + value_l_, + sampling_grid_l_, + mode="bilinear", + padding_mode="zeros", + align_corners=False, + ) + sampling_value_list.append(sampling_value_l_) + # (batch_size, num_queries, num_heads, num_levels, num_points) + # -> (batch_size, num_heads, num_queries, num_levels, num_points) + # -> (batch_size, num_heads, 1, num_queries, num_levels*num_points) + attention_weights = attention_weights.transpose(1, 2).reshape( + batch_size * num_heads, 1, num_queries, num_levels * num_points + ) + output = ( + (torch.stack(sampling_value_list, dim=-2).flatten(-2) * attention_weights) + .sum(-1) + .view(batch_size, num_heads * hidden_dim, num_queries) + ) + return output.transpose(1, 2).contiguous() + + +@dataclass +class DeformableDetrDecoderOutput(ModelOutput): + """ + Base class for outputs of the DeformableDetrDecoder. This class adds two attributes to + BaseModelOutputWithCrossAttentions, namely: + - a stacked tensor of intermediate decoder hidden states (i.e. the output of each decoder layer) + - a stacked tensor of intermediate reference points. + + Args: + last_hidden_state (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`): + Sequence of hidden-states at the output of the last layer of the model. + intermediate_hidden_states (`torch.FloatTensor` of shape `(batch_size, config.decoder_layers, num_queries, hidden_size)`): + Stacked intermediate hidden states (output of each layer of the decoder). + intermediate_reference_points (`torch.FloatTensor` of shape `(batch_size, config.decoder_layers, sequence_length, hidden_size)`): + Stacked intermediate reference points (reference points of each layer of the decoder). + hidden_states (`tuple(torch.FloatTensor)`, *optional*, returned when `output_hidden_states=True` is passed or when `config.output_hidden_states=True`): + Tuple of `torch.FloatTensor` (one for the output of the embeddings + one for the output of each layer) of + shape `(batch_size, sequence_length, hidden_size)`. Hidden-states of the model at the output of each layer + plus the initial embedding outputs. + attentions (`tuple(torch.FloatTensor)`, *optional*, returned when `output_attentions=True` is passed or when `config.output_attentions=True`): + Tuple of `torch.FloatTensor` (one for each layer) of shape `(batch_size, num_heads, sequence_length, + sequence_length)`. Attentions weights after the attention softmax, used to compute the weighted average in + the self-attention heads. + cross_attentions (`tuple(torch.FloatTensor)`, *optional*, returned when `output_attentions=True` and `config.add_cross_attention=True` is passed or when `config.output_attentions=True`): + Tuple of `torch.FloatTensor` (one for each layer) of shape `(batch_size, num_heads, sequence_length, + sequence_length)`. Attentions weights of the decoder's cross-attention layer, after the attention softmax, + used to compute the weighted average in the cross-attention heads. + """ + + last_hidden_state: torch.FloatTensor = None + intermediate_hidden_states: torch.FloatTensor = None + intermediate_reference_points: torch.FloatTensor = None + hidden_states: Optional[Tuple[torch.FloatTensor]] = None + attentions: Optional[Tuple[torch.FloatTensor]] = None + cross_attentions: Optional[Tuple[torch.FloatTensor]] = None + + +@dataclass +class DeformableDetrModelOutput(ModelOutput): + """ + Base class for outputs of the Deformable DETR encoder-decoder model. + + Args: + init_reference_points (`torch.FloatTensor` of shape `(batch_size, num_queries, 4)`): + Initial reference points sent through the Transformer decoder. + last_hidden_state (`torch.FloatTensor` of shape `(batch_size, num_queries, hidden_size)`): + Sequence of hidden-states at the output of the last layer of the decoder of the model. + intermediate_hidden_states (`torch.FloatTensor` of shape `(batch_size, config.decoder_layers, num_queries, hidden_size)`): + Stacked intermediate hidden states (output of each layer of the decoder). + intermediate_reference_points (`torch.FloatTensor` of shape `(batch_size, config.decoder_layers, num_queries, 4)`): + Stacked intermediate reference points (reference points of each layer of the decoder). + decoder_hidden_states (`tuple(torch.FloatTensor)`, *optional*, returned when `output_hidden_states=True` is passed or when `config.output_hidden_states=True`): + Tuple of `torch.FloatTensor` (one for the output of the embeddings + one for the output of each layer) of + shape `(batch_size, num_queries, hidden_size)`. Hidden-states of the decoder at the output of each layer + plus the initial embedding outputs. + decoder_attentions (`tuple(torch.FloatTensor)`, *optional*, returned when `output_attentions=True` is passed or when `config.output_attentions=True`): + Tuple of `torch.FloatTensor` (one for each layer) of shape `(batch_size, num_heads, num_queries, + num_queries)`. Attentions weights of the decoder, after the attention softmax, used to compute the weighted + average in the self-attention heads. + cross_attentions (`tuple(torch.FloatTensor)`, *optional*, returned when `output_attentions=True` is passed or when `config.output_attentions=True`): + Tuple of `torch.FloatTensor` (one for each layer) of shape `(batch_size, num_queries, num_heads, 4, 4)`. + Attentions weights of the decoder's cross-attention layer, after the attention softmax, used to compute the + weighted average in the cross-attention heads. + encoder_last_hidden_state (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`, *optional*): + Sequence of hidden-states at the output of the last layer of the encoder of the model. + encoder_hidden_states (`tuple(torch.FloatTensor)`, *optional*, returned when `output_hidden_states=True` is passed or when `config.output_hidden_states=True`): + Tuple of `torch.FloatTensor` (one for the output of the embeddings + one for the output of each layer) of + shape `(batch_size, sequence_length, hidden_size)`. Hidden-states of the encoder at the output of each + layer plus the initial embedding outputs. + encoder_attentions (`tuple(torch.FloatTensor)`, *optional*, returned when `output_attentions=True` is passed or when `config.output_attentions=True`): + Tuple of `torch.FloatTensor` (one for each layer) of shape `(batch_size, num_queries, num_heads, 4, 4)`. + Attentions weights of the encoder, after the attention softmax, used to compute the weighted average in the + self-attention heads. + enc_outputs_class (`torch.FloatTensor` of shape `(batch_size, sequence_length, config.num_labels)`, *optional*, returned when `config.with_box_refine=True` and `config.two_stage=True`): + Predicted bounding boxes scores where the top `config.two_stage_num_proposals` scoring bounding boxes are + picked as region proposals in the first stage. Output of bounding box binary classification (i.e. + foreground and background). + enc_outputs_coord_logits (`torch.FloatTensor` of shape `(batch_size, sequence_length, 4)`, *optional*, returned when `config.with_box_refine=True` and `config.two_stage=True`): + Logits of predicted bounding boxes coordinates in the first stage. + """ + + init_reference_points: torch.FloatTensor = None + last_hidden_state: torch.FloatTensor = None + intermediate_hidden_states: torch.FloatTensor = None + intermediate_reference_points: torch.FloatTensor = None + decoder_hidden_states: Optional[Tuple[torch.FloatTensor]] = None + decoder_attentions: Optional[Tuple[torch.FloatTensor]] = None + cross_attentions: Optional[Tuple[torch.FloatTensor]] = None + encoder_last_hidden_state: Optional[torch.FloatTensor] = None + encoder_hidden_states: Optional[Tuple[torch.FloatTensor]] = None + encoder_attentions: Optional[Tuple[torch.FloatTensor]] = None + enc_outputs_class: Optional[torch.FloatTensor] = None + enc_outputs_coord_logits: Optional[torch.FloatTensor] = None + + +@dataclass +class DeformableDetrObjectDetectionOutput(ModelOutput): + """ + Output type of [`DeformableDetrForObjectDetection`]. + + Args: + loss (`torch.FloatTensor` of shape `(1,)`, *optional*, returned when `labels` are provided)): + Total loss as a linear combination of a negative log-likehood (cross-entropy) for class prediction and a + bounding box loss. The latter is defined as a linear combination of the L1 loss and the generalized + scale-invariant IoU loss. + loss_dict (`Dict`, *optional*): + A dictionary containing the individual losses. Useful for logging. + logits (`torch.FloatTensor` of shape `(batch_size, num_queries, num_classes + 1)`): + Classification logits (including no-object) for all queries. + pred_boxes (`torch.FloatTensor` of shape `(batch_size, num_queries, 4)`): + Normalized boxes coordinates for all queries, represented as (center_x, center_y, width, height). These + values are normalized in [0, 1], relative to the size of each individual image in the batch (disregarding + possible padding). You can use [`~DeformableDetrProcessor.post_process_object_detection`] to retrieve the + unnormalized bounding boxes. + auxiliary_outputs (`list[Dict]`, *optional*): + Optional, only returned when auxilary losses are activated (i.e. `config.auxiliary_loss` is set to `True`) + and labels are provided. It is a list of dictionaries containing the two above keys (`logits` and + `pred_boxes`) for each decoder layer. + last_hidden_state (`torch.FloatTensor` of shape `(batch_size, num_queries, hidden_size)`, *optional*): + Sequence of hidden-states at the output of the last layer of the decoder of the model. + decoder_hidden_states (`tuple(torch.FloatTensor)`, *optional*, returned when `output_hidden_states=True` is passed or when `config.output_hidden_states=True`): + Tuple of `torch.FloatTensor` (one for the output of the embeddings + one for the output of each layer) of + shape `(batch_size, num_queries, hidden_size)`. Hidden-states of the decoder at the output of each layer + plus the initial embedding outputs. + decoder_attentions (`tuple(torch.FloatTensor)`, *optional*, returned when `output_attentions=True` is passed or when `config.output_attentions=True`): + Tuple of `torch.FloatTensor` (one for each layer) of shape `(batch_size, num_heads, num_queries, + num_queries)`. Attentions weights of the decoder, after the attention softmax, used to compute the weighted + average in the self-attention heads. + cross_attentions (`tuple(torch.FloatTensor)`, *optional*, returned when `output_attentions=True` is passed or when `config.output_attentions=True`): + Tuple of `torch.FloatTensor` (one for each layer) of shape `(batch_size, num_queries, num_heads, 4, 4)`. + Attentions weights of the decoder's cross-attention layer, after the attention softmax, used to compute the + weighted average in the cross-attention heads. + encoder_last_hidden_state (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`, *optional*): + Sequence of hidden-states at the output of the last layer of the encoder of the model. + encoder_hidden_states (`tuple(torch.FloatTensor)`, *optional*, returned when `output_hidden_states=True` is passed or when `config.output_hidden_states=True`): + Tuple of `torch.FloatTensor` (one for the output of the embeddings + one for the output of each layer) of + shape `(batch_size, sequence_length, hidden_size)`. Hidden-states of the encoder at the output of each + layer plus the initial embedding outputs. + encoder_attentions (`tuple(torch.FloatTensor)`, *optional*, returned when `output_attentions=True` is passed or when `config.output_attentions=True`): + Tuple of `torch.FloatTensor` (one for each layer) of shape `(batch_size, sequence_length, num_heads, 4, + 4)`. Attentions weights of the encoder, after the attention softmax, used to compute the weighted average + in the self-attention heads. + intermediate_hidden_states (`torch.FloatTensor` of shape `(batch_size, config.decoder_layers, num_queries, hidden_size)`): + Stacked intermediate hidden states (output of each layer of the decoder). + intermediate_reference_points (`torch.FloatTensor` of shape `(batch_size, config.decoder_layers, num_queries, 4)`): + Stacked intermediate reference points (reference points of each layer of the decoder). + init_reference_points (`torch.FloatTensor` of shape `(batch_size, num_queries, 4)`): + Initial reference points sent through the Transformer decoder. + enc_outputs_class (`torch.FloatTensor` of shape `(batch_size, sequence_length, config.num_labels)`, *optional*, returned when `config.with_box_refine=True` and `config.two_stage=True`): + Predicted bounding boxes scores where the top `config.two_stage_num_proposals` scoring bounding boxes are + picked as region proposals in the first stage. Output of bounding box binary classification (i.e. + foreground and background). + enc_outputs_coord_logits (`torch.FloatTensor` of shape `(batch_size, sequence_length, 4)`, *optional*, returned when `config.with_box_refine=True` and `config.two_stage=True`): + Logits of predicted bounding boxes coordinates in the first stage. + """ + + loss: Optional[torch.FloatTensor] = None + loss_dict: Optional[Dict] = None + logits: torch.FloatTensor = None + pred_boxes: torch.FloatTensor = None + auxiliary_outputs: Optional[List[Dict]] = None + init_reference_points: Optional[torch.FloatTensor] = None + last_hidden_state: Optional[torch.FloatTensor] = None + intermediate_hidden_states: Optional[torch.FloatTensor] = None + intermediate_reference_points: Optional[torch.FloatTensor] = None + decoder_hidden_states: Optional[Tuple[torch.FloatTensor]] = None + decoder_attentions: Optional[Tuple[torch.FloatTensor]] = None + cross_attentions: Optional[Tuple[torch.FloatTensor]] = None + encoder_last_hidden_state: Optional[torch.FloatTensor] = None + encoder_hidden_states: Optional[Tuple[torch.FloatTensor]] = None + encoder_attentions: Optional[Tuple[torch.FloatTensor]] = None + enc_outputs_class: Optional = None + enc_outputs_coord_logits: Optional = None + + +def _get_clones(module, N): + return nn.ModuleList([copy.deepcopy(module) for i in range(N)]) + + +def inverse_sigmoid(x, eps=1e-5): + x = x.clamp(min=0, max=1) + x1 = x.clamp(min=eps) + x2 = (1 - x).clamp(min=eps) + return torch.log(x1 / x2) + + +# Copied from transformers.models.detr.modeling_detr.DetrFrozenBatchNorm2d with Detr->DeformableDetr +class DeformableDetrFrozenBatchNorm2d(nn.Module): + """ + BatchNorm2d where the batch statistics and the affine parameters are fixed. + + Copy-paste from torchvision.misc.ops with added eps before rqsrt, without which any other models than + torchvision.models.resnet[18,34,50,101] produce nans. + """ + + def __init__(self, n): + super().__init__() + self.register_buffer("weight", torch.ones(n)) + self.register_buffer("bias", torch.zeros(n)) + self.register_buffer("running_mean", torch.zeros(n)) + self.register_buffer("running_var", torch.ones(n)) + + def _load_from_state_dict( + self, state_dict, prefix, local_metadata, strict, missing_keys, unexpected_keys, error_msgs + ): + num_batches_tracked_key = prefix + "num_batches_tracked" + if num_batches_tracked_key in state_dict: + del state_dict[num_batches_tracked_key] + + super()._load_from_state_dict( + state_dict, prefix, local_metadata, strict, missing_keys, unexpected_keys, error_msgs + ) + + def forward(self, x): + # move reshapes to the beginning + # to make it user-friendly + weight = self.weight.reshape(1, -1, 1, 1) + bias = self.bias.reshape(1, -1, 1, 1) + running_var = self.running_var.reshape(1, -1, 1, 1) + running_mean = self.running_mean.reshape(1, -1, 1, 1) + epsilon = 1e-5 + scale = weight * (running_var + epsilon).rsqrt() + bias = bias - running_mean * scale + return x * scale + bias + + +# Copied from transformers.models.detr.modeling_detr.replace_batch_norm with Detr->DeformableDetr +def replace_batch_norm(model): + r""" + Recursively replace all `torch.nn.BatchNorm2d` with `DeformableDetrFrozenBatchNorm2d`. + + Args: + model (torch.nn.Module): + input model + """ + for name, module in model.named_children(): + if isinstance(module, nn.BatchNorm2d): + new_module = DeformableDetrFrozenBatchNorm2d(module.num_features) + + if not module.weight.device == torch.device("meta"): + new_module.weight.data.copy_(module.weight) + new_module.bias.data.copy_(module.bias) + new_module.running_mean.data.copy_(module.running_mean) + new_module.running_var.data.copy_(module.running_var) + + model._modules[name] = new_module + + if len(list(module.children())) > 0: + replace_batch_norm(module) + + +class DeformableDetrConvEncoder(nn.Module): + """ + Convolutional backbone, using either the AutoBackbone API or one from the timm library. + + nn.BatchNorm2d layers are replaced by DeformableDetrFrozenBatchNorm2d as defined above. + + """ + + def __init__(self, config): + super().__init__() + + self.config = config + + # For backwards compatibility we have to use the timm library directly instead of the AutoBackbone API + if config.use_timm_backbone: + # We default to values which were previously hard-coded. This enables configurability from the config + # using backbone arguments, while keeping the default behavior the same. + requires_backends(self, ["timm"]) + kwargs = getattr(config, "backbone_kwargs", {}) + kwargs = {} if kwargs is None else kwargs.copy() + out_indices = kwargs.pop("out_indices", (2, 3, 4) if config.num_feature_levels > 1 else (4,)) + num_channels = kwargs.pop("in_chans", config.num_channels) + if config.dilation: + kwargs["output_stride"] = kwargs.get("output_stride", 16) + backbone = create_model( + config.backbone, + pretrained=config.use_pretrained_backbone, + features_only=True, + out_indices=out_indices, + in_chans=num_channels, + **kwargs, + ) + else: + backbone = load_backbone(config) + + # replace batch norm by frozen batch norm + with torch.no_grad(): + replace_batch_norm(backbone) + self.model = backbone + self.intermediate_channel_sizes = ( + self.model.feature_info.channels() if config.use_timm_backbone else self.model.channels + ) + + backbone_model_type = None + if config.backbone is not None: + backbone_model_type = config.backbone + elif config.backbone_config is not None: + backbone_model_type = config.backbone_config.model_type + else: + raise ValueError("Either `backbone` or `backbone_config` should be provided in the config") + + if "resnet" in backbone_model_type: + for name, parameter in self.model.named_parameters(): + if config.use_timm_backbone: + if "layer2" not in name and "layer3" not in name and "layer4" not in name: + parameter.requires_grad_(False) + else: + if "stage.1" not in name and "stage.2" not in name and "stage.3" not in name: + parameter.requires_grad_(False) + + # Copied from transformers.models.detr.modeling_detr.DetrConvEncoder.forward with Detr->DeformableDetr + def forward(self, pixel_values: torch.Tensor, pixel_mask: torch.Tensor): + # send pixel_values through the model to get list of feature maps + features = self.model(pixel_values) if self.config.use_timm_backbone else self.model(pixel_values).feature_maps + + out = [] + for feature_map in features: + # downsample pixel_mask to match shape of corresponding feature_map + mask = nn.functional.interpolate(pixel_mask[None].float(), size=feature_map.shape[-2:]).to(torch.bool)[0] + out.append((feature_map, mask)) + return out + + +# Copied from transformers.models.detr.modeling_detr.DetrConvModel with Detr->DeformableDetr +class DeformableDetrConvModel(nn.Module): + """ + This module adds 2D position embeddings to all intermediate feature maps of the convolutional encoder. + """ + + def __init__(self, conv_encoder, position_embedding): + super().__init__() + self.conv_encoder = conv_encoder + self.position_embedding = position_embedding + + def forward(self, pixel_values, pixel_mask): + # send pixel_values and pixel_mask through backbone to get list of (feature_map, pixel_mask) tuples + out = self.conv_encoder(pixel_values, pixel_mask) + pos = [] + for feature_map, mask in out: + # position encoding + pos.append(self.position_embedding(feature_map, mask).to(feature_map.dtype)) + + return out, pos + + +class DeformableDetrSinePositionEmbedding(nn.Module): + """ + This is a more standard version of the position embedding, very similar to the one used by the Attention is all you + need paper, generalized to work on images. + """ + + def __init__(self, embedding_dim=64, temperature=10000, normalize=False, scale=None): + super().__init__() + self.embedding_dim = embedding_dim + self.temperature = temperature + self.normalize = normalize + if scale is not None and normalize is False: + raise ValueError("normalize should be True if scale is passed") + if scale is None: + scale = 2 * math.pi + self.scale = scale + + def forward(self, pixel_values, pixel_mask): + if pixel_mask is None: + raise ValueError("No pixel mask provided") + y_embed = pixel_mask.cumsum(1, dtype=pixel_values.dtype) + x_embed = pixel_mask.cumsum(2, dtype=pixel_values.dtype) + if self.normalize: + eps = 1e-6 + y_embed = (y_embed - 0.5) / (y_embed[:, -1:, :] + eps) * self.scale + x_embed = (x_embed - 0.5) / (x_embed[:, :, -1:] + eps) * self.scale + + dim_t = torch.arange(self.embedding_dim, dtype=pixel_values.dtype, device=pixel_values.device) + dim_t = self.temperature ** (2 * torch.div(dim_t, 2, rounding_mode="floor") / self.embedding_dim) + + pos_x = x_embed[:, :, :, None] / dim_t + pos_y = y_embed[:, :, :, None] / dim_t + pos_x = torch.stack((pos_x[:, :, :, 0::2].sin(), pos_x[:, :, :, 1::2].cos()), dim=4).flatten(3) + pos_y = torch.stack((pos_y[:, :, :, 0::2].sin(), pos_y[:, :, :, 1::2].cos()), dim=4).flatten(3) + pos = torch.cat((pos_y, pos_x), dim=3).permute(0, 3, 1, 2) + return pos + + +# Copied from transformers.models.detr.modeling_detr.DetrLearnedPositionEmbedding +class DeformableDetrLearnedPositionEmbedding(nn.Module): + """ + This module learns positional embeddings up to a fixed maximum size. + """ + + def __init__(self, embedding_dim=256): + super().__init__() + self.row_embeddings = nn.Embedding(50, embedding_dim) + self.column_embeddings = nn.Embedding(50, embedding_dim) + + def forward(self, pixel_values, pixel_mask=None): + height, width = pixel_values.shape[-2:] + width_values = torch.arange(width, device=pixel_values.device) + height_values = torch.arange(height, device=pixel_values.device) + x_emb = self.column_embeddings(width_values) + y_emb = self.row_embeddings(height_values) + pos = torch.cat([x_emb.unsqueeze(0).repeat(height, 1, 1), y_emb.unsqueeze(1).repeat(1, width, 1)], dim=-1) + pos = pos.permute(2, 0, 1) + pos = pos.unsqueeze(0) + pos = pos.repeat(pixel_values.shape[0], 1, 1, 1) + return pos + + +# Copied from transformers.models.detr.modeling_detr.build_position_encoding with Detr->DeformableDetr +def build_position_encoding(config): + n_steps = config.d_model // 2 + if config.position_embedding_type == "sine": + # TODO find a better way of exposing other arguments + position_embedding = DeformableDetrSinePositionEmbedding(n_steps, normalize=True) + elif config.position_embedding_type == "learned": + position_embedding = DeformableDetrLearnedPositionEmbedding(n_steps) + else: + raise ValueError(f"Not supported {config.position_embedding_type}") + + return position_embedding + + +class DeformableDetrMultiscaleDeformableAttention(nn.Module): + """ + Multiscale deformable attention as proposed in Deformable DETR. + """ + + def __init__(self, config: DeformableDetrConfig, num_heads: int, n_points: int): + super().__init__() + + self.attn = MultiScaleDeformableAttention() + + if config.d_model % num_heads != 0: + raise ValueError( + f"embed_dim (d_model) must be divisible by num_heads, but got {config.d_model} and {num_heads}" + ) + dim_per_head = config.d_model // num_heads + # check if dim_per_head is power of 2 + if not ((dim_per_head & (dim_per_head - 1) == 0) and dim_per_head != 0): + warnings.warn( + "You'd better set embed_dim (d_model) in DeformableDetrMultiscaleDeformableAttention to make the" + " dimension of each attention head a power of 2 which is more efficient in the authors' CUDA" + " implementation." + ) + + self.im2col_step = 64 + + self.d_model = config.d_model + self.n_levels = config.num_feature_levels + self.n_heads = num_heads + self.n_points = n_points + + self.sampling_offsets = nn.Linear(config.d_model, num_heads * self.n_levels * n_points * 2) + self.attention_weights = nn.Linear(config.d_model, num_heads * self.n_levels * n_points) + self.value_proj = nn.Linear(config.d_model, config.d_model) + self.output_proj = nn.Linear(config.d_model, config.d_model) + + self.disable_custom_kernels = config.disable_custom_kernels + + def with_pos_embed(self, tensor: torch.Tensor, position_embeddings: Optional[Tensor]): + return tensor if position_embeddings is None else tensor + position_embeddings + + def forward( + self, + hidden_states: torch.Tensor, + attention_mask: Optional[torch.Tensor] = None, + encoder_hidden_states=None, + encoder_attention_mask=None, + position_embeddings: Optional[torch.Tensor] = None, + reference_points=None, + spatial_shapes=None, + spatial_shapes_list=None, + level_start_index=None, + output_attentions: bool = False, + ): + # add position embeddings to the hidden states before projecting to queries and keys + if position_embeddings is not None: + hidden_states = self.with_pos_embed(hidden_states, position_embeddings) + + batch_size, num_queries, _ = hidden_states.shape + batch_size, sequence_length, _ = encoder_hidden_states.shape + total_elements = sum(height * width for height, width in spatial_shapes_list) + if total_elements != sequence_length: + raise ValueError( + "Make sure to align the spatial shapes with the sequence length of the encoder hidden states" + ) + + value = self.value_proj(encoder_hidden_states) + if attention_mask is not None: + # we invert the attention_mask + value = value.masked_fill(~attention_mask[..., None], float(0)) + value = value.view(batch_size, sequence_length, self.n_heads, self.d_model // self.n_heads) + sampling_offsets = self.sampling_offsets(hidden_states).view( + batch_size, num_queries, self.n_heads, self.n_levels, self.n_points, 2 + ) + attention_weights = self.attention_weights(hidden_states).view( + batch_size, num_queries, self.n_heads, self.n_levels * self.n_points + ) + attention_weights = F.softmax(attention_weights, -1).view( + batch_size, num_queries, self.n_heads, self.n_levels, self.n_points + ) + # batch_size, num_queries, n_heads, n_levels, n_points, 2 + num_coordinates = reference_points.shape[-1] + if num_coordinates == 2: + offset_normalizer = torch.stack([spatial_shapes[..., 1], spatial_shapes[..., 0]], -1) + sampling_locations = ( + reference_points[:, :, None, :, None, :] + + sampling_offsets / offset_normalizer[None, None, None, :, None, :] + ) + elif num_coordinates == 4: + sampling_locations = ( + reference_points[:, :, None, :, None, :2] + + sampling_offsets / self.n_points * reference_points[:, :, None, :, None, 2:] * 0.5 + ) + else: + raise ValueError(f"Last dim of reference_points must be 2 or 4, but got {reference_points.shape[-1]}") + + output = self.attn( + value, + spatial_shapes, + spatial_shapes_list, + level_start_index, + sampling_locations, + attention_weights, + self.im2col_step, + ) + + output = self.output_proj(output) + + return output, attention_weights + + +class DeformableDetrMultiheadAttention(nn.Module): + """ + Multi-headed attention from 'Attention Is All You Need' paper. + + Here, we add position embeddings to the queries and keys (as explained in the Deformable DETR paper). + """ + + def __init__( + self, + embed_dim: int, + num_heads: int, + dropout: float = 0.0, + bias: bool = True, + ): + super().__init__() + self.embed_dim = embed_dim + self.num_heads = num_heads + self.dropout = dropout + self.head_dim = embed_dim // num_heads + if self.head_dim * num_heads != self.embed_dim: + raise ValueError( + f"embed_dim must be divisible by num_heads (got `embed_dim`: {self.embed_dim} and `num_heads`:" + f" {num_heads})." + ) + self.scaling = self.head_dim**-0.5 + + self.k_proj = nn.Linear(embed_dim, embed_dim, bias=bias) + self.v_proj = nn.Linear(embed_dim, embed_dim, bias=bias) + self.q_proj = nn.Linear(embed_dim, embed_dim, bias=bias) + self.out_proj = nn.Linear(embed_dim, embed_dim, bias=bias) + + def _shape(self, tensor: torch.Tensor, seq_len: int, batch_size: int): + return tensor.view(batch_size, seq_len, self.num_heads, self.head_dim).transpose(1, 2).contiguous() + + def with_pos_embed(self, tensor: torch.Tensor, position_embeddings: Optional[Tensor]): + return tensor if position_embeddings is None else tensor + position_embeddings + + def forward( + self, + hidden_states: torch.Tensor, + attention_mask: Optional[torch.Tensor] = None, + position_embeddings: Optional[torch.Tensor] = None, + output_attentions: bool = False, + ) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor]]]: + """Input shape: Batch x Time x Channel""" + + batch_size, target_len, embed_dim = hidden_states.size() + # add position embeddings to the hidden states before projecting to queries and keys + if position_embeddings is not None: + hidden_states_original = hidden_states + hidden_states = self.with_pos_embed(hidden_states, position_embeddings) + + # get queries, keys and values + query_states = self.q_proj(hidden_states) * self.scaling + key_states = self._shape(self.k_proj(hidden_states), -1, batch_size) + value_states = self._shape(self.v_proj(hidden_states_original), -1, batch_size) + + proj_shape = (batch_size * self.num_heads, -1, self.head_dim) + query_states = self._shape(query_states, target_len, batch_size).view(*proj_shape) + key_states = key_states.view(*proj_shape) + value_states = value_states.view(*proj_shape) + + source_len = key_states.size(1) + + attn_weights = torch.bmm(query_states, key_states.transpose(1, 2)) + + if attn_weights.size() != (batch_size * self.num_heads, target_len, source_len): + raise ValueError( + f"Attention weights should be of size {(batch_size * self.num_heads, target_len, source_len)}, but is" + f" {attn_weights.size()}" + ) + + # expand attention_mask + if attention_mask is not None: + # [batch_size, seq_len] -> [batch_size, 1, target_seq_len, source_seq_len] + attention_mask = _prepare_4d_attention_mask(attention_mask, hidden_states.dtype) + + if attention_mask is not None: + if attention_mask.size() != (batch_size, 1, target_len, source_len): + raise ValueError( + f"Attention mask should be of size {(batch_size, 1, target_len, source_len)}, but is" + f" {attention_mask.size()}" + ) + attn_weights = attn_weights.view(batch_size, self.num_heads, target_len, source_len) + attention_mask + attn_weights = attn_weights.view(batch_size * self.num_heads, target_len, source_len) + + attn_weights = nn.functional.softmax(attn_weights, dim=-1) + + if output_attentions: + # this operation is a bit awkward, but it's required to + # make sure that attn_weights keeps its gradient. + # In order to do so, attn_weights have to reshaped + # twice and have to be reused in the following + attn_weights_reshaped = attn_weights.view(batch_size, self.num_heads, target_len, source_len) + attn_weights = attn_weights_reshaped.view(batch_size * self.num_heads, target_len, source_len) + else: + attn_weights_reshaped = None + + attn_probs = nn.functional.dropout(attn_weights, p=self.dropout, training=self.training) + + attn_output = torch.bmm(attn_probs, value_states) + + if attn_output.size() != ( + batch_size * self.num_heads, + target_len, + self.head_dim, + ): + raise ValueError( + f"`attn_output` should be of size {(batch_size, self.num_heads, target_len, self.head_dim)}, but is" + f" {attn_output.size()}" + ) + + attn_output = attn_output.view(batch_size, self.num_heads, target_len, self.head_dim) + attn_output = attn_output.transpose(1, 2) + attn_output = attn_output.reshape(batch_size, target_len, embed_dim) + + attn_output = self.out_proj(attn_output) + + return attn_output, attn_weights_reshaped + + +class DeformableDetrEncoderLayer(nn.Module): + def __init__(self, config: DeformableDetrConfig): + super().__init__() + self.embed_dim = config.d_model + self.self_attn = DeformableDetrMultiscaleDeformableAttention( + config, + num_heads=config.encoder_attention_heads, + n_points=config.encoder_n_points, + ) + self.self_attn_layer_norm = nn.LayerNorm(self.embed_dim) + self.dropout = config.dropout + self.activation_fn = ACT2FN[config.activation_function] + self.activation_dropout = config.activation_dropout + self.fc1 = nn.Linear(self.embed_dim, config.encoder_ffn_dim) + self.fc2 = nn.Linear(config.encoder_ffn_dim, self.embed_dim) + self.final_layer_norm = nn.LayerNorm(self.embed_dim) + + def forward( + self, + hidden_states: torch.Tensor, + attention_mask: torch.Tensor, + position_embeddings: torch.Tensor = None, + reference_points=None, + spatial_shapes=None, + spatial_shapes_list=None, + level_start_index=None, + output_attentions: bool = False, + ): + """ + Args: + hidden_states (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`): + Input to the layer. + attention_mask (`torch.FloatTensor` of shape `(batch_size, sequence_length)`): + Attention mask. + position_embeddings (`torch.FloatTensor`, *optional*): + Position embeddings, to be added to `hidden_states`. + reference_points (`torch.FloatTensor`, *optional*): + Reference points. + spatial_shapes (`torch.LongTensor`, *optional*): + Spatial shapes of the backbone feature maps. + level_start_index (`torch.LongTensor`, *optional*): + Level start index. + output_attentions (`bool`, *optional*): + Whether or not to return the attentions tensors of all attention layers. See `attentions` under + returned tensors for more detail. + """ + residual = hidden_states + + # Apply Multi-scale Deformable Attention Module on the multi-scale feature maps. + hidden_states, attn_weights = self.self_attn( + hidden_states=hidden_states, + attention_mask=attention_mask, + encoder_hidden_states=hidden_states, + encoder_attention_mask=attention_mask, + position_embeddings=position_embeddings, + reference_points=reference_points, + spatial_shapes=spatial_shapes, + spatial_shapes_list=spatial_shapes_list, + level_start_index=level_start_index, + output_attentions=output_attentions, + ) + + hidden_states = nn.functional.dropout(hidden_states, p=self.dropout, training=self.training) + hidden_states = residual + hidden_states + hidden_states = self.self_attn_layer_norm(hidden_states) + + residual = hidden_states + hidden_states = self.activation_fn(self.fc1(hidden_states)) + hidden_states = nn.functional.dropout(hidden_states, p=self.activation_dropout, training=self.training) + + hidden_states = self.fc2(hidden_states) + hidden_states = nn.functional.dropout(hidden_states, p=self.dropout, training=self.training) + + hidden_states = residual + hidden_states + hidden_states = self.final_layer_norm(hidden_states) + + if self.training: + if torch.isinf(hidden_states).any() or torch.isnan(hidden_states).any(): + clamp_value = torch.finfo(hidden_states.dtype).max - 1000 + hidden_states = torch.clamp(hidden_states, min=-clamp_value, max=clamp_value) + + outputs = (hidden_states,) + + if output_attentions: + outputs += (attn_weights,) + + return outputs + + +class DeformableDetrDecoderLayer(nn.Module): + def __init__(self, config: DeformableDetrConfig): + super().__init__() + self.embed_dim = config.d_model + + # self-attention + self.self_attn = DeformableDetrMultiheadAttention( + embed_dim=self.embed_dim, + num_heads=config.decoder_attention_heads, + dropout=config.attention_dropout, + ) + self.dropout = config.dropout + self.activation_fn = ACT2FN[config.activation_function] + self.activation_dropout = config.activation_dropout + + self.self_attn_layer_norm = nn.LayerNorm(self.embed_dim) + # cross-attention + self.encoder_attn = DeformableDetrMultiscaleDeformableAttention( + config, + num_heads=config.decoder_attention_heads, + n_points=config.decoder_n_points, + ) + self.encoder_attn_layer_norm = nn.LayerNorm(self.embed_dim) + # feedforward neural networks + self.fc1 = nn.Linear(self.embed_dim, config.decoder_ffn_dim) + self.fc2 = nn.Linear(config.decoder_ffn_dim, self.embed_dim) + self.final_layer_norm = nn.LayerNorm(self.embed_dim) + + def forward( + self, + hidden_states: torch.Tensor, + position_embeddings: Optional[torch.Tensor] = None, + reference_points=None, + spatial_shapes=None, + spatial_shapes_list=None, + level_start_index=None, + encoder_hidden_states: Optional[torch.Tensor] = None, + encoder_attention_mask: Optional[torch.Tensor] = None, + output_attentions: Optional[bool] = False, + ): + """ + Args: + hidden_states (`torch.FloatTensor`): + Input to the layer of shape `(seq_len, batch, embed_dim)`. + position_embeddings (`torch.FloatTensor`, *optional*): + Position embeddings that are added to the queries and keys in the self-attention layer. + reference_points (`torch.FloatTensor`, *optional*): + Reference points. + spatial_shapes (`torch.LongTensor`, *optional*): + Spatial shapes. + level_start_index (`torch.LongTensor`, *optional*): + Level start index. + encoder_hidden_states (`torch.FloatTensor`): + cross attention input to the layer of shape `(seq_len, batch, embed_dim)` + encoder_attention_mask (`torch.FloatTensor`): encoder attention mask of size + `(batch, 1, target_len, source_len)` where padding elements are indicated by very large negative + values. + output_attentions (`bool`, *optional*): + Whether or not to return the attentions tensors of all attention layers. See `attentions` under + returned tensors for more detail. + """ + residual = hidden_states + + # Self Attention + hidden_states, self_attn_weights = self.self_attn( + hidden_states=hidden_states, + position_embeddings=position_embeddings, + output_attentions=output_attentions, + ) + + hidden_states = nn.functional.dropout(hidden_states, p=self.dropout, training=self.training) + hidden_states = residual + hidden_states + hidden_states = self.self_attn_layer_norm(hidden_states) + + second_residual = hidden_states + + # Cross-Attention + cross_attn_weights = None + hidden_states, cross_attn_weights = self.encoder_attn( + hidden_states=hidden_states, + attention_mask=encoder_attention_mask, + encoder_hidden_states=encoder_hidden_states, + encoder_attention_mask=encoder_attention_mask, + position_embeddings=position_embeddings, + reference_points=reference_points, + spatial_shapes=spatial_shapes, + spatial_shapes_list=spatial_shapes_list, + level_start_index=level_start_index, + output_attentions=output_attentions, + ) + + hidden_states = nn.functional.dropout(hidden_states, p=self.dropout, training=self.training) + hidden_states = second_residual + hidden_states + + hidden_states = self.encoder_attn_layer_norm(hidden_states) + + # Fully Connected + residual = hidden_states + hidden_states = self.activation_fn(self.fc1(hidden_states)) + hidden_states = nn.functional.dropout(hidden_states, p=self.activation_dropout, training=self.training) + hidden_states = self.fc2(hidden_states) + hidden_states = nn.functional.dropout(hidden_states, p=self.dropout, training=self.training) + hidden_states = residual + hidden_states + hidden_states = self.final_layer_norm(hidden_states) + + outputs = (hidden_states,) + + if output_attentions: + outputs += (self_attn_weights, cross_attn_weights) + + return outputs + + +class DeformableDetrPreTrainedModel(PreTrainedModel): + config_class = DeformableDetrConfig + base_model_prefix = "model" + main_input_name = "pixel_values" + supports_gradient_checkpointing = True + _no_split_modules = [ + r"DeformableDetrConvEncoder", + r"DeformableDetrEncoderLayer", + r"DeformableDetrDecoderLayer", + ] + + def _init_weights(self, module): + std = self.config.init_std + + if isinstance(module, DeformableDetrLearnedPositionEmbedding): + nn.init.uniform_(module.row_embeddings.weight) + nn.init.uniform_(module.column_embeddings.weight) + elif isinstance(module, DeformableDetrMultiscaleDeformableAttention): + nn.init.constant_(module.sampling_offsets.weight.data, 0.0) + default_dtype = torch.get_default_dtype() + thetas = torch.arange(module.n_heads, dtype=torch.int64).to(default_dtype) * ( + 2.0 * math.pi / module.n_heads + ) + grid_init = torch.stack([thetas.cos(), thetas.sin()], -1) + grid_init = ( + (grid_init / grid_init.abs().max(-1, keepdim=True)[0]) + .view(module.n_heads, 1, 1, 2) + .repeat(1, module.n_levels, module.n_points, 1) + ) + for i in range(module.n_points): + grid_init[:, :, i, :] *= i + 1 + with torch.no_grad(): + module.sampling_offsets.bias = nn.Parameter(grid_init.view(-1)) + nn.init.constant_(module.attention_weights.weight.data, 0.0) + nn.init.constant_(module.attention_weights.bias.data, 0.0) + nn.init.xavier_uniform_(module.value_proj.weight.data) + nn.init.constant_(module.value_proj.bias.data, 0.0) + nn.init.xavier_uniform_(module.output_proj.weight.data) + nn.init.constant_(module.output_proj.bias.data, 0.0) + elif isinstance(module, (nn.Linear, nn.Conv2d, nn.BatchNorm2d)): + # Slightly different from the TF version which uses truncated_normal for initialization + # cf https://github.com/pytorch/pytorch/pull/5617 + module.weight.data.normal_(mean=0.0, std=std) + if module.bias is not None: + module.bias.data.zero_() + elif isinstance(module, nn.Embedding): + module.weight.data.normal_(mean=0.0, std=std) + if module.padding_idx is not None: + module.weight.data[module.padding_idx].zero_() + if hasattr(module, "reference_points") and not self.config.two_stage: + nn.init.xavier_uniform_(module.reference_points.weight.data, gain=1.0) + nn.init.constant_(module.reference_points.bias.data, 0.0) + if hasattr(module, "level_embed"): + nn.init.normal_(module.level_embed) + + +DEFORMABLE_DETR_START_DOCSTRING = r""" + This model inherits from [`PreTrainedModel`]. Check the superclass documentation for the generic methods the + library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads + etc.) + + This model is also a PyTorch [torch.nn.Module](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) subclass. + Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage + and behavior. + + Parameters: + config ([`DeformableDetrConfig`]): + Model configuration class with all the parameters of the model. Initializing with a config file does not + load the weights associated with the model, only the configuration. Check out the + [`~PreTrainedModel.from_pretrained`] method to load the model weights. +""" + +DEFORMABLE_DETR_INPUTS_DOCSTRING = r""" + Args: + pixel_values (`torch.FloatTensor` of shape `(batch_size, num_channels, height, width)`): + Pixel values. Padding will be ignored by default should you provide it. + + Pixel values can be obtained using [`AutoImageProcessor`]. See [`DeformableDetrImageProcessor.__call__`] + for details. + + pixel_mask (`torch.LongTensor` of shape `(batch_size, height, width)`, *optional*): + Mask to avoid performing attention on padding pixel values. Mask values selected in `[0, 1]`: + + - 1 for pixels that are real (i.e. **not masked**), + - 0 for pixels that are padding (i.e. **masked**). + + [What are attention masks?](../glossary#attention-mask) + + decoder_attention_mask (`torch.FloatTensor` of shape `(batch_size, num_queries)`, *optional*): + Not used by default. Can be used to mask object queries. + encoder_outputs (`tuple(tuple(torch.FloatTensor)`, *optional*): + Tuple consists of (`last_hidden_state`, *optional*: `hidden_states`, *optional*: `attentions`) + `last_hidden_state` of shape `(batch_size, sequence_length, hidden_size)`, *optional*) is a sequence of + hidden-states at the output of the last layer of the encoder. Used in the cross-attention of the decoder. + inputs_embeds (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`, *optional*): + Optionally, instead of passing the flattened feature map (output of the backbone + projection layer), you + can choose to directly pass a flattened representation of an image. + decoder_inputs_embeds (`torch.FloatTensor` of shape `(batch_size, num_queries, hidden_size)`, *optional*): + Optionally, instead of initializing the queries with a tensor of zeros, you can choose to directly pass an + embedded representation. + output_attentions (`bool`, *optional*): + Whether or not to return the attentions tensors of all attention layers. See `attentions` under returned + tensors for more detail. + output_hidden_states (`bool`, *optional*): + Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for + more detail. + return_dict (`bool`, *optional*): + Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple. +""" + + +class DeformableDetrEncoder(DeformableDetrPreTrainedModel): + """ + Transformer encoder consisting of *config.encoder_layers* deformable attention layers. Each layer is a + [`DeformableDetrEncoderLayer`]. + + The encoder updates the flattened multi-scale feature maps through multiple deformable attention layers. + + Args: + config: DeformableDetrConfig + """ + + def __init__(self, config: DeformableDetrConfig): + super().__init__(config) + self.gradient_checkpointing = False + + self.dropout = config.dropout + self.layers = nn.ModuleList([DeformableDetrEncoderLayer(config) for _ in range(config.encoder_layers)]) + + # Initialize weights and apply final processing + self.post_init() + + @staticmethod + def get_reference_points(spatial_shapes, valid_ratios, device): + """ + Get reference points for each feature map. Used in decoder. + + Args: + spatial_shapes (`torch.LongTensor` of shape `(num_feature_levels, 2)`): + Spatial shapes of each feature map. + valid_ratios (`torch.FloatTensor` of shape `(batch_size, num_feature_levels, 2)`): + Valid ratios of each feature map. + device (`torch.device`): + Device on which to create the tensors. + Returns: + `torch.FloatTensor` of shape `(batch_size, num_queries, num_feature_levels, 2)` + """ + reference_points_list = [] + for level, (height, width) in enumerate(spatial_shapes): + ref_y, ref_x = meshgrid( + torch.linspace(0.5, height - 0.5, height, dtype=valid_ratios.dtype, device=device), + torch.linspace(0.5, width - 0.5, width, dtype=valid_ratios.dtype, device=device), + indexing="ij", + ) + # TODO: valid_ratios could be useless here. check https://github.com/fundamentalvision/Deformable-DETR/issues/36 + ref_y = ref_y.reshape(-1)[None] / (valid_ratios[:, None, level, 1] * height) + ref_x = ref_x.reshape(-1)[None] / (valid_ratios[:, None, level, 0] * width) + ref = torch.stack((ref_x, ref_y), -1) + reference_points_list.append(ref) + reference_points = torch.cat(reference_points_list, 1) + reference_points = reference_points[:, :, None] * valid_ratios[:, None] + return reference_points + + def forward( + self, + inputs_embeds=None, + attention_mask=None, + position_embeddings=None, + spatial_shapes=None, + spatial_shapes_list=None, + level_start_index=None, + valid_ratios=None, + output_attentions=None, + output_hidden_states=None, + return_dict=None, + ): + r""" + Args: + inputs_embeds (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`): + Flattened feature map (output of the backbone + projection layer) that is passed to the encoder. + attention_mask (`torch.Tensor` of shape `(batch_size, sequence_length)`, *optional*): + Mask to avoid performing attention on padding pixel features. Mask values selected in `[0, 1]`: + - 1 for pixel features that are real (i.e. **not masked**), + - 0 for pixel features that are padding (i.e. **masked**). + [What are attention masks?](../glossary#attention-mask) + position_embeddings (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`): + Position embeddings that are added to the queries and keys in each self-attention layer. + spatial_shapes (`torch.LongTensor` of shape `(num_feature_levels, 2)`): + Spatial shapes of each feature map. + level_start_index (`torch.LongTensor` of shape `(num_feature_levels)`): + Starting index of each feature map. + valid_ratios (`torch.FloatTensor` of shape `(batch_size, num_feature_levels, 2)`): + Ratio of valid area in each feature level. + output_attentions (`bool`, *optional*): + Whether or not to return the attentions tensors of all attention layers. See `attentions` under + returned tensors for more detail. + output_hidden_states (`bool`, *optional*): + Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors + for more detail. + return_dict (`bool`, *optional*): + Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple. + """ + output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions + output_hidden_states = ( + output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states + ) + return_dict = return_dict if return_dict is not None else self.config.use_return_dict + + hidden_states = inputs_embeds + hidden_states = nn.functional.dropout(hidden_states, p=self.dropout, training=self.training) + + spatial_shapes_tuple = tuple(spatial_shapes_list) + reference_points = self.get_reference_points(spatial_shapes_tuple, valid_ratios, device=inputs_embeds.device) + + encoder_states = () if output_hidden_states else None + all_attentions = () if output_attentions else None + for i, encoder_layer in enumerate(self.layers): + if output_hidden_states: + encoder_states = encoder_states + (hidden_states,) + if self.gradient_checkpointing and self.training: + layer_outputs = self._gradient_checkpointing_func( + encoder_layer.__call__, + hidden_states, + attention_mask, + position_embeddings, + reference_points, + spatial_shapes, + spatial_shapes_list, + level_start_index, + output_attentions, + ) + else: + layer_outputs = encoder_layer( + hidden_states, + attention_mask, + position_embeddings=position_embeddings, + reference_points=reference_points, + spatial_shapes=spatial_shapes, + spatial_shapes_list=spatial_shapes_list, + level_start_index=level_start_index, + output_attentions=output_attentions, + ) + + hidden_states = layer_outputs[0] + + if output_attentions: + all_attentions = all_attentions + (layer_outputs[1],) + + if output_hidden_states: + encoder_states = encoder_states + (hidden_states,) + + if not return_dict: + return tuple(v for v in [hidden_states, encoder_states, all_attentions] if v is not None) + return BaseModelOutput( + last_hidden_state=hidden_states, + hidden_states=encoder_states, + attentions=all_attentions, + ) + + +class DeformableDetrDecoder(DeformableDetrPreTrainedModel): + """ + Transformer decoder consisting of *config.decoder_layers* layers. Each layer is a [`DeformableDetrDecoderLayer`]. + + The decoder updates the query embeddings through multiple self-attention and cross-attention layers. + + Some tweaks for Deformable DETR: + + - `position_embeddings`, `reference_points`, `spatial_shapes` and `valid_ratios` are added to the forward pass. + - it also returns a stack of intermediate outputs and reference points from all decoding layers. + + Args: + config: DeformableDetrConfig + """ + + def __init__(self, config: DeformableDetrConfig): + super().__init__(config) + + self.dropout = config.dropout + self.layers = nn.ModuleList([DeformableDetrDecoderLayer(config) for _ in range(config.decoder_layers)]) + self.gradient_checkpointing = False + + # hack implementation for iterative bounding box refinement and two-stage Deformable DETR + self.bbox_embed = None + self.class_embed = None + + # Initialize weights and apply final processing + self.post_init() + + def forward( + self, + inputs_embeds=None, + encoder_hidden_states=None, + encoder_attention_mask=None, + position_embeddings=None, + reference_points=None, + spatial_shapes=None, + spatial_shapes_list=None, + level_start_index=None, + valid_ratios=None, + output_attentions=None, + output_hidden_states=None, + return_dict=None, + ): + r""" + Args: + inputs_embeds (`torch.FloatTensor` of shape `(batch_size, num_queries, hidden_size)`): + The query embeddings that are passed into the decoder. + encoder_hidden_states (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`, *optional*): + Sequence of hidden-states at the output of the last layer of the encoder. Used in the cross-attention + of the decoder. + encoder_attention_mask (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*): + Mask to avoid performing cross-attention on padding pixel_values of the encoder. Mask values selected + in `[0, 1]`: + - 1 for pixels that are real (i.e. **not masked**), + - 0 for pixels that are padding (i.e. **masked**). + position_embeddings (`torch.FloatTensor` of shape `(batch_size, num_queries, hidden_size)`, *optional*): + Position embeddings that are added to the queries and keys in each self-attention layer. + reference_points (`torch.FloatTensor` of shape `(batch_size, num_queries, 4)` is `as_two_stage` else `(batch_size, num_queries, 2)` or , *optional*): + Reference point in range `[0, 1]`, top-left (0,0), bottom-right (1, 1), including padding area. + spatial_shapes (`torch.FloatTensor` of shape `(num_feature_levels, 2)`): + Spatial shapes of the feature maps. + level_start_index (`torch.LongTensor` of shape `(num_feature_levels)`, *optional*): + Indexes for the start of each feature level. In range `[0, sequence_length]`. + valid_ratios (`torch.FloatTensor` of shape `(batch_size, num_feature_levels, 2)`, *optional*): + Ratio of valid area in each feature level. + + output_attentions (`bool`, *optional*): + Whether or not to return the attentions tensors of all attention layers. See `attentions` under + returned tensors for more detail. + output_hidden_states (`bool`, *optional*): + Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors + for more detail. + return_dict (`bool`, *optional*): + Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple. + """ + output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions + output_hidden_states = ( + output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states + ) + return_dict = return_dict if return_dict is not None else self.config.use_return_dict + + if inputs_embeds is not None: + hidden_states = inputs_embeds + + # decoder layers + all_hidden_states = () if output_hidden_states else None + all_self_attns = () if output_attentions else None + all_cross_attentions = () if (output_attentions and encoder_hidden_states is not None) else None + intermediate = () + intermediate_reference_points = () + + for idx, decoder_layer in enumerate(self.layers): + num_coordinates = reference_points.shape[-1] + if num_coordinates == 4: + reference_points_input = ( + reference_points[:, :, None] * torch.cat([valid_ratios, valid_ratios], -1)[:, None] + ) + elif reference_points.shape[-1] == 2: + reference_points_input = reference_points[:, :, None] * valid_ratios[:, None] + else: + raise ValueError("Reference points' last dimension must be of size 2") + + if output_hidden_states: + all_hidden_states += (hidden_states,) + + if self.gradient_checkpointing and self.training: + layer_outputs = self._gradient_checkpointing_func( + decoder_layer.__call__, + hidden_states, + position_embeddings, + reference_points_input, + spatial_shapes, + spatial_shapes_list, + level_start_index, + encoder_hidden_states, + encoder_attention_mask, + output_attentions, + ) + else: + layer_outputs = decoder_layer( + hidden_states, + position_embeddings=position_embeddings, + encoder_hidden_states=encoder_hidden_states, + reference_points=reference_points_input, + spatial_shapes=spatial_shapes, + spatial_shapes_list=spatial_shapes_list, + level_start_index=level_start_index, + encoder_attention_mask=encoder_attention_mask, + output_attentions=output_attentions, + ) + + hidden_states = layer_outputs[0] + + # hack implementation for iterative bounding box refinement + if self.bbox_embed is not None: + tmp = self.bbox_embed[idx](hidden_states) + num_coordinates = reference_points.shape[-1] + if num_coordinates == 4: + new_reference_points = tmp + inverse_sigmoid(reference_points) + new_reference_points = new_reference_points.sigmoid() + elif num_coordinates == 2: + new_reference_points = tmp + new_reference_points[..., :2] = tmp[..., :2] + inverse_sigmoid(reference_points) + new_reference_points = new_reference_points.sigmoid() + else: + raise ValueError( + f"Last dim of reference_points must be 2 or 4, but got {reference_points.shape[-1]}" + ) + reference_points = new_reference_points.detach() + + intermediate += (hidden_states,) + intermediate_reference_points += (reference_points,) + + if output_attentions: + all_self_attns += (layer_outputs[1],) + + if encoder_hidden_states is not None: + all_cross_attentions += (layer_outputs[2],) + + # Keep batch_size as first dimension + intermediate = torch.stack(intermediate, dim=1) + intermediate_reference_points = torch.stack(intermediate_reference_points, dim=1) + + # add hidden states from the last decoder layer + if output_hidden_states: + all_hidden_states += (hidden_states,) + + if not return_dict: + return tuple( + v + for v in [ + hidden_states, + intermediate, + intermediate_reference_points, + all_hidden_states, + all_self_attns, + all_cross_attentions, + ] + if v is not None + ) + return DeformableDetrDecoderOutput( + last_hidden_state=hidden_states, + intermediate_hidden_states=intermediate, + intermediate_reference_points=intermediate_reference_points, + hidden_states=all_hidden_states, + attentions=all_self_attns, + cross_attentions=all_cross_attentions, + ) + + +@add_start_docstrings( + """ + The bare Deformable DETR Model (consisting of a backbone and encoder-decoder Transformer) outputting raw + hidden-states without any specific head on top. + """, + DEFORMABLE_DETR_START_DOCSTRING, +) +class DeformableDetrModel(DeformableDetrPreTrainedModel): + def __init__(self, config: DeformableDetrConfig): + super().__init__(config) + + # Create backbone + positional encoding + backbone = DeformableDetrConvEncoder(config) + position_embeddings = build_position_encoding(config) + self.backbone = DeformableDetrConvModel(backbone, position_embeddings) + + # Create input projection layers + if config.num_feature_levels > 1: + num_backbone_outs = len(backbone.intermediate_channel_sizes) + input_proj_list = [] + for _ in range(num_backbone_outs): + in_channels = backbone.intermediate_channel_sizes[_] + input_proj_list.append( + nn.Sequential( + nn.Conv2d(in_channels, config.d_model, kernel_size=1), + nn.GroupNorm(32, config.d_model), + ) + ) + for _ in range(config.num_feature_levels - num_backbone_outs): + input_proj_list.append( + nn.Sequential( + nn.Conv2d( + in_channels, + config.d_model, + kernel_size=3, + stride=2, + padding=1, + ), + nn.GroupNorm(32, config.d_model), + ) + ) + in_channels = config.d_model + self.input_proj = nn.ModuleList(input_proj_list) + else: + self.input_proj = nn.ModuleList( + [ + nn.Sequential( + nn.Conv2d( + backbone.intermediate_channel_sizes[-1], + config.d_model, + kernel_size=1, + ), + nn.GroupNorm(32, config.d_model), + ) + ] + ) + + if not config.two_stage: + self.query_position_embeddings = nn.Embedding(config.num_queries, config.d_model * 2) + + self.encoder = DeformableDetrEncoder(config) + self.decoder = DeformableDetrDecoder(config) + + self.level_embed = nn.Parameter(torch.Tensor(config.num_feature_levels, config.d_model)) + + if config.two_stage: + self.enc_output = nn.Linear(config.d_model, config.d_model) + self.enc_output_norm = nn.LayerNorm(config.d_model) + self.pos_trans = nn.Linear(config.d_model * 2, config.d_model * 2) + self.pos_trans_norm = nn.LayerNorm(config.d_model * 2) + else: + self.reference_points = nn.Linear(config.d_model, 2) + + self.post_init() + + def get_encoder(self): + return self.encoder + + def get_decoder(self): + return self.decoder + + def freeze_backbone(self): + for name, param in self.backbone.conv_encoder.model.named_parameters(): + param.requires_grad_(False) + + def unfreeze_backbone(self): + for name, param in self.backbone.conv_encoder.model.named_parameters(): + param.requires_grad_(True) + + def get_valid_ratio(self, mask, dtype=torch.float32): + """Get the valid ratio of all feature maps.""" + + _, height, width = mask.shape + valid_height = torch.sum(mask[:, :, 0], 1) + valid_width = torch.sum(mask[:, 0, :], 1) + valid_ratio_height = valid_height.to(dtype) / height + valid_ratio_width = valid_width.to(dtype) / width + valid_ratio = torch.stack([valid_ratio_width, valid_ratio_height], -1) + return valid_ratio + + def get_proposal_pos_embed(self, proposals): + """Get the position embedding of the proposals.""" + + num_pos_feats = self.config.d_model // 2 + temperature = 10000 + scale = 2 * math.pi + + dim_t = torch.arange(num_pos_feats, dtype=proposals.dtype, device=proposals.device) + dim_t = temperature ** (2 * torch.div(dim_t, 2, rounding_mode="floor") / num_pos_feats) + # batch_size, num_queries, 4 + proposals = proposals.sigmoid() * scale + # batch_size, num_queries, 4, 128 + pos = proposals[:, :, :, None] / dim_t + # batch_size, num_queries, 4, 64, 2 -> batch_size, num_queries, 512 + pos = torch.stack((pos[:, :, :, 0::2].sin(), pos[:, :, :, 1::2].cos()), dim=4).flatten(2) + return pos + + def gen_encoder_output_proposals(self, enc_output, padding_mask, spatial_shapes): + """Generate the encoder output proposals from encoded enc_output. + + Args: + enc_output (Tensor[batch_size, sequence_length, hidden_size]): Output of the encoder. + padding_mask (Tensor[batch_size, sequence_length]): Padding mask for `enc_output`. + spatial_shapes (List[Tuple[int, int]]): Spatial shapes of the feature maps. + + Returns: + `tuple(torch.FloatTensor)`: A tuple of feature map and bbox prediction. + - object_query (Tensor[batch_size, sequence_length, hidden_size]): Object query features. Later used to + directly predict a bounding box. (without the need of a decoder) + - output_proposals (Tensor[batch_size, sequence_length, 4]): Normalized proposals, after an inverse + sigmoid. + """ + batch_size = enc_output.shape[0] + proposals = [] + _cur = 0 + for level, (height, width) in enumerate(spatial_shapes): + mask_flatten_ = padding_mask[:, _cur : (_cur + height * width)].view(batch_size, height, width, 1) + valid_height = torch.sum(~mask_flatten_[:, :, 0, 0], 1) + valid_width = torch.sum(~mask_flatten_[:, 0, :, 0], 1) + + grid_y, grid_x = meshgrid( + torch.linspace( + 0, + height - 1, + height, + dtype=enc_output.dtype, + device=enc_output.device, + ), + torch.linspace( + 0, + width - 1, + width, + dtype=enc_output.dtype, + device=enc_output.device, + ), + indexing="ij", + ) + grid = torch.cat([grid_x.unsqueeze(-1), grid_y.unsqueeze(-1)], -1) + + scale = torch.cat([valid_width.unsqueeze(-1), valid_height.unsqueeze(-1)], 1).view(batch_size, 1, 1, 2) + grid = (grid.unsqueeze(0).expand(batch_size, -1, -1, -1) + 0.5) / scale + width_heigth = torch.ones_like(grid) * 0.05 * (2.0**level) + proposal = torch.cat((grid, width_heigth), -1).view(batch_size, -1, 4) + proposals.append(proposal) + _cur += height * width + output_proposals = torch.cat(proposals, 1) + output_proposals_valid = ((output_proposals > 0.01) & (output_proposals < 0.99)).all(-1, keepdim=True) + output_proposals = torch.log(output_proposals / (1 - output_proposals)) # inverse sigmoid + output_proposals = output_proposals.masked_fill(padding_mask.unsqueeze(-1), float("inf")) + output_proposals = output_proposals.masked_fill(~output_proposals_valid, float("inf")) + + # assign each pixel as an object query + object_query = enc_output + object_query = object_query.masked_fill(padding_mask.unsqueeze(-1), float(0)) + object_query = object_query.masked_fill(~output_proposals_valid, float(0)) + object_query = self.enc_output_norm(self.enc_output(object_query)) + return object_query, output_proposals + + @add_start_docstrings_to_model_forward(DEFORMABLE_DETR_INPUTS_DOCSTRING) + @replace_return_docstrings(output_type=DeformableDetrModelOutput, config_class=_CONFIG_FOR_DOC) + def forward( + self, + pixel_values: torch.FloatTensor, + pixel_mask: Optional[torch.LongTensor] = None, + decoder_attention_mask: Optional[torch.FloatTensor] = None, + encoder_outputs: Optional[torch.FloatTensor] = None, + inputs_embeds: Optional[torch.FloatTensor] = None, + decoder_inputs_embeds: Optional[torch.FloatTensor] = None, + output_attentions: Optional[bool] = None, + output_hidden_states: Optional[bool] = None, + return_dict: Optional[bool] = None, + ) -> Union[Tuple[torch.FloatTensor], DeformableDetrModelOutput]: + r""" + Returns: + + Examples: + + ```python + >>> from transformers import AutoImageProcessor, DeformableDetrModel + >>> from PIL import Image + >>> import requests + + >>> url = "http://images.cocodataset.org/val2017/000000039769.jpg" + >>> image = Image.open(requests.get(url, stream=True).raw) + + >>> image_processor = AutoImageProcessor.from_pretrained("SenseTime/deformable-detr") + >>> model = DeformableDetrModel.from_pretrained("SenseTime/deformable-detr") + + >>> inputs = image_processor(images=image, return_tensors="pt") + + >>> outputs = model(**inputs) + + >>> last_hidden_states = outputs.last_hidden_state + >>> list(last_hidden_states.shape) + [1, 300, 256] + ```""" + output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions + output_hidden_states = ( + output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states + ) + return_dict = return_dict if return_dict is not None else self.config.use_return_dict + + batch_size, num_channels, height, width = pixel_values.shape + device = pixel_values.device + + if pixel_mask is None: + pixel_mask = torch.ones(((batch_size, height, width)), dtype=torch.long, device=device) + + # Extract multi-scale feature maps of same resolution `config.d_model` (cf Figure 4 in paper) + # First, sent pixel_values + pixel_mask through Backbone to obtain the features + # which is a list of tuples + features, position_embeddings_list = self.backbone(pixel_values, pixel_mask) + + # Then, apply 1x1 convolution to reduce the channel dimension to d_model (256 by default) + sources = [] + masks = [] + for level, (source, mask) in enumerate(features): + sources.append(self.input_proj[level](source)) + masks.append(mask) + if mask is None: + raise ValueError("No attention mask was provided") + + # Lowest resolution feature maps are obtained via 3x3 stride 2 convolutions on the final stage + if self.config.num_feature_levels > len(sources): + _len_sources = len(sources) + for level in range(_len_sources, self.config.num_feature_levels): + if level == _len_sources: + source = self.input_proj[level](features[-1][0]) + else: + source = self.input_proj[level](sources[-1]) + mask = nn.functional.interpolate(pixel_mask[None].to(pixel_values.dtype), size=source.shape[-2:]).to( + torch.bool + )[0] + pos_l = self.backbone.position_embedding(source, mask).to(source.dtype) + sources.append(source) + masks.append(mask) + position_embeddings_list.append(pos_l) + + # Create queries + query_embeds = None + if not self.config.two_stage: + query_embeds = self.query_position_embeddings.weight + + # Prepare encoder inputs (by flattening) + source_flatten = [] + mask_flatten = [] + lvl_pos_embed_flatten = [] + spatial_shapes_list = [] + for level, (source, mask, pos_embed) in enumerate(zip(sources, masks, position_embeddings_list)): + batch_size, num_channels, height, width = source.shape + spatial_shape = (height, width) + spatial_shapes_list.append(spatial_shape) + source = source.flatten(2).transpose(1, 2) + mask = mask.flatten(1) + pos_embed = pos_embed.flatten(2).transpose(1, 2) + lvl_pos_embed = pos_embed + self.level_embed[level].view(1, 1, -1) + lvl_pos_embed_flatten.append(lvl_pos_embed) + source_flatten.append(source) + mask_flatten.append(mask) + source_flatten = torch.cat(source_flatten, 1) + mask_flatten = torch.cat(mask_flatten, 1) + lvl_pos_embed_flatten = torch.cat(lvl_pos_embed_flatten, 1) + spatial_shapes = torch.as_tensor(spatial_shapes_list, dtype=torch.long, device=source_flatten.device) + level_start_index = torch.cat((spatial_shapes.new_zeros((1,)), spatial_shapes.prod(1).cumsum(0)[:-1])) + valid_ratios = torch.stack([self.get_valid_ratio(m, dtype=source_flatten.dtype) for m in masks], 1) + + # Fourth, sent source_flatten + mask_flatten + lvl_pos_embed_flatten (backbone + proj layer output) through encoder + # Also provide spatial_shapes, level_start_index and valid_ratios + if encoder_outputs is None: + encoder_outputs = self.encoder( + inputs_embeds=source_flatten, + attention_mask=mask_flatten, + position_embeddings=lvl_pos_embed_flatten, + spatial_shapes=spatial_shapes, + spatial_shapes_list=spatial_shapes_list, + level_start_index=level_start_index, + valid_ratios=valid_ratios, + output_attentions=output_attentions, + output_hidden_states=output_hidden_states, + return_dict=return_dict, + ) + # If the user passed a tuple for encoder_outputs, we wrap it in a BaseModelOutput when return_dict=True + elif return_dict and not isinstance(encoder_outputs, BaseModelOutput): + encoder_outputs = BaseModelOutput( + last_hidden_state=encoder_outputs[0], + hidden_states=encoder_outputs[1] if len(encoder_outputs) > 1 else None, + attentions=encoder_outputs[2] if len(encoder_outputs) > 2 else None, + ) + + # Fifth, prepare decoder inputs + batch_size, _, num_channels = encoder_outputs[0].shape + enc_outputs_class = None + enc_outputs_coord_logits = None + if self.config.two_stage: + object_query_embedding, output_proposals = self.gen_encoder_output_proposals( + encoder_outputs[0], ~mask_flatten, spatial_shapes_list + ) + + # hack implementation for two-stage Deformable DETR + # apply a detection head to each pixel (A.4 in paper) + # linear projection for bounding box binary classification (i.e. foreground and background) + enc_outputs_class = self.decoder.class_embed[-1](object_query_embedding) + # 3-layer FFN to predict bounding boxes coordinates (bbox regression branch) + delta_bbox = self.decoder.bbox_embed[-1](object_query_embedding) + enc_outputs_coord_logits = delta_bbox + output_proposals + + # only keep top scoring `config.two_stage_num_proposals` proposals + topk = self.config.two_stage_num_proposals + topk_proposals = torch.topk(enc_outputs_class[..., 0], topk, dim=1)[1] + topk_coords_logits = torch.gather( + enc_outputs_coord_logits, + 1, + topk_proposals.unsqueeze(-1).repeat(1, 1, 4), + ) + + topk_coords_logits = topk_coords_logits.detach() + reference_points = topk_coords_logits.sigmoid() + init_reference_points = reference_points + pos_trans_out = self.pos_trans_norm(self.pos_trans(self.get_proposal_pos_embed(topk_coords_logits))) + query_embed, target = torch.split(pos_trans_out, num_channels, dim=2) + else: + query_embed, target = torch.split(query_embeds, num_channels, dim=1) + query_embed = query_embed.unsqueeze(0).expand(batch_size, -1, -1) + target = target.unsqueeze(0).expand(batch_size, -1, -1) + reference_points = self.reference_points(query_embed).sigmoid() + init_reference_points = reference_points + + decoder_outputs = self.decoder( + inputs_embeds=target, + position_embeddings=query_embed, + encoder_hidden_states=encoder_outputs[0], + encoder_attention_mask=mask_flatten, + reference_points=reference_points, + spatial_shapes=spatial_shapes, + spatial_shapes_list=spatial_shapes_list, + level_start_index=level_start_index, + valid_ratios=valid_ratios, + output_attentions=output_attentions, + output_hidden_states=output_hidden_states, + return_dict=return_dict, + ) + + if not return_dict: + enc_outputs = tuple(value for value in [enc_outputs_class, enc_outputs_coord_logits] if value is not None) + tuple_outputs = (init_reference_points,) + decoder_outputs + encoder_outputs + enc_outputs + + return tuple_outputs + + return DeformableDetrModelOutput( + init_reference_points=init_reference_points, + last_hidden_state=decoder_outputs.last_hidden_state, + intermediate_hidden_states=decoder_outputs.intermediate_hidden_states, + intermediate_reference_points=decoder_outputs.intermediate_reference_points, + decoder_hidden_states=decoder_outputs.hidden_states, + decoder_attentions=decoder_outputs.attentions, + cross_attentions=decoder_outputs.cross_attentions, + encoder_last_hidden_state=encoder_outputs.last_hidden_state, + encoder_hidden_states=encoder_outputs.hidden_states, + encoder_attentions=encoder_outputs.attentions, + enc_outputs_class=enc_outputs_class, + enc_outputs_coord_logits=enc_outputs_coord_logits, + ) + + +# Copied from transformers.models.detr.modeling_detr.DetrMLPPredictionHead +class DeformableDetrMLPPredictionHead(nn.Module): + """ + Very simple multi-layer perceptron (MLP, also called FFN), used to predict the normalized center coordinates, + height and width of a bounding box w.r.t. an image. + + Copied from https://github.com/facebookresearch/detr/blob/master/models/detr.py + + """ + + def __init__(self, input_dim, hidden_dim, output_dim, num_layers): + super().__init__() + self.num_layers = num_layers + h = [hidden_dim] * (num_layers - 1) + self.layers = nn.ModuleList(nn.Linear(n, k) for n, k in zip([input_dim] + h, h + [output_dim])) + + def forward(self, x): + for i, layer in enumerate(self.layers): + x = nn.functional.relu(layer(x)) if i < self.num_layers - 1 else layer(x) + return x + + +@add_start_docstrings( + """ + Deformable DETR Model (consisting of a backbone and encoder-decoder Transformer) with object detection heads on + top, for tasks such as COCO detection. + """, + DEFORMABLE_DETR_START_DOCSTRING, +) +class DeformableDetrForObjectDetection(DeformableDetrPreTrainedModel): + # When using clones, all layers > 0 will be clones, but layer 0 *is* required + _tied_weights_keys = [r"bbox_embed\.[1-9]\d*", r"class_embed\.[1-9]\d*"] + # We can't initialize the model on meta device as some weights are modified during the initialization + _no_split_modules = None + + def __init__(self, config: DeformableDetrConfig): + super().__init__(config) + + # Deformable DETR encoder-decoder model + self.model = DeformableDetrModel(config) + # Detection heads on top + self.class_embed = nn.Linear(config.d_model, config.num_labels) + self.bbox_embed = DeformableDetrMLPPredictionHead( + input_dim=config.d_model, + hidden_dim=config.d_model, + output_dim=4, + num_layers=3, + ) + + prior_prob = 0.01 + bias_value = -math.log((1 - prior_prob) / prior_prob) + self.class_embed.bias.data = torch.ones(config.num_labels) * bias_value + nn.init.constant_(self.bbox_embed.layers[-1].weight.data, 0) + nn.init.constant_(self.bbox_embed.layers[-1].bias.data, 0) + + # if two-stage, the last class_embed and bbox_embed is for region proposal generation + num_pred = (config.decoder_layers + 1) if config.two_stage else config.decoder_layers + if config.with_box_refine: + self.class_embed = _get_clones(self.class_embed, num_pred) + self.bbox_embed = _get_clones(self.bbox_embed, num_pred) + nn.init.constant_(self.bbox_embed[0].layers[-1].bias.data[2:], -2.0) + # hack implementation for iterative bounding box refinement + self.model.decoder.bbox_embed = self.bbox_embed + else: + nn.init.constant_(self.bbox_embed.layers[-1].bias.data[2:], -2.0) + self.class_embed = nn.ModuleList([self.class_embed for _ in range(num_pred)]) + self.bbox_embed = nn.ModuleList([self.bbox_embed for _ in range(num_pred)]) + self.model.decoder.bbox_embed = None + if config.two_stage: + # hack implementation for two-stage + self.model.decoder.class_embed = self.class_embed + for box_embed in self.bbox_embed: + nn.init.constant_(box_embed.layers[-1].bias.data[2:], 0.0) + + # Initialize weights and apply final processing + self.post_init() + + @add_start_docstrings_to_model_forward(DEFORMABLE_DETR_INPUTS_DOCSTRING) + @replace_return_docstrings(output_type=DeformableDetrObjectDetectionOutput, config_class=_CONFIG_FOR_DOC) + def forward( + self, + pixel_values: torch.FloatTensor, + pixel_mask: Optional[torch.LongTensor] = None, + decoder_attention_mask: Optional[torch.FloatTensor] = None, + encoder_outputs: Optional[torch.FloatTensor] = None, + inputs_embeds: Optional[torch.FloatTensor] = None, + decoder_inputs_embeds: Optional[torch.FloatTensor] = None, + labels: Optional[List[dict]] = None, + output_attentions: Optional[bool] = None, + output_hidden_states: Optional[bool] = None, + return_dict: Optional[bool] = None, + ) -> Union[Tuple[torch.FloatTensor], DeformableDetrObjectDetectionOutput]: + r""" + labels (`List[Dict]` of len `(batch_size,)`, *optional*): + Labels for computing the bipartite matching loss. List of dicts, each dictionary containing at least the + following 2 keys: 'class_labels' and 'boxes' (the class labels and bounding boxes of an image in the batch + respectively). The class labels themselves should be a `torch.LongTensor` of len `(number of bounding boxes + in the image,)` and the boxes a `torch.FloatTensor` of shape `(number of bounding boxes in the image, 4)`. + + Returns: + + Examples: + + ```python + >>> from transformers import AutoImageProcessor, DeformableDetrForObjectDetection + >>> from PIL import Image + >>> import requests + + >>> url = "http://images.cocodataset.org/val2017/000000039769.jpg" + >>> image = Image.open(requests.get(url, stream=True).raw) + + >>> image_processor = AutoImageProcessor.from_pretrained("SenseTime/deformable-detr") + >>> model = DeformableDetrForObjectDetection.from_pretrained("SenseTime/deformable-detr") + + >>> inputs = image_processor(images=image, return_tensors="pt") + >>> outputs = model(**inputs) + + >>> # convert outputs (bounding boxes and class logits) to Pascal VOC format (xmin, ymin, xmax, ymax) + >>> target_sizes = torch.tensor([image.size[::-1]]) + >>> results = image_processor.post_process_object_detection(outputs, threshold=0.5, target_sizes=target_sizes)[ + ... 0 + ... ] + >>> for score, label, box in zip(results["scores"], results["labels"], results["boxes"]): + ... box = [round(i, 2) for i in box.tolist()] + ... print( + ... f"Detected {model.config.id2label[label.item()]} with confidence " + ... f"{round(score.item(), 3)} at location {box}" + ... ) + Detected cat with confidence 0.8 at location [16.5, 52.84, 318.25, 470.78] + Detected cat with confidence 0.789 at location [342.19, 24.3, 640.02, 372.25] + Detected remote with confidence 0.633 at location [40.79, 72.78, 176.76, 117.25] + ```""" + return_dict = return_dict if return_dict is not None else self.config.use_return_dict + + # First, sent images through DETR base model to obtain encoder + decoder outputs + outputs = self.model( + pixel_values, + pixel_mask=pixel_mask, + decoder_attention_mask=decoder_attention_mask, + encoder_outputs=encoder_outputs, + inputs_embeds=inputs_embeds, + decoder_inputs_embeds=decoder_inputs_embeds, + output_attentions=output_attentions, + output_hidden_states=output_hidden_states, + return_dict=return_dict, + ) + + hidden_states = outputs.intermediate_hidden_states if return_dict else outputs[2] + init_reference = outputs.init_reference_points if return_dict else outputs[0] + inter_references = outputs.intermediate_reference_points if return_dict else outputs[3] + + # class logits + predicted bounding boxes + outputs_classes = [] + outputs_coords = [] + + for level in range(hidden_states.shape[1]): + if level == 0: + reference = init_reference + else: + reference = inter_references[:, level - 1] + reference = inverse_sigmoid(reference) + outputs_class = self.class_embed[level](hidden_states[:, level]) + delta_bbox = self.bbox_embed[level](hidden_states[:, level]) + if reference.shape[-1] == 4: + outputs_coord_logits = delta_bbox + reference + elif reference.shape[-1] == 2: + delta_bbox[..., :2] += reference + outputs_coord_logits = delta_bbox + else: + raise ValueError(f"reference.shape[-1] should be 4 or 2, but got {reference.shape[-1]}") + outputs_coord = outputs_coord_logits.sigmoid() + outputs_classes.append(outputs_class) + outputs_coords.append(outputs_coord) + outputs_class = torch.stack(outputs_classes) + outputs_coord = torch.stack(outputs_coords) + + logits = outputs_class[-1] + pred_boxes = outputs_coord[-1] + + loss, loss_dict, auxiliary_outputs = None, None, None + if labels is not None: + loss, loss_dict, auxiliary_outputs = self.loss_function( + logits, + labels, + self.device, + pred_boxes, + self.config, + outputs_class, + outputs_coord, + ) + if not return_dict: + if auxiliary_outputs is not None: + output = (logits, pred_boxes) + auxiliary_outputs + outputs + else: + output = (logits, pred_boxes) + outputs + tuple_outputs = ((loss, loss_dict) + output) if loss is not None else output + + return tuple_outputs + + dict_outputs = DeformableDetrObjectDetectionOutput( + loss=loss, + loss_dict=loss_dict, + logits=logits, + pred_boxes=pred_boxes, + auxiliary_outputs=auxiliary_outputs, + last_hidden_state=outputs.last_hidden_state, + decoder_hidden_states=outputs.decoder_hidden_states, + decoder_attentions=outputs.decoder_attentions, + cross_attentions=outputs.cross_attentions, + encoder_last_hidden_state=outputs.encoder_last_hidden_state, + encoder_hidden_states=outputs.encoder_hidden_states, + encoder_attentions=outputs.encoder_attentions, + intermediate_hidden_states=outputs.intermediate_hidden_states, + intermediate_reference_points=outputs.intermediate_reference_points, + init_reference_points=outputs.init_reference_points, + enc_outputs_class=outputs.enc_outputs_class, + enc_outputs_coord_logits=outputs.enc_outputs_coord_logits, + ) + + return dict_outputs + + +__all__ = [ + "DeformableDetrForObjectDetection", + "DeformableDetrModel", + "DeformableDetrPreTrainedModel", +] diff --git a/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/modular_deformable_detr.py b/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/modular_deformable_detr.py new file mode 100644 index 0000000000000000000000000000000000000000..5dfb6048c3431b151ae1cf47157abf7e917c536c --- /dev/null +++ b/vllm/lib/python3.10/site-packages/transformers/models/deformable_detr/modular_deformable_detr.py @@ -0,0 +1,144 @@ +from typing import List, Tuple, Union + +from transformers.models.detr.image_processing_detr_fast import DetrImageProcessorFast + +from ...image_transforms import center_to_corners_format +from ...utils import ( + TensorType, + is_torch_available, + logging, +) + + +if is_torch_available(): + import torch + + +logger = logging.get_logger(__name__) + + +class DeformableDetrImageProcessorFast(DetrImageProcessorFast): + def post_process(self, outputs, target_sizes): + """ + Converts the raw output of [`DeformableDetrForObjectDetection`] into final bounding boxes in (top_left_x, + top_left_y, bottom_right_x, bottom_right_y) format. Only supports PyTorch. + + Args: + outputs ([`DeformableDetrObjectDetectionOutput`]): + Raw outputs of the model. + target_sizes (`torch.Tensor` of shape `(batch_size, 2)`): + Tensor containing the size (height, width) of each image of the batch. For evaluation, this must be the + original image size (before any data augmentation). For visualization, this should be the image size + after data augment, but before padding. + Returns: + `List[Dict]`: A list of dictionaries, each dictionary containing the scores, labels and boxes for an image + in the batch as predicted by the model. + """ + logger.warning_once( + "`post_process` is deprecated and will be removed in v5 of Transformers, please use" + " `post_process_object_detection` instead, with `threshold=0.` for equivalent results.", + ) + + out_logits, out_bbox = outputs.logits, outputs.pred_boxes + + if len(out_logits) != len(target_sizes): + raise ValueError("Make sure that you pass in as many target sizes as the batch dimension of the logits") + if target_sizes.shape[1] != 2: + raise ValueError("Each element of target_sizes must contain the size (h, w) of each image of the batch") + + prob = out_logits.sigmoid() + topk_values, topk_indexes = torch.topk(prob.view(out_logits.shape[0], -1), 100, dim=1) + scores = topk_values + topk_boxes = torch.div(topk_indexes, out_logits.shape[2], rounding_mode="floor") + labels = topk_indexes % out_logits.shape[2] + boxes = center_to_corners_format(out_bbox) + boxes = torch.gather(boxes, 1, topk_boxes.unsqueeze(-1).repeat(1, 1, 4)) + + # and from relative [0, 1] to absolute [0, height] coordinates + img_h, img_w = target_sizes.unbind(1) + scale_fct = torch.stack([img_w, img_h, img_w, img_h], dim=1) + boxes = boxes * scale_fct[:, None, :] + + results = [{"scores": s, "labels": l, "boxes": b} for s, l, b in zip(scores, labels, boxes)] + + return results + + def post_process_object_detection( + self, outputs, threshold: float = 0.5, target_sizes: Union[TensorType, List[Tuple]] = None, top_k: int = 100 + ): + """ + Converts the raw output of [`DeformableDetrForObjectDetection`] into final bounding boxes in (top_left_x, + top_left_y, bottom_right_x, bottom_right_y) format. Only supports PyTorch. + + Args: + outputs ([`DetrObjectDetectionOutput`]): + Raw outputs of the model. + threshold (`float`, *optional*): + Score threshold to keep object detection predictions. + target_sizes (`torch.Tensor` or `List[Tuple[int, int]]`, *optional*): + Tensor of shape `(batch_size, 2)` or list of tuples (`Tuple[int, int]`) containing the target size + (height, width) of each image in the batch. If left to None, predictions will not be resized. + top_k (`int`, *optional*, defaults to 100): + Keep only top k bounding boxes before filtering by thresholding. + + Returns: + `List[Dict]`: A list of dictionaries, each dictionary containing the scores, labels and boxes for an image + in the batch as predicted by the model. + """ + out_logits, out_bbox = outputs.logits, outputs.pred_boxes + + if target_sizes is not None: + if len(out_logits) != len(target_sizes): + raise ValueError( + "Make sure that you pass in as many target sizes as the batch dimension of the logits" + ) + + prob = out_logits.sigmoid() + prob = prob.view(out_logits.shape[0], -1) + k_value = min(top_k, prob.size(1)) + topk_values, topk_indexes = torch.topk(prob, k_value, dim=1) + scores = topk_values + topk_boxes = torch.div(topk_indexes, out_logits.shape[2], rounding_mode="floor") + labels = topk_indexes % out_logits.shape[2] + boxes = center_to_corners_format(out_bbox) + boxes = torch.gather(boxes, 1, topk_boxes.unsqueeze(-1).repeat(1, 1, 4)) + + # and from relative [0, 1] to absolute [0, height] coordinates + if target_sizes is not None: + if isinstance(target_sizes, List): + img_h = torch.Tensor([i[0] for i in target_sizes]) + img_w = torch.Tensor([i[1] for i in target_sizes]) + else: + img_h, img_w = target_sizes.unbind(1) + scale_fct = torch.stack([img_w, img_h, img_w, img_h], dim=1).to(boxes.device) + boxes = boxes * scale_fct[:, None, :] + + results = [] + for s, l, b in zip(scores, labels, boxes): + score = s[s > threshold] + label = l[s > threshold] + box = b[s > threshold] + results.append({"scores": score, "labels": label, "boxes": box}) + + return results + + def post_process_segmentation(): + raise NotImplementedError("Segmentation post-processing is not implemented for Deformable DETR yet.") + + def post_process_instance(): + raise NotImplementedError("Instance post-processing is not implemented for Deformable DETR yet.") + + def post_process_panoptic(): + raise NotImplementedError("Panoptic post-processing is not implemented for Deformable DETR yet.") + + def post_process_instance_segmentation(): + raise NotImplementedError("Segmentation post-processing is not implemented for Deformable DETR yet.") + + def post_process_semantic_segmentation(): + raise NotImplementedError("Semantic segmentation post-processing is not implemented for Deformable DETR yet.") + + def post_process_panoptic_segmentation(): + raise NotImplementedError("Panoptic segmentation post-processing is not implemented for Deformable DETR yet.") + + +__all__ = ["DeformableDetrImageProcessorFast"] diff --git a/vllm/lib/python3.10/site-packages/transformers/models/mobilenet_v2/__init__.py b/vllm/lib/python3.10/site-packages/transformers/models/mobilenet_v2/__init__.py new file mode 100644 index 0000000000000000000000000000000000000000..c29b5fc245ef93341fe476e148f4957da3ff145a --- /dev/null +++ b/vllm/lib/python3.10/site-packages/transformers/models/mobilenet_v2/__init__.py @@ -0,0 +1,29 @@ +# Copyright 2024 The HuggingFace Team. All rights reserved. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +from typing import TYPE_CHECKING + +from ...utils import _LazyModule +from ...utils.import_utils import define_import_structure + + +if TYPE_CHECKING: + from .configuration_mobilenet_v2 import * + from .feature_extraction_mobilenet_v2 import * + from .image_processing_mobilenet_v2 import * + from .modeling_mobilenet_v2 import * +else: + import sys + + _file = globals()["__file__"] + sys.modules[__name__] = _LazyModule(__name__, _file, define_import_structure(_file), module_spec=__spec__) diff --git a/vllm/lib/python3.10/site-packages/transformers/models/mobilenet_v2/__pycache__/__init__.cpython-310.pyc b/vllm/lib/python3.10/site-packages/transformers/models/mobilenet_v2/__pycache__/__init__.cpython-310.pyc new file mode 100644 index 0000000000000000000000000000000000000000..c5a4ede62362c2451d06ebce6da5fbae9ceeda75 Binary files /dev/null and b/vllm/lib/python3.10/site-packages/transformers/models/mobilenet_v2/__pycache__/__init__.cpython-310.pyc differ diff --git a/vllm/lib/python3.10/site-packages/transformers/models/mobilenet_v2/__pycache__/configuration_mobilenet_v2.cpython-310.pyc b/vllm/lib/python3.10/site-packages/transformers/models/mobilenet_v2/__pycache__/configuration_mobilenet_v2.cpython-310.pyc new file mode 100644 index 0000000000000000000000000000000000000000..a891792b96cfbfd7ee432131005324bb482b3c62 Binary files /dev/null and b/vllm/lib/python3.10/site-packages/transformers/models/mobilenet_v2/__pycache__/configuration_mobilenet_v2.cpython-310.pyc differ diff --git a/vllm/lib/python3.10/site-packages/transformers/models/mobilenet_v2/__pycache__/feature_extraction_mobilenet_v2.cpython-310.pyc b/vllm/lib/python3.10/site-packages/transformers/models/mobilenet_v2/__pycache__/feature_extraction_mobilenet_v2.cpython-310.pyc new file mode 100644 index 0000000000000000000000000000000000000000..d737289550b3b19f2d04e92947d743e8d2d06ba1 Binary files /dev/null and b/vllm/lib/python3.10/site-packages/transformers/models/mobilenet_v2/__pycache__/feature_extraction_mobilenet_v2.cpython-310.pyc differ diff --git a/vllm/lib/python3.10/site-packages/transformers/models/mobilenet_v2/__pycache__/image_processing_mobilenet_v2.cpython-310.pyc b/vllm/lib/python3.10/site-packages/transformers/models/mobilenet_v2/__pycache__/image_processing_mobilenet_v2.cpython-310.pyc new file mode 100644 index 0000000000000000000000000000000000000000..3b8221a3acef2eb56a79d95eb4394d138db99ce1 Binary files /dev/null and b/vllm/lib/python3.10/site-packages/transformers/models/mobilenet_v2/__pycache__/image_processing_mobilenet_v2.cpython-310.pyc differ diff --git a/vllm/lib/python3.10/site-packages/transformers/models/mobilenet_v2/__pycache__/modeling_mobilenet_v2.cpython-310.pyc b/vllm/lib/python3.10/site-packages/transformers/models/mobilenet_v2/__pycache__/modeling_mobilenet_v2.cpython-310.pyc new file mode 100644 index 0000000000000000000000000000000000000000..98257f03e6853d86ae625703031dc4c34372911a Binary files /dev/null and b/vllm/lib/python3.10/site-packages/transformers/models/mobilenet_v2/__pycache__/modeling_mobilenet_v2.cpython-310.pyc differ diff --git a/vllm/lib/python3.10/site-packages/transformers/models/mobilenet_v2/configuration_mobilenet_v2.py b/vllm/lib/python3.10/site-packages/transformers/models/mobilenet_v2/configuration_mobilenet_v2.py new file mode 100644 index 0000000000000000000000000000000000000000..6cf64847e2cd644ebbde26bc972b9344705941e6 --- /dev/null +++ b/vllm/lib/python3.10/site-packages/transformers/models/mobilenet_v2/configuration_mobilenet_v2.py @@ -0,0 +1,154 @@ +# coding=utf-8 +# Copyright 2022 The HuggingFace Inc. team. All rights reserved. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +"""MobileNetV2 model configuration""" + +from collections import OrderedDict +from typing import Mapping + +from packaging import version + +from ...configuration_utils import PretrainedConfig +from ...onnx import OnnxConfig +from ...utils import logging + + +logger = logging.get_logger(__name__) + + +class MobileNetV2Config(PretrainedConfig): + r""" + This is the configuration class to store the configuration of a [`MobileNetV2Model`]. It is used to instantiate a + MobileNetV2 model according to the specified arguments, defining the model architecture. Instantiating a + configuration with the defaults will yield a similar configuration to that of the MobileNetV2 + [google/mobilenet_v2_1.0_224](https://huggingface.co/google/mobilenet_v2_1.0_224) architecture. + + Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the + documentation from [`PretrainedConfig`] for more information. + + Args: + num_channels (`int`, *optional*, defaults to 3): + The number of input channels. + image_size (`int`, *optional*, defaults to 224): + The size (resolution) of each image. + depth_multiplier (`float`, *optional*, defaults to 1.0): + Shrinks or expands the number of channels in each layer. Default is 1.0, which starts the network with 32 + channels. This is sometimes also called "alpha" or "width multiplier". + depth_divisible_by (`int`, *optional*, defaults to 8): + The number of channels in each layer will always be a multiple of this number. + min_depth (`int`, *optional*, defaults to 8): + All layers will have at least this many channels. + expand_ratio (`float`, *optional*, defaults to 6.0): + The number of output channels of the first layer in each block is input channels times expansion ratio. + output_stride (`int`, *optional*, defaults to 32): + The ratio between the spatial resolution of the input and output feature maps. By default the model reduces + the input dimensions by a factor of 32. If `output_stride` is 8 or 16, the model uses dilated convolutions + on the depthwise layers instead of regular convolutions, so that the feature maps never become more than 8x + or 16x smaller than the input image. + first_layer_is_expansion (`bool`, *optional*, defaults to `True`): + True if the very first convolution layer is also the expansion layer for the first expansion block. + finegrained_output (`bool`, *optional*, defaults to `True`): + If true, the number of output channels in the final convolution layer will stay large (1280) even if + `depth_multiplier` is less than 1. + hidden_act (`str` or `function`, *optional*, defaults to `"relu6"`): + The non-linear activation function (function or string) in the Transformer encoder and convolution layers. + tf_padding (`bool`, *optional*, defaults to `True`): + Whether to use TensorFlow padding rules on the convolution layers. + classifier_dropout_prob (`float`, *optional*, defaults to 0.8): + The dropout ratio for attached classifiers. + initializer_range (`float`, *optional*, defaults to 0.02): + The standard deviation of the truncated_normal_initializer for initializing all weight matrices. + layer_norm_eps (`float`, *optional*, defaults to 0.001): + The epsilon used by the layer normalization layers. + semantic_loss_ignore_index (`int`, *optional*, defaults to 255): + The index that is ignored by the loss function of the semantic segmentation model. + + Example: + + ```python + >>> from transformers import MobileNetV2Config, MobileNetV2Model + + >>> # Initializing a "mobilenet_v2_1.0_224" style configuration + >>> configuration = MobileNetV2Config() + + >>> # Initializing a model from the "mobilenet_v2_1.0_224" style configuration + >>> model = MobileNetV2Model(configuration) + + >>> # Accessing the model configuration + >>> configuration = model.config + ```""" + + model_type = "mobilenet_v2" + + def __init__( + self, + num_channels=3, + image_size=224, + depth_multiplier=1.0, + depth_divisible_by=8, + min_depth=8, + expand_ratio=6.0, + output_stride=32, + first_layer_is_expansion=True, + finegrained_output=True, + hidden_act="relu6", + tf_padding=True, + classifier_dropout_prob=0.8, + initializer_range=0.02, + layer_norm_eps=0.001, + semantic_loss_ignore_index=255, + **kwargs, + ): + super().__init__(**kwargs) + + if depth_multiplier <= 0: + raise ValueError("depth_multiplier must be greater than zero.") + + self.num_channels = num_channels + self.image_size = image_size + self.depth_multiplier = depth_multiplier + self.depth_divisible_by = depth_divisible_by + self.min_depth = min_depth + self.expand_ratio = expand_ratio + self.output_stride = output_stride + self.first_layer_is_expansion = first_layer_is_expansion + self.finegrained_output = finegrained_output + self.hidden_act = hidden_act + self.tf_padding = tf_padding + self.classifier_dropout_prob = classifier_dropout_prob + self.initializer_range = initializer_range + self.layer_norm_eps = layer_norm_eps + self.semantic_loss_ignore_index = semantic_loss_ignore_index + + +class MobileNetV2OnnxConfig(OnnxConfig): + torch_onnx_minimum_version = version.parse("1.11") + + @property + def inputs(self) -> Mapping[str, Mapping[int, str]]: + return OrderedDict([("pixel_values", {0: "batch"})]) + + @property + def outputs(self) -> Mapping[str, Mapping[int, str]]: + if self.task == "image-classification": + return OrderedDict([("logits", {0: "batch"})]) + else: + return OrderedDict([("last_hidden_state", {0: "batch"}), ("pooler_output", {0: "batch"})]) + + @property + def atol_for_validation(self) -> float: + return 1e-4 + + +__all__ = ["MobileNetV2Config", "MobileNetV2OnnxConfig"] diff --git a/vllm/lib/python3.10/site-packages/transformers/models/mobilenet_v2/feature_extraction_mobilenet_v2.py b/vllm/lib/python3.10/site-packages/transformers/models/mobilenet_v2/feature_extraction_mobilenet_v2.py new file mode 100644 index 0000000000000000000000000000000000000000..09043aa4834e8c97a83e06f92ab5b4f018ceecb6 --- /dev/null +++ b/vllm/lib/python3.10/site-packages/transformers/models/mobilenet_v2/feature_extraction_mobilenet_v2.py @@ -0,0 +1,36 @@ +# coding=utf-8 +# Copyright 2022 The HuggingFace Inc. team. All rights reserved. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +"""Feature extractor class for MobileNetV2.""" + +import warnings + +from ...utils import logging +from .image_processing_mobilenet_v2 import MobileNetV2ImageProcessor + + +logger = logging.get_logger(__name__) + + +class MobileNetV2FeatureExtractor(MobileNetV2ImageProcessor): + def __init__(self, *args, **kwargs) -> None: + warnings.warn( + "The class MobileNetV2FeatureExtractor is deprecated and will be removed in version 5 of Transformers." + " Please use MobileNetV2ImageProcessor instead.", + FutureWarning, + ) + super().__init__(*args, **kwargs) + + +__all__ = ["MobileNetV2FeatureExtractor"] diff --git a/vllm/lib/python3.10/site-packages/transformers/models/mobilenet_v2/image_processing_mobilenet_v2.py b/vllm/lib/python3.10/site-packages/transformers/models/mobilenet_v2/image_processing_mobilenet_v2.py new file mode 100644 index 0000000000000000000000000000000000000000..c93f95a2058e05d3a418c2ced90ff2ef5c9ce795 --- /dev/null +++ b/vllm/lib/python3.10/site-packages/transformers/models/mobilenet_v2/image_processing_mobilenet_v2.py @@ -0,0 +1,352 @@ +# coding=utf-8 +# Copyright 2022 The HuggingFace Inc. team. All rights reserved. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +"""Image processor class for MobileNetV2.""" + +from typing import Dict, List, Optional, Tuple, Union + +import numpy as np + +from ...image_processing_utils import BaseImageProcessor, BatchFeature, get_size_dict +from ...image_transforms import ( + get_resize_output_image_size, + resize, + to_channel_dimension_format, +) +from ...image_utils import ( + IMAGENET_STANDARD_MEAN, + IMAGENET_STANDARD_STD, + ChannelDimension, + ImageInput, + PILImageResampling, + infer_channel_dimension_format, + is_scaled_image, + make_list_of_images, + to_numpy_array, + valid_images, + validate_preprocess_arguments, +) +from ...utils import TensorType, filter_out_non_signature_kwargs, is_torch_available, is_torch_tensor, logging + + +if is_torch_available(): + import torch + + +logger = logging.get_logger(__name__) + + +class MobileNetV2ImageProcessor(BaseImageProcessor): + r""" + Constructs a MobileNetV2 image processor. + + Args: + do_resize (`bool`, *optional*, defaults to `True`): + Whether to resize the image's (height, width) dimensions to the specified `size`. Can be overridden by + `do_resize` in the `preprocess` method. + size (`Dict[str, int]` *optional*, defaults to `{"shortest_edge": 256}`): + Size of the image after resizing. The shortest edge of the image is resized to size["shortest_edge"], with + the longest edge resized to keep the input aspect ratio. Can be overridden by `size` in the `preprocess` + method. + resample (`PILImageResampling`, *optional*, defaults to `PILImageResampling.BILINEAR`): + Resampling filter to use if resizing the image. Can be overridden by the `resample` parameter in the + `preprocess` method. + do_center_crop (`bool`, *optional*, defaults to `True`): + Whether to center crop the image. If the input size is smaller than `crop_size` along any edge, the image + is padded with 0's and then center cropped. Can be overridden by the `do_center_crop` parameter in the + `preprocess` method. + crop_size (`Dict[str, int]`, *optional*, defaults to `{"height": 224, "width": 224}`): + Desired output size when applying center-cropping. Only has an effect if `do_center_crop` is set to `True`. + Can be overridden by the `crop_size` parameter in the `preprocess` method. + do_rescale (`bool`, *optional*, defaults to `True`): + Whether to rescale the image by the specified scale `rescale_factor`. Can be overridden by the `do_rescale` + parameter in the `preprocess` method. + rescale_factor (`int` or `float`, *optional*, defaults to `1/255`): + Scale factor to use if rescaling the image. Can be overridden by the `rescale_factor` parameter in the + `preprocess` method. + do_normalize: + Whether to normalize the image. Can be overridden by the `do_normalize` parameter in the `preprocess` + method. + image_mean (`float` or `List[float]`, *optional*, defaults to `IMAGENET_STANDARD_MEAN`): + Mean to use if normalizing the image. This is a float or list of floats the length of the number of + channels in the image. Can be overridden by the `image_mean` parameter in the `preprocess` method. + image_std (`float` or `List[float]`, *optional*, defaults to `IMAGENET_STANDARD_STD`): + Standard deviation to use if normalizing the image. This is a float or list of floats the length of the + number of channels in the image. Can be overridden by the `image_std` parameter in the `preprocess` method. + """ + + model_input_names = ["pixel_values"] + + def __init__( + self, + do_resize: bool = True, + size: Optional[Dict[str, int]] = None, + resample: PILImageResampling = PILImageResampling.BILINEAR, + do_center_crop: bool = True, + crop_size: Dict[str, int] = None, + do_rescale: bool = True, + rescale_factor: Union[int, float] = 1 / 255, + do_normalize: bool = True, + image_mean: Optional[Union[float, List[float]]] = None, + image_std: Optional[Union[float, List[float]]] = None, + **kwargs, + ) -> None: + super().__init__(**kwargs) + size = size if size is not None else {"shortest_edge": 256} + size = get_size_dict(size, default_to_square=False) + crop_size = crop_size if crop_size is not None else {"height": 224, "width": 224} + crop_size = get_size_dict(crop_size, param_name="crop_size") + self.do_resize = do_resize + self.size = size + self.resample = resample + self.do_center_crop = do_center_crop + self.crop_size = crop_size + self.do_rescale = do_rescale + self.rescale_factor = rescale_factor + self.do_normalize = do_normalize + self.image_mean = image_mean if image_mean is not None else IMAGENET_STANDARD_MEAN + self.image_std = image_std if image_std is not None else IMAGENET_STANDARD_STD + + # Copied from transformers.models.mobilenet_v1.image_processing_mobilenet_v1.MobileNetV1ImageProcessor.resize + def resize( + self, + image: np.ndarray, + size: Dict[str, int], + resample: PILImageResampling = PILImageResampling.BICUBIC, + data_format: Optional[Union[str, ChannelDimension]] = None, + input_data_format: Optional[Union[str, ChannelDimension]] = None, + **kwargs, + ) -> np.ndarray: + """ + Resize an image. The shortest edge of the image is resized to size["shortest_edge"], with the longest edge + resized to keep the input aspect ratio. + + Args: + image (`np.ndarray`): + Image to resize. + size (`Dict[str, int]`): + Size of the output image. + resample (`PILImageResampling`, *optional*, defaults to `PILImageResampling.BICUBIC`): + Resampling filter to use when resiizing the image. + data_format (`str` or `ChannelDimension`, *optional*): + The channel dimension format of the image. If not provided, it will be the same as the input image. + input_data_format (`ChannelDimension` or `str`, *optional*): + The channel dimension format of the input image. If not provided, it will be inferred. + """ + default_to_square = True + if "shortest_edge" in size: + size = size["shortest_edge"] + default_to_square = False + elif "height" in size and "width" in size: + size = (size["height"], size["width"]) + else: + raise ValueError("Size must contain either 'shortest_edge' or 'height' and 'width'.") + + output_size = get_resize_output_image_size( + image, + size=size, + default_to_square=default_to_square, + input_data_format=input_data_format, + ) + return resize( + image, + size=output_size, + resample=resample, + data_format=data_format, + input_data_format=input_data_format, + **kwargs, + ) + + @filter_out_non_signature_kwargs() + def preprocess( + self, + images: ImageInput, + do_resize: Optional[bool] = None, + size: Dict[str, int] = None, + resample: PILImageResampling = None, + do_center_crop: bool = None, + crop_size: Dict[str, int] = None, + do_rescale: Optional[bool] = None, + rescale_factor: Optional[float] = None, + do_normalize: Optional[bool] = None, + image_mean: Optional[Union[float, List[float]]] = None, + image_std: Optional[Union[float, List[float]]] = None, + return_tensors: Optional[Union[str, TensorType]] = None, + data_format: Union[str, ChannelDimension] = ChannelDimension.FIRST, + input_data_format: Optional[Union[str, ChannelDimension]] = None, + ): + """ + Preprocess an image or batch of images. + + Args: + images (`ImageInput`): + Image to preprocess. Expects a single or batch of images with pixel values ranging from 0 to 255. If + passing in images with pixel values between 0 and 1, set `do_rescale=False`. + do_resize (`bool`, *optional*, defaults to `self.do_resize`): + Whether to resize the image. + size (`Dict[str, int]`, *optional*, defaults to `self.size`): + Size of the image after resizing. Shortest edge of the image is resized to size["shortest_edge"], with + the longest edge resized to keep the input aspect ratio. + resample (`PILImageResampling` filter, *optional*, defaults to `self.resample`): + `PILImageResampling` filter to use if resizing the image e.g. `PILImageResampling.BILINEAR`. Only has + an effect if `do_resize` is set to `True`. + do_center_crop (`bool`, *optional*, defaults to `self.do_center_crop`): + Whether to center crop the image. + crop_size (`Dict[str, int]`, *optional*, defaults to `self.crop_size`): + Size of the center crop. Only has an effect if `do_center_crop` is set to `True`. + do_rescale (`bool`, *optional*, defaults to `self.do_rescale`): + Whether to rescale the image values between [0 - 1]. + rescale_factor (`float`, *optional*, defaults to `self.rescale_factor`): + Rescale factor to rescale the image by if `do_rescale` is set to `True`. + do_normalize (`bool`, *optional*, defaults to `self.do_normalize`): + Whether to normalize the image. + image_mean (`float` or `List[float]`, *optional*, defaults to `self.image_mean`): + Image mean to use if `do_normalize` is set to `True`. + image_std (`float` or `List[float]`, *optional*, defaults to `self.image_std`): + Image standard deviation to use if `do_normalize` is set to `True`. + return_tensors (`str` or `TensorType`, *optional*): + The type of tensors to return. Can be one of: + - Unset: Return a list of `np.ndarray`. + - `TensorType.TENSORFLOW` or `'tf'`: Return a batch of type `tf.Tensor`. + - `TensorType.PYTORCH` or `'pt'`: Return a batch of type `torch.Tensor`. + - `TensorType.NUMPY` or `'np'`: Return a batch of type `np.ndarray`. + - `TensorType.JAX` or `'jax'`: Return a batch of type `jax.numpy.ndarray`. + data_format (`ChannelDimension` or `str`, *optional*, defaults to `ChannelDimension.FIRST`): + The channel dimension format for the output image. Can be one of: + - `"channels_first"` or `ChannelDimension.FIRST`: image in (num_channels, height, width) format. + - `"channels_last"` or `ChannelDimension.LAST`: image in (height, width, num_channels) format. + - Unset: Use the channel dimension format of the input image. + input_data_format (`ChannelDimension` or `str`, *optional*): + The channel dimension format for the input image. If unset, the channel dimension format is inferred + from the input image. Can be one of: + - `"channels_first"` or `ChannelDimension.FIRST`: image in (num_channels, height, width) format. + - `"channels_last"` or `ChannelDimension.LAST`: image in (height, width, num_channels) format. + - `"none"` or `ChannelDimension.NONE`: image in (height, width) format. + """ + do_resize = do_resize if do_resize is not None else self.do_resize + size = size if size is not None else self.size + size = get_size_dict(size, default_to_square=False) + resample = resample if resample is not None else self.resample + do_center_crop = do_center_crop if do_center_crop is not None else self.do_center_crop + crop_size = crop_size if crop_size is not None else self.crop_size + crop_size = get_size_dict(crop_size, param_name="crop_size") + do_rescale = do_rescale if do_rescale is not None else self.do_rescale + rescale_factor = rescale_factor if rescale_factor is not None else self.rescale_factor + do_normalize = do_normalize if do_normalize is not None else self.do_normalize + image_mean = image_mean if image_mean is not None else self.image_mean + image_std = image_std if image_std is not None else self.image_std + + images = make_list_of_images(images) + + if not valid_images(images): + raise ValueError( + "Invalid image type. Must be of type PIL.Image.Image, numpy.ndarray, " + "torch.Tensor, tf.Tensor or jax.ndarray." + ) + validate_preprocess_arguments( + do_rescale=do_rescale, + rescale_factor=rescale_factor, + do_normalize=do_normalize, + image_mean=image_mean, + image_std=image_std, + do_center_crop=do_center_crop, + crop_size=crop_size, + do_resize=do_resize, + size=size, + resample=resample, + ) + # All transformations expect numpy arrays. + images = [to_numpy_array(image) for image in images] + + if do_rescale and is_scaled_image(images[0]): + logger.warning_once( + "It looks like you are trying to rescale already rescaled images. If the input" + " images have pixel values between 0 and 1, set `do_rescale=False` to avoid rescaling them again." + ) + + if input_data_format is None: + # We assume that all images have the same channel dimension format. + input_data_format = infer_channel_dimension_format(images[0]) + + all_images = [] + for image in images: + if do_resize: + image = self.resize(image=image, size=size, resample=resample, input_data_format=input_data_format) + + if do_center_crop: + image = self.center_crop(image=image, size=crop_size, input_data_format=input_data_format) + + if do_rescale: + image = self.rescale(image=image, scale=rescale_factor, input_data_format=input_data_format) + + if do_normalize: + image = self.normalize( + image=image, mean=image_mean, std=image_std, input_data_format=input_data_format + ) + + all_images.append(image) + images = [ + to_channel_dimension_format(image, data_format, input_channel_dim=input_data_format) + for image in all_images + ] + + data = {"pixel_values": images} + return BatchFeature(data=data, tensor_type=return_tensors) + + # Copied from transformers.models.beit.image_processing_beit.BeitImageProcessor.post_process_semantic_segmentation with Beit->MobileNetV2 + def post_process_semantic_segmentation(self, outputs, target_sizes: List[Tuple] = None): + """ + Converts the output of [`MobileNetV2ForSemanticSegmentation`] into semantic segmentation maps. Only supports PyTorch. + + Args: + outputs ([`MobileNetV2ForSemanticSegmentation`]): + Raw outputs of the model. + target_sizes (`List[Tuple]` of length `batch_size`, *optional*): + List of tuples corresponding to the requested final size (height, width) of each prediction. If unset, + predictions will not be resized. + + Returns: + semantic_segmentation: `List[torch.Tensor]` of length `batch_size`, where each item is a semantic + segmentation map of shape (height, width) corresponding to the target_sizes entry (if `target_sizes` is + specified). Each entry of each `torch.Tensor` correspond to a semantic class id. + """ + # TODO: add support for other frameworks + logits = outputs.logits + + # Resize logits and compute semantic segmentation maps + if target_sizes is not None: + if len(logits) != len(target_sizes): + raise ValueError( + "Make sure that you pass in as many target sizes as the batch dimension of the logits" + ) + + if is_torch_tensor(target_sizes): + target_sizes = target_sizes.numpy() + + semantic_segmentation = [] + + for idx in range(len(logits)): + resized_logits = torch.nn.functional.interpolate( + logits[idx].unsqueeze(dim=0), size=target_sizes[idx], mode="bilinear", align_corners=False + ) + semantic_map = resized_logits[0].argmax(dim=0) + semantic_segmentation.append(semantic_map) + else: + semantic_segmentation = logits.argmax(dim=1) + semantic_segmentation = [semantic_segmentation[i] for i in range(semantic_segmentation.shape[0])] + + return semantic_segmentation + + +__all__ = ["MobileNetV2ImageProcessor"] diff --git a/vllm/lib/python3.10/site-packages/transformers/models/mobilenet_v2/modeling_mobilenet_v2.py b/vllm/lib/python3.10/site-packages/transformers/models/mobilenet_v2/modeling_mobilenet_v2.py new file mode 100644 index 0000000000000000000000000000000000000000..bd36d111f8af174a09f3211a969bb396c5e274e6 --- /dev/null +++ b/vllm/lib/python3.10/site-packages/transformers/models/mobilenet_v2/modeling_mobilenet_v2.py @@ -0,0 +1,868 @@ +# coding=utf-8 +# Copyright 2022 Apple Inc. and The HuggingFace Inc. team. All rights reserved. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +"""PyTorch MobileNetV2 model.""" + +from typing import Optional, Union + +import torch +from torch import nn +from torch.nn import BCEWithLogitsLoss, CrossEntropyLoss, MSELoss + +from ...activations import ACT2FN +from ...modeling_outputs import ( + BaseModelOutputWithPoolingAndNoAttention, + ImageClassifierOutputWithNoAttention, + SemanticSegmenterOutput, +) +from ...modeling_utils import PreTrainedModel +from ...utils import ( + add_code_sample_docstrings, + add_start_docstrings, + add_start_docstrings_to_model_forward, + logging, + replace_return_docstrings, +) +from .configuration_mobilenet_v2 import MobileNetV2Config + + +logger = logging.get_logger(__name__) + + +# General docstring +_CONFIG_FOR_DOC = "MobileNetV2Config" + +# Base docstring +_CHECKPOINT_FOR_DOC = "google/mobilenet_v2_1.0_224" +_EXPECTED_OUTPUT_SHAPE = [1, 1280, 7, 7] + +# Image classification docstring +_IMAGE_CLASS_CHECKPOINT = "google/mobilenet_v2_1.0_224" +_IMAGE_CLASS_EXPECTED_OUTPUT = "tabby, tabby cat" + + +def _build_tf_to_pytorch_map(model, config, tf_weights=None): + """ + A map of modules from TF to PyTorch. + """ + + tf_to_pt_map = {} + + if isinstance(model, (MobileNetV2ForImageClassification, MobileNetV2ForSemanticSegmentation)): + backbone = model.mobilenet_v2 + else: + backbone = model + + # Use the EMA weights if available + def ema(x): + return x + "/ExponentialMovingAverage" if x + "/ExponentialMovingAverage" in tf_weights else x + + prefix = "MobilenetV2/Conv/" + tf_to_pt_map[ema(prefix + "weights")] = backbone.conv_stem.first_conv.convolution.weight + tf_to_pt_map[ema(prefix + "BatchNorm/beta")] = backbone.conv_stem.first_conv.normalization.bias + tf_to_pt_map[ema(prefix + "BatchNorm/gamma")] = backbone.conv_stem.first_conv.normalization.weight + tf_to_pt_map[prefix + "BatchNorm/moving_mean"] = backbone.conv_stem.first_conv.normalization.running_mean + tf_to_pt_map[prefix + "BatchNorm/moving_variance"] = backbone.conv_stem.first_conv.normalization.running_var + + prefix = "MobilenetV2/expanded_conv/depthwise/" + tf_to_pt_map[ema(prefix + "depthwise_weights")] = backbone.conv_stem.conv_3x3.convolution.weight + tf_to_pt_map[ema(prefix + "BatchNorm/beta")] = backbone.conv_stem.conv_3x3.normalization.bias + tf_to_pt_map[ema(prefix + "BatchNorm/gamma")] = backbone.conv_stem.conv_3x3.normalization.weight + tf_to_pt_map[prefix + "BatchNorm/moving_mean"] = backbone.conv_stem.conv_3x3.normalization.running_mean + tf_to_pt_map[prefix + "BatchNorm/moving_variance"] = backbone.conv_stem.conv_3x3.normalization.running_var + + prefix = "MobilenetV2/expanded_conv/project/" + tf_to_pt_map[ema(prefix + "weights")] = backbone.conv_stem.reduce_1x1.convolution.weight + tf_to_pt_map[ema(prefix + "BatchNorm/beta")] = backbone.conv_stem.reduce_1x1.normalization.bias + tf_to_pt_map[ema(prefix + "BatchNorm/gamma")] = backbone.conv_stem.reduce_1x1.normalization.weight + tf_to_pt_map[prefix + "BatchNorm/moving_mean"] = backbone.conv_stem.reduce_1x1.normalization.running_mean + tf_to_pt_map[prefix + "BatchNorm/moving_variance"] = backbone.conv_stem.reduce_1x1.normalization.running_var + + for i in range(16): + tf_index = i + 1 + pt_index = i + pointer = backbone.layer[pt_index] + + prefix = f"MobilenetV2/expanded_conv_{tf_index}/expand/" + tf_to_pt_map[ema(prefix + "weights")] = pointer.expand_1x1.convolution.weight + tf_to_pt_map[ema(prefix + "BatchNorm/beta")] = pointer.expand_1x1.normalization.bias + tf_to_pt_map[ema(prefix + "BatchNorm/gamma")] = pointer.expand_1x1.normalization.weight + tf_to_pt_map[prefix + "BatchNorm/moving_mean"] = pointer.expand_1x1.normalization.running_mean + tf_to_pt_map[prefix + "BatchNorm/moving_variance"] = pointer.expand_1x1.normalization.running_var + + prefix = f"MobilenetV2/expanded_conv_{tf_index}/depthwise/" + tf_to_pt_map[ema(prefix + "depthwise_weights")] = pointer.conv_3x3.convolution.weight + tf_to_pt_map[ema(prefix + "BatchNorm/beta")] = pointer.conv_3x3.normalization.bias + tf_to_pt_map[ema(prefix + "BatchNorm/gamma")] = pointer.conv_3x3.normalization.weight + tf_to_pt_map[prefix + "BatchNorm/moving_mean"] = pointer.conv_3x3.normalization.running_mean + tf_to_pt_map[prefix + "BatchNorm/moving_variance"] = pointer.conv_3x3.normalization.running_var + + prefix = f"MobilenetV2/expanded_conv_{tf_index}/project/" + tf_to_pt_map[ema(prefix + "weights")] = pointer.reduce_1x1.convolution.weight + tf_to_pt_map[ema(prefix + "BatchNorm/beta")] = pointer.reduce_1x1.normalization.bias + tf_to_pt_map[ema(prefix + "BatchNorm/gamma")] = pointer.reduce_1x1.normalization.weight + tf_to_pt_map[prefix + "BatchNorm/moving_mean"] = pointer.reduce_1x1.normalization.running_mean + tf_to_pt_map[prefix + "BatchNorm/moving_variance"] = pointer.reduce_1x1.normalization.running_var + + prefix = "MobilenetV2/Conv_1/" + tf_to_pt_map[ema(prefix + "weights")] = backbone.conv_1x1.convolution.weight + tf_to_pt_map[ema(prefix + "BatchNorm/beta")] = backbone.conv_1x1.normalization.bias + tf_to_pt_map[ema(prefix + "BatchNorm/gamma")] = backbone.conv_1x1.normalization.weight + tf_to_pt_map[prefix + "BatchNorm/moving_mean"] = backbone.conv_1x1.normalization.running_mean + tf_to_pt_map[prefix + "BatchNorm/moving_variance"] = backbone.conv_1x1.normalization.running_var + + if isinstance(model, MobileNetV2ForImageClassification): + prefix = "MobilenetV2/Logits/Conv2d_1c_1x1/" + tf_to_pt_map[ema(prefix + "weights")] = model.classifier.weight + tf_to_pt_map[ema(prefix + "biases")] = model.classifier.bias + + if isinstance(model, MobileNetV2ForSemanticSegmentation): + prefix = "image_pooling/" + tf_to_pt_map[prefix + "weights"] = model.segmentation_head.conv_pool.convolution.weight + tf_to_pt_map[prefix + "BatchNorm/beta"] = model.segmentation_head.conv_pool.normalization.bias + tf_to_pt_map[prefix + "BatchNorm/gamma"] = model.segmentation_head.conv_pool.normalization.weight + tf_to_pt_map[prefix + "BatchNorm/moving_mean"] = model.segmentation_head.conv_pool.normalization.running_mean + tf_to_pt_map[prefix + "BatchNorm/moving_variance"] = ( + model.segmentation_head.conv_pool.normalization.running_var + ) + + prefix = "aspp0/" + tf_to_pt_map[prefix + "weights"] = model.segmentation_head.conv_aspp.convolution.weight + tf_to_pt_map[prefix + "BatchNorm/beta"] = model.segmentation_head.conv_aspp.normalization.bias + tf_to_pt_map[prefix + "BatchNorm/gamma"] = model.segmentation_head.conv_aspp.normalization.weight + tf_to_pt_map[prefix + "BatchNorm/moving_mean"] = model.segmentation_head.conv_aspp.normalization.running_mean + tf_to_pt_map[prefix + "BatchNorm/moving_variance"] = ( + model.segmentation_head.conv_aspp.normalization.running_var + ) + + prefix = "concat_projection/" + tf_to_pt_map[prefix + "weights"] = model.segmentation_head.conv_projection.convolution.weight + tf_to_pt_map[prefix + "BatchNorm/beta"] = model.segmentation_head.conv_projection.normalization.bias + tf_to_pt_map[prefix + "BatchNorm/gamma"] = model.segmentation_head.conv_projection.normalization.weight + tf_to_pt_map[prefix + "BatchNorm/moving_mean"] = ( + model.segmentation_head.conv_projection.normalization.running_mean + ) + tf_to_pt_map[prefix + "BatchNorm/moving_variance"] = ( + model.segmentation_head.conv_projection.normalization.running_var + ) + + prefix = "logits/semantic/" + tf_to_pt_map[ema(prefix + "weights")] = model.segmentation_head.classifier.convolution.weight + tf_to_pt_map[ema(prefix + "biases")] = model.segmentation_head.classifier.convolution.bias + + return tf_to_pt_map + + +def load_tf_weights_in_mobilenet_v2(model, config, tf_checkpoint_path): + """Load TensorFlow checkpoints in a PyTorch model.""" + try: + import numpy as np + import tensorflow as tf + except ImportError: + logger.error( + "Loading a TensorFlow models in PyTorch, requires TensorFlow to be installed. Please see " + "https://www.tensorflow.org/install/ for installation instructions." + ) + raise + + # Load weights from TF model + init_vars = tf.train.list_variables(tf_checkpoint_path) + tf_weights = {} + for name, shape in init_vars: + logger.info(f"Loading TF weight {name} with shape {shape}") + array = tf.train.load_variable(tf_checkpoint_path, name) + tf_weights[name] = array + + # Build TF to PyTorch weights loading map + tf_to_pt_map = _build_tf_to_pytorch_map(model, config, tf_weights) + + for name, pointer in tf_to_pt_map.items(): + logger.info(f"Importing {name}") + if name not in tf_weights: + logger.info(f"{name} not in tf pre-trained weights, skipping") + continue + + array = tf_weights[name] + + if "depthwise_weights" in name: + logger.info("Transposing depthwise") + array = np.transpose(array, (2, 3, 0, 1)) + elif "weights" in name: + logger.info("Transposing") + if len(pointer.shape) == 2: # copying into linear layer + array = array.squeeze().transpose() + else: + array = np.transpose(array, (3, 2, 0, 1)) + + if pointer.shape != array.shape: + raise ValueError(f"Pointer shape {pointer.shape} and array shape {array.shape} mismatched") + + logger.info(f"Initialize PyTorch weight {name} {array.shape}") + pointer.data = torch.from_numpy(array) + + tf_weights.pop(name, None) + tf_weights.pop(name + "/RMSProp", None) + tf_weights.pop(name + "/RMSProp_1", None) + tf_weights.pop(name + "/ExponentialMovingAverage", None) + tf_weights.pop(name + "/Momentum", None) + + logger.info(f"Weights not copied to PyTorch model: {', '.join(tf_weights.keys())}") + return model + + +def make_divisible(value: int, divisor: int = 8, min_value: Optional[int] = None) -> int: + """ + Ensure that all layers have a channel count that is divisible by `divisor`. This function is taken from the + original TensorFlow repo. It can be seen here: + https://github.com/tensorflow/models/blob/master/research/slim/nets/mobilenet/mobilenet.py + """ + if min_value is None: + min_value = divisor + new_value = max(min_value, int(value + divisor / 2) // divisor * divisor) + # Make sure that round down does not go down by more than 10%. + if new_value < 0.9 * value: + new_value += divisor + return int(new_value) + + +def apply_depth_multiplier(config: MobileNetV2Config, channels: int) -> int: + return make_divisible(int(round(channels * config.depth_multiplier)), config.depth_divisible_by, config.min_depth) + + +def apply_tf_padding(features: torch.Tensor, conv_layer: nn.Conv2d) -> torch.Tensor: + """ + Apply TensorFlow-style "SAME" padding to a convolution layer. See the notes at: + https://www.tensorflow.org/api_docs/python/tf/nn#notes_on_padding_2 + """ + in_height = int(features.shape[-2]) + in_width = int(features.shape[-1]) + stride_height, stride_width = conv_layer.stride + kernel_height, kernel_width = conv_layer.kernel_size + dilation_height, dilation_width = conv_layer.dilation + + if in_height % stride_height == 0: + pad_along_height = max(kernel_height - stride_height, 0) + else: + pad_along_height = max(kernel_height - (in_height % stride_height), 0) + + if in_width % stride_width == 0: + pad_along_width = max(kernel_width - stride_width, 0) + else: + pad_along_width = max(kernel_width - (in_width % stride_width), 0) + + pad_left = pad_along_width // 2 + pad_right = pad_along_width - pad_left + pad_top = pad_along_height // 2 + pad_bottom = pad_along_height - pad_top + + padding = ( + pad_left * dilation_width, + pad_right * dilation_width, + pad_top * dilation_height, + pad_bottom * dilation_height, + ) + return nn.functional.pad(features, padding, "constant", 0.0) + + +class MobileNetV2ConvLayer(nn.Module): + def __init__( + self, + config: MobileNetV2Config, + in_channels: int, + out_channels: int, + kernel_size: int, + stride: int = 1, + groups: int = 1, + bias: bool = False, + dilation: int = 1, + use_normalization: bool = True, + use_activation: Union[bool, str] = True, + layer_norm_eps: Optional[float] = None, + ) -> None: + super().__init__() + self.config = config + + if in_channels % groups != 0: + raise ValueError(f"Input channels ({in_channels}) are not divisible by {groups} groups.") + if out_channels % groups != 0: + raise ValueError(f"Output channels ({out_channels}) are not divisible by {groups} groups.") + + padding = 0 if config.tf_padding else int((kernel_size - 1) / 2) * dilation + + self.convolution = nn.Conv2d( + in_channels=in_channels, + out_channels=out_channels, + kernel_size=kernel_size, + stride=stride, + padding=padding, + dilation=dilation, + groups=groups, + bias=bias, + padding_mode="zeros", + ) + + if use_normalization: + self.normalization = nn.BatchNorm2d( + num_features=out_channels, + eps=config.layer_norm_eps if layer_norm_eps is None else layer_norm_eps, + momentum=0.997, + affine=True, + track_running_stats=True, + ) + else: + self.normalization = None + + if use_activation: + if isinstance(use_activation, str): + self.activation = ACT2FN[use_activation] + elif isinstance(config.hidden_act, str): + self.activation = ACT2FN[config.hidden_act] + else: + self.activation = config.hidden_act + else: + self.activation = None + + def forward(self, features: torch.Tensor) -> torch.Tensor: + if self.config.tf_padding: + features = apply_tf_padding(features, self.convolution) + features = self.convolution(features) + if self.normalization is not None: + features = self.normalization(features) + if self.activation is not None: + features = self.activation(features) + return features + + +class MobileNetV2InvertedResidual(nn.Module): + def __init__( + self, config: MobileNetV2Config, in_channels: int, out_channels: int, stride: int, dilation: int = 1 + ) -> None: + super().__init__() + + expanded_channels = make_divisible( + int(round(in_channels * config.expand_ratio)), config.depth_divisible_by, config.min_depth + ) + + if stride not in [1, 2]: + raise ValueError(f"Invalid stride {stride}.") + + self.use_residual = (stride == 1) and (in_channels == out_channels) + + self.expand_1x1 = MobileNetV2ConvLayer( + config, in_channels=in_channels, out_channels=expanded_channels, kernel_size=1 + ) + + self.conv_3x3 = MobileNetV2ConvLayer( + config, + in_channels=expanded_channels, + out_channels=expanded_channels, + kernel_size=3, + stride=stride, + groups=expanded_channels, + dilation=dilation, + ) + + self.reduce_1x1 = MobileNetV2ConvLayer( + config, + in_channels=expanded_channels, + out_channels=out_channels, + kernel_size=1, + use_activation=False, + ) + + def forward(self, features: torch.Tensor) -> torch.Tensor: + residual = features + + features = self.expand_1x1(features) + features = self.conv_3x3(features) + features = self.reduce_1x1(features) + + return residual + features if self.use_residual else features + + +class MobileNetV2Stem(nn.Module): + def __init__(self, config: MobileNetV2Config, in_channels: int, expanded_channels: int, out_channels: int) -> None: + super().__init__() + + # The very first layer is a regular 3x3 convolution with stride 2 that expands to 32 channels. + # All other expansion layers use the expansion factor to compute the number of output channels. + self.first_conv = MobileNetV2ConvLayer( + config, + in_channels=in_channels, + out_channels=expanded_channels, + kernel_size=3, + stride=2, + ) + + if config.first_layer_is_expansion: + self.expand_1x1 = None + else: + self.expand_1x1 = MobileNetV2ConvLayer( + config, in_channels=expanded_channels, out_channels=expanded_channels, kernel_size=1 + ) + + self.conv_3x3 = MobileNetV2ConvLayer( + config, + in_channels=expanded_channels, + out_channels=expanded_channels, + kernel_size=3, + stride=1, + groups=expanded_channels, + ) + + self.reduce_1x1 = MobileNetV2ConvLayer( + config, + in_channels=expanded_channels, + out_channels=out_channels, + kernel_size=1, + use_activation=False, + ) + + def forward(self, features: torch.Tensor) -> torch.Tensor: + features = self.first_conv(features) + if self.expand_1x1 is not None: + features = self.expand_1x1(features) + features = self.conv_3x3(features) + features = self.reduce_1x1(features) + return features + + +class MobileNetV2PreTrainedModel(PreTrainedModel): + """ + An abstract class to handle weights initialization and a simple interface for downloading and loading pretrained + models. + """ + + config_class = MobileNetV2Config + load_tf_weights = load_tf_weights_in_mobilenet_v2 + base_model_prefix = "mobilenet_v2" + main_input_name = "pixel_values" + supports_gradient_checkpointing = False + _no_split_modules = [] + + def _init_weights(self, module: Union[nn.Linear, nn.Conv2d]) -> None: + """Initialize the weights""" + if isinstance(module, (nn.Linear, nn.Conv2d)): + module.weight.data.normal_(mean=0.0, std=self.config.initializer_range) + if module.bias is not None: + module.bias.data.zero_() + elif isinstance(module, nn.BatchNorm2d): + module.bias.data.zero_() + module.weight.data.fill_(1.0) + + +MOBILENET_V2_START_DOCSTRING = r""" + This model is a PyTorch [torch.nn.Module](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) subclass. Use it + as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and + behavior. + + Parameters: + config ([`MobileNetV2Config`]): Model configuration class with all the parameters of the model. + Initializing with a config file does not load the weights associated with the model, only the + configuration. Check out the [`~PreTrainedModel.from_pretrained`] method to load the model weights. +""" + +MOBILENET_V2_INPUTS_DOCSTRING = r""" + Args: + pixel_values (`torch.FloatTensor` of shape `(batch_size, num_channels, height, width)`): + Pixel values. Pixel values can be obtained using [`AutoImageProcessor`]. See + [`MobileNetV2ImageProcessor.__call__`] for details. + output_hidden_states (`bool`, *optional*): + Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for + more detail. + return_dict (`bool`, *optional*): + Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple. +""" + + +@add_start_docstrings( + "The bare MobileNetV2 model outputting raw hidden-states without any specific head on top.", + MOBILENET_V2_START_DOCSTRING, +) +class MobileNetV2Model(MobileNetV2PreTrainedModel): + def __init__(self, config: MobileNetV2Config, add_pooling_layer: bool = True): + super().__init__(config) + self.config = config + + # Output channels for the projection layers + channels = [16, 24, 24, 32, 32, 32, 64, 64, 64, 64, 96, 96, 96, 160, 160, 160, 320] + channels = [apply_depth_multiplier(config, x) for x in channels] + + # Strides for the depthwise layers + strides = [2, 1, 2, 1, 1, 2, 1, 1, 1, 1, 1, 1, 2, 1, 1, 1] + + self.conv_stem = MobileNetV2Stem( + config, + in_channels=config.num_channels, + expanded_channels=apply_depth_multiplier(config, 32), + out_channels=channels[0], + ) + + current_stride = 2 # first conv layer has stride 2 + dilation = 1 + + self.layer = nn.ModuleList() + for i in range(16): + # Keep making the feature maps smaller or use dilated convolution? + if current_stride == config.output_stride: + layer_stride = 1 + layer_dilation = dilation + dilation *= strides[i] # larger dilation starts in next block + else: + layer_stride = strides[i] + layer_dilation = 1 + current_stride *= layer_stride + + self.layer.append( + MobileNetV2InvertedResidual( + config, + in_channels=channels[i], + out_channels=channels[i + 1], + stride=layer_stride, + dilation=layer_dilation, + ) + ) + + if config.finegrained_output and config.depth_multiplier < 1.0: + output_channels = 1280 + else: + output_channels = apply_depth_multiplier(config, 1280) + + self.conv_1x1 = MobileNetV2ConvLayer( + config, + in_channels=channels[-1], + out_channels=output_channels, + kernel_size=1, + ) + + self.pooler = nn.AdaptiveAvgPool2d((1, 1)) if add_pooling_layer else None + + # Initialize weights and apply final processing + self.post_init() + + def _prune_heads(self, heads_to_prune): + raise NotImplementedError + + @add_start_docstrings_to_model_forward(MOBILENET_V2_INPUTS_DOCSTRING) + @add_code_sample_docstrings( + checkpoint=_CHECKPOINT_FOR_DOC, + output_type=BaseModelOutputWithPoolingAndNoAttention, + config_class=_CONFIG_FOR_DOC, + modality="vision", + expected_output=_EXPECTED_OUTPUT_SHAPE, + ) + def forward( + self, + pixel_values: Optional[torch.Tensor] = None, + output_hidden_states: Optional[bool] = None, + return_dict: Optional[bool] = None, + ) -> Union[tuple, BaseModelOutputWithPoolingAndNoAttention]: + output_hidden_states = ( + output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states + ) + return_dict = return_dict if return_dict is not None else self.config.use_return_dict + + if pixel_values is None: + raise ValueError("You have to specify pixel_values") + + hidden_states = self.conv_stem(pixel_values) + + all_hidden_states = () if output_hidden_states else None + + for i, layer_module in enumerate(self.layer): + hidden_states = layer_module(hidden_states) + + if output_hidden_states: + all_hidden_states = all_hidden_states + (hidden_states,) + + last_hidden_state = self.conv_1x1(hidden_states) + + if self.pooler is not None: + pooled_output = torch.flatten(self.pooler(last_hidden_state), start_dim=1) + else: + pooled_output = None + + if not return_dict: + return tuple(v for v in [last_hidden_state, pooled_output, all_hidden_states] if v is not None) + + return BaseModelOutputWithPoolingAndNoAttention( + last_hidden_state=last_hidden_state, + pooler_output=pooled_output, + hidden_states=all_hidden_states, + ) + + +@add_start_docstrings( + """ + MobileNetV2 model with an image classification head on top (a linear layer on top of the pooled features), e.g. for + ImageNet. + """, + MOBILENET_V2_START_DOCSTRING, +) +class MobileNetV2ForImageClassification(MobileNetV2PreTrainedModel): + def __init__(self, config: MobileNetV2Config) -> None: + super().__init__(config) + + self.num_labels = config.num_labels + self.mobilenet_v2 = MobileNetV2Model(config) + + last_hidden_size = self.mobilenet_v2.conv_1x1.convolution.out_channels + + # Classifier head + self.dropout = nn.Dropout(config.classifier_dropout_prob, inplace=True) + self.classifier = nn.Linear(last_hidden_size, config.num_labels) if config.num_labels > 0 else nn.Identity() + + # Initialize weights and apply final processing + self.post_init() + + @add_start_docstrings_to_model_forward(MOBILENET_V2_INPUTS_DOCSTRING) + @add_code_sample_docstrings( + checkpoint=_CHECKPOINT_FOR_DOC, + output_type=ImageClassifierOutputWithNoAttention, + config_class=_CONFIG_FOR_DOC, + expected_output=_IMAGE_CLASS_EXPECTED_OUTPUT, + ) + def forward( + self, + pixel_values: Optional[torch.Tensor] = None, + output_hidden_states: Optional[bool] = None, + labels: Optional[torch.Tensor] = None, + return_dict: Optional[bool] = None, + ) -> Union[tuple, ImageClassifierOutputWithNoAttention]: + r""" + labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*): + Labels for computing the image classification/regression loss. Indices should be in `[0, ..., + config.num_labels - 1]`. If `config.num_labels == 1` a regression loss is computed (Mean-Square loss). If + `config.num_labels > 1` a classification loss is computed (Cross-Entropy). + """ + return_dict = return_dict if return_dict is not None else self.config.use_return_dict + + outputs = self.mobilenet_v2(pixel_values, output_hidden_states=output_hidden_states, return_dict=return_dict) + + pooled_output = outputs.pooler_output if return_dict else outputs[1] + + logits = self.classifier(self.dropout(pooled_output)) + + loss = None + if labels is not None: + if self.config.problem_type is None: + if self.num_labels == 1: + self.config.problem_type = "regression" + elif self.num_labels > 1 and (labels.dtype == torch.long or labels.dtype == torch.int): + self.config.problem_type = "single_label_classification" + else: + self.config.problem_type = "multi_label_classification" + + if self.config.problem_type == "regression": + loss_fct = MSELoss() + if self.num_labels == 1: + loss = loss_fct(logits.squeeze(), labels.squeeze()) + else: + loss = loss_fct(logits, labels) + elif self.config.problem_type == "single_label_classification": + loss_fct = CrossEntropyLoss() + loss = loss_fct(logits.view(-1, self.num_labels), labels.view(-1)) + elif self.config.problem_type == "multi_label_classification": + loss_fct = BCEWithLogitsLoss() + loss = loss_fct(logits, labels) + + if not return_dict: + output = (logits,) + outputs[2:] + return ((loss,) + output) if loss is not None else output + + return ImageClassifierOutputWithNoAttention( + loss=loss, + logits=logits, + hidden_states=outputs.hidden_states, + ) + + +class MobileNetV2DeepLabV3Plus(nn.Module): + """ + The neural network from the paper "Encoder-Decoder with Atrous Separable Convolution for Semantic Image + Segmentation" https://arxiv.org/abs/1802.02611 + """ + + def __init__(self, config: MobileNetV2Config) -> None: + super().__init__() + + self.avg_pool = nn.AdaptiveAvgPool2d(output_size=1) + + self.conv_pool = MobileNetV2ConvLayer( + config, + in_channels=apply_depth_multiplier(config, 320), + out_channels=256, + kernel_size=1, + stride=1, + use_normalization=True, + use_activation="relu", + layer_norm_eps=1e-5, + ) + + self.conv_aspp = MobileNetV2ConvLayer( + config, + in_channels=apply_depth_multiplier(config, 320), + out_channels=256, + kernel_size=1, + stride=1, + use_normalization=True, + use_activation="relu", + layer_norm_eps=1e-5, + ) + + self.conv_projection = MobileNetV2ConvLayer( + config, + in_channels=512, + out_channels=256, + kernel_size=1, + stride=1, + use_normalization=True, + use_activation="relu", + layer_norm_eps=1e-5, + ) + + self.dropout = nn.Dropout2d(config.classifier_dropout_prob) + + self.classifier = MobileNetV2ConvLayer( + config, + in_channels=256, + out_channels=config.num_labels, + kernel_size=1, + use_normalization=False, + use_activation=False, + bias=True, + ) + + def forward(self, features: torch.Tensor) -> torch.Tensor: + spatial_size = features.shape[-2:] + + features_pool = self.avg_pool(features) + features_pool = self.conv_pool(features_pool) + features_pool = nn.functional.interpolate( + features_pool, size=spatial_size, mode="bilinear", align_corners=True + ) + + features_aspp = self.conv_aspp(features) + + features = torch.cat([features_pool, features_aspp], dim=1) + + features = self.conv_projection(features) + features = self.dropout(features) + features = self.classifier(features) + return features + + +@add_start_docstrings( + """ + MobileNetV2 model with a semantic segmentation head on top, e.g. for Pascal VOC. + """, + MOBILENET_V2_START_DOCSTRING, +) +class MobileNetV2ForSemanticSegmentation(MobileNetV2PreTrainedModel): + def __init__(self, config: MobileNetV2Config) -> None: + super().__init__(config) + + self.num_labels = config.num_labels + self.mobilenet_v2 = MobileNetV2Model(config, add_pooling_layer=False) + self.segmentation_head = MobileNetV2DeepLabV3Plus(config) + + # Initialize weights and apply final processing + self.post_init() + + @add_start_docstrings_to_model_forward(MOBILENET_V2_INPUTS_DOCSTRING) + @replace_return_docstrings(output_type=SemanticSegmenterOutput, config_class=_CONFIG_FOR_DOC) + def forward( + self, + pixel_values: Optional[torch.Tensor] = None, + labels: Optional[torch.Tensor] = None, + output_hidden_states: Optional[bool] = None, + return_dict: Optional[bool] = None, + ) -> Union[tuple, SemanticSegmenterOutput]: + r""" + labels (`torch.LongTensor` of shape `(batch_size, height, width)`, *optional*): + Ground truth semantic segmentation maps for computing the loss. Indices should be in `[0, ..., + config.num_labels - 1]`. If `config.num_labels > 1`, a classification loss is computed (Cross-Entropy). + + Returns: + + Examples: + + ```python + >>> from transformers import AutoImageProcessor, MobileNetV2ForSemanticSegmentation + >>> from PIL import Image + >>> import requests + + >>> url = "http://images.cocodataset.org/val2017/000000039769.jpg" + >>> image = Image.open(requests.get(url, stream=True).raw) + + >>> image_processor = AutoImageProcessor.from_pretrained("google/deeplabv3_mobilenet_v2_1.0_513") + >>> model = MobileNetV2ForSemanticSegmentation.from_pretrained("google/deeplabv3_mobilenet_v2_1.0_513") + + >>> inputs = image_processor(images=image, return_tensors="pt") + + >>> with torch.no_grad(): + ... outputs = model(**inputs) + + >>> # logits are of shape (batch_size, num_labels, height, width) + >>> logits = outputs.logits + ```""" + output_hidden_states = ( + output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states + ) + return_dict = return_dict if return_dict is not None else self.config.use_return_dict + + if labels is not None and self.config.num_labels == 1: + raise ValueError("The number of labels should be greater than one") + + outputs = self.mobilenet_v2( + pixel_values, + output_hidden_states=True, # we need the intermediate hidden states + return_dict=return_dict, + ) + + encoder_hidden_states = outputs.hidden_states if return_dict else outputs[1] + + logits = self.segmentation_head(encoder_hidden_states[-1]) + + loss = None + if labels is not None: + # upsample logits to the images' original size + upsampled_logits = nn.functional.interpolate( + logits, size=labels.shape[-2:], mode="bilinear", align_corners=False + ) + loss_fct = CrossEntropyLoss(ignore_index=self.config.semantic_loss_ignore_index) + loss = loss_fct(upsampled_logits, labels) + + if not return_dict: + if output_hidden_states: + output = (logits,) + outputs[1:] + else: + output = (logits,) + outputs[2:] + return ((loss,) + output) if loss is not None else output + + return SemanticSegmenterOutput( + loss=loss, + logits=logits, + hidden_states=outputs.hidden_states if output_hidden_states else None, + attentions=None, + ) + + +__all__ = [ + "MobileNetV2ForImageClassification", + "MobileNetV2ForSemanticSegmentation", + "MobileNetV2Model", + "MobileNetV2PreTrainedModel", + "load_tf_weights_in_mobilenet_v2", +] diff --git a/vllm/lib/python3.10/site-packages/transformers/models/modernbert/__init__.py b/vllm/lib/python3.10/site-packages/transformers/models/modernbert/__init__.py new file mode 100644 index 0000000000000000000000000000000000000000..18317742981909460973138da806584ddfc4a390 --- /dev/null +++ b/vllm/lib/python3.10/site-packages/transformers/models/modernbert/__init__.py @@ -0,0 +1,27 @@ +# Copyright 2024 The HuggingFace Inc. team. All rights reserved. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +from typing import TYPE_CHECKING + +from ...utils import _LazyModule +from ...utils.import_utils import define_import_structure + + +if TYPE_CHECKING: + from .configuration_modernbert import * + from .modeling_modernbert import * +else: + import sys + + _file = globals()["__file__"] + sys.modules[__name__] = _LazyModule(__name__, _file, define_import_structure(_file), module_spec=__spec__) diff --git a/vllm/lib/python3.10/site-packages/transformers/models/modernbert/__pycache__/__init__.cpython-310.pyc b/vllm/lib/python3.10/site-packages/transformers/models/modernbert/__pycache__/__init__.cpython-310.pyc new file mode 100644 index 0000000000000000000000000000000000000000..58d1046474064dc39466e7f6e7f018b1b65bc5cc Binary files /dev/null and b/vllm/lib/python3.10/site-packages/transformers/models/modernbert/__pycache__/__init__.cpython-310.pyc differ diff --git a/vllm/lib/python3.10/site-packages/transformers/models/modernbert/__pycache__/configuration_modernbert.cpython-310.pyc b/vllm/lib/python3.10/site-packages/transformers/models/modernbert/__pycache__/configuration_modernbert.cpython-310.pyc new file mode 100644 index 0000000000000000000000000000000000000000..cf077a92042cfce8c6a8e7b29ba9489baeaaeca3 Binary files /dev/null and b/vllm/lib/python3.10/site-packages/transformers/models/modernbert/__pycache__/configuration_modernbert.cpython-310.pyc differ diff --git a/vllm/lib/python3.10/site-packages/transformers/models/modernbert/__pycache__/modeling_modernbert.cpython-310.pyc b/vllm/lib/python3.10/site-packages/transformers/models/modernbert/__pycache__/modeling_modernbert.cpython-310.pyc new file mode 100644 index 0000000000000000000000000000000000000000..053ca8f814316bceb3b6c95f7c31d37068e7aa5f Binary files /dev/null and b/vllm/lib/python3.10/site-packages/transformers/models/modernbert/__pycache__/modeling_modernbert.cpython-310.pyc differ diff --git a/vllm/lib/python3.10/site-packages/transformers/models/modernbert/__pycache__/modular_modernbert.cpython-310.pyc b/vllm/lib/python3.10/site-packages/transformers/models/modernbert/__pycache__/modular_modernbert.cpython-310.pyc new file mode 100644 index 0000000000000000000000000000000000000000..f04c599a22dc81193a563c71e5b862d7c4eb2687 Binary files /dev/null and b/vllm/lib/python3.10/site-packages/transformers/models/modernbert/__pycache__/modular_modernbert.cpython-310.pyc differ diff --git a/vllm/lib/python3.10/site-packages/transformers/models/modernbert/configuration_modernbert.py b/vllm/lib/python3.10/site-packages/transformers/models/modernbert/configuration_modernbert.py new file mode 100644 index 0000000000000000000000000000000000000000..cc0295c25b55d7e5ac431b83c0d5c0b4fdf9762b --- /dev/null +++ b/vllm/lib/python3.10/site-packages/transformers/models/modernbert/configuration_modernbert.py @@ -0,0 +1,218 @@ +# 🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨 +# This file was automatically generated from src/transformers/models/modernbert/modular_modernbert.py. +# Do NOT edit this file manually as any edits will be overwritten by the generation of +# the file from the modular. If any change should be done, please apply the change to the +# modular_modernbert.py file directly. One of our CI enforces this. +# 🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨 +# Copyright 2024 Answer.AI, LightOn, and contributors, and the HuggingFace Inc. team. All rights reserved. +# +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +from typing import Literal + +from ...configuration_utils import PretrainedConfig + + +class ModernBertConfig(PretrainedConfig): + r""" + This is the configuration class to store the configuration of a [`ModernBertModel`]. It is used to instantiate an ModernBert + model according to the specified arguments, defining the model architecture. Instantiating a configuration with the + defaults will yield a similar configuration to that of the ModernBERT-base. + e.g. [answerdotai/ModernBERT-base](https://huggingface.co/answerdotai/ModernBERT-base) + + Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the + documentation from [`PretrainedConfig`] for more information. + + Args: + vocab_size (`int`, *optional*, defaults to 50368): + Vocabulary size of the ModernBert model. Defines the number of different tokens that can be represented by the + `inputs_ids` passed when calling [`ModernBertModel`] + hidden_size (`int`, *optional*, defaults to 768): + Dimension of the hidden representations. + intermediate_size (`int`, *optional*, defaults to 1152): + Dimension of the MLP representations. + num_hidden_layers (`int`, *optional*, defaults to 22): + Number of hidden layers in the Transformer decoder. + num_attention_heads (`int`, *optional*, defaults to 12): + Number of attention heads for each attention layer in the Transformer decoder. + hidden_activation (`str` or `function`, *optional*, defaults to `"gelu"`): + The non-linear activation function (function or string) in the decoder. Will default to `"gelu"` + if not specified. + max_position_embeddings (`int`, *optional*, defaults to 8192): + The maximum sequence length that this model might ever be used with. + initializer_range (`float`, *optional*, defaults to 0.02): + The standard deviation of the truncated_normal_initializer for initializing all weight matrices. + initializer_cutoff_factor (`float`, *optional*, defaults to 2.0): + The cutoff factor for the truncated_normal_initializer for initializing all weight matrices. + norm_eps (`float`, *optional*, defaults to 1e-05): + The epsilon used by the rms normalization layers. + norm_bias (`bool`, *optional*, defaults to `False`): + Whether to use bias in the normalization layers. + pad_token_id (`int`, *optional*, defaults to 50283): + Padding token id. + eos_token_id (`int`, *optional*, defaults to 50282): + End of stream token id. + bos_token_id (`int`, *optional*, defaults to 50281): + Beginning of stream token id. + cls_token_id (`int`, *optional*, defaults to 50281): + Classification token id. + sep_token_id (`int`, *optional*, defaults to 50282): + Separation token id. + global_rope_theta (`float`, *optional*, defaults to 160000.0): + The base period of the global RoPE embeddings. + attention_bias (`bool`, *optional*, defaults to `False`): + Whether to use a bias in the query, key, value and output projection layers during self-attention. + attention_dropout (`float`, *optional*, defaults to 0.0): + The dropout ratio for the attention probabilities. + global_attn_every_n_layers (`int`, *optional*, defaults to 3): + The number of layers between global attention layers. + local_attention (`int`, *optional*, defaults to 128): + The window size for local attention. + local_rope_theta (`float`, *optional*, defaults to 10000.0): + The base period of the local RoPE embeddings. + embedding_dropout (`float`, *optional*, defaults to 0.0): + The dropout ratio for the embeddings. + mlp_bias (`bool`, *optional*, defaults to `False`): + Whether to use bias in the MLP layers. + mlp_dropout (`float`, *optional*, defaults to 0.0): + The dropout ratio for the MLP layers. + decoder_bias (`bool`, *optional*, defaults to `True`): + Whether to use bias in the decoder layers. + classifier_pooling (`str`, *optional*, defaults to `"cls"`): + The pooling method for the classifier. Should be either `"cls"` or `"mean"`. In local attention layers, the + CLS token doesn't attend to all tokens on long sequences. + classifier_dropout (`float`, *optional*, defaults to 0.0): + The dropout ratio for the classifier. + classifier_bias (`bool`, *optional*, defaults to `False`): + Whether to use bias in the classifier. + classifier_activation (`str`, *optional*, defaults to `"gelu"`): + The activation function for the classifier. + deterministic_flash_attn (`bool`, *optional*, defaults to `False`): + Whether to use deterministic flash attention. If `False`, inference will be faster but not deterministic. + sparse_prediction (`bool`, *optional*, defaults to `False`): + Whether to use sparse prediction for the masked language model instead of returning the full dense logits. + sparse_pred_ignore_index (`int`, *optional*, defaults to -100): + The index to ignore for the sparse prediction. + reference_compile (`bool`, *optional*): + Whether to compile the layers of the model which were compiled during pretraining. If `None`, then parts of + the model will be compiled if 1) `triton` is installed, 2) the model is not on MPS, 3) the model is not + shared between devices, and 4) the model is not resized after initialization. If `True`, then the model may + be faster in some scenarios. + repad_logits_with_grad (`bool`, *optional*, defaults to `False`): + When True, ModernBertForMaskedLM keeps track of the logits' gradient when repadding for output. This only + applies when using Flash Attention 2 with passed labels. Otherwise output logits always have a gradient. + + Examples: + + ```python + >>> from transformers import ModernBertModel, ModernBertConfig + + >>> # Initializing a ModernBert style configuration + >>> configuration = ModernBertConfig() + + >>> # Initializing a model from the modernbert-base style configuration + >>> model = ModernBertModel(configuration) + + >>> # Accessing the model configuration + >>> configuration = model.config + ```""" + + model_type = "modernbert" + keys_to_ignore_at_inference = ["past_key_values"] + + def __init__( + self, + vocab_size=50368, + hidden_size=768, + intermediate_size=1152, + num_hidden_layers=22, + num_attention_heads=12, + hidden_activation="gelu", + max_position_embeddings=8192, + initializer_range=0.02, + initializer_cutoff_factor=2.0, + norm_eps=1e-5, + norm_bias=False, + pad_token_id=50283, + eos_token_id=50282, + bos_token_id=50281, + cls_token_id=50281, + sep_token_id=50282, + global_rope_theta=160000.0, + attention_bias=False, + attention_dropout=0.0, + global_attn_every_n_layers=3, + local_attention=128, + local_rope_theta=10000.0, + embedding_dropout=0.0, + mlp_bias=False, + mlp_dropout=0.0, + decoder_bias=True, + classifier_pooling: Literal["cls", "mean"] = "cls", + classifier_dropout=0.0, + classifier_bias=False, + classifier_activation="gelu", + deterministic_flash_attn=False, + sparse_prediction=False, + sparse_pred_ignore_index=-100, + reference_compile=None, + repad_logits_with_grad=False, + **kwargs, + ): + super().__init__( + pad_token_id=pad_token_id, + bos_token_id=bos_token_id, + eos_token_id=eos_token_id, + cls_token_id=cls_token_id, + sep_token_id=sep_token_id, + **kwargs, + ) + self.vocab_size = vocab_size + self.max_position_embeddings = max_position_embeddings + self.hidden_size = hidden_size + self.intermediate_size = intermediate_size + self.num_hidden_layers = num_hidden_layers + self.num_attention_heads = num_attention_heads + self.initializer_range = initializer_range + self.initializer_cutoff_factor = initializer_cutoff_factor + self.norm_eps = norm_eps + self.norm_bias = norm_bias + self.global_rope_theta = global_rope_theta + self.attention_bias = attention_bias + self.attention_dropout = attention_dropout + self.hidden_activation = hidden_activation + self.global_attn_every_n_layers = global_attn_every_n_layers + self.local_attention = local_attention + self.local_rope_theta = local_rope_theta + self.embedding_dropout = embedding_dropout + self.mlp_bias = mlp_bias + self.mlp_dropout = mlp_dropout + self.decoder_bias = decoder_bias + self.classifier_pooling = classifier_pooling + self.classifier_dropout = classifier_dropout + self.classifier_bias = classifier_bias + self.classifier_activation = classifier_activation + self.deterministic_flash_attn = deterministic_flash_attn + self.sparse_prediction = sparse_prediction + self.sparse_pred_ignore_index = sparse_pred_ignore_index + self.reference_compile = reference_compile + self.repad_logits_with_grad = repad_logits_with_grad + + if self.classifier_pooling not in ["cls", "mean"]: + raise ValueError( + f'Invalid value for `classifier_pooling`, should be either "cls" or "mean", but is {self.classifier_pooling}.' + ) + + +__all__ = ["ModernBertConfig"] diff --git a/vllm/lib/python3.10/site-packages/transformers/models/modernbert/modeling_modernbert.py b/vllm/lib/python3.10/site-packages/transformers/models/modernbert/modeling_modernbert.py new file mode 100644 index 0000000000000000000000000000000000000000..3d39041253d849551ecab00b943af65e31a77481 --- /dev/null +++ b/vllm/lib/python3.10/site-packages/transformers/models/modernbert/modeling_modernbert.py @@ -0,0 +1,1393 @@ +# 🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨 +# This file was automatically generated from src/transformers/models/modernbert/modular_modernbert.py. +# Do NOT edit this file manually as any edits will be overwritten by the generation of +# the file from the modular. If any change should be done, please apply the change to the +# modular_modernbert.py file directly. One of our CI enforces this. +# 🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨 +# Copyright 2024 Answer.AI, LightOn, and contributors, and the HuggingFace Inc. team. All rights reserved. +# +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +import math +from contextlib import nullcontext +from typing import Dict, Optional, Tuple, Union + +import torch +import torch.nn.functional as F +from torch import nn +from torch.nn import BCEWithLogitsLoss, CrossEntropyLoss, MSELoss + +from ...activations import ACT2FN +from ...modeling_attn_mask_utils import _prepare_4d_attention_mask +from ...modeling_outputs import BaseModelOutput, MaskedLMOutput, SequenceClassifierOutput, TokenClassifierOutput +from ...modeling_rope_utils import ROPE_INIT_FUNCTIONS +from ...modeling_utils import PreTrainedModel +from ...utils import ( + add_code_sample_docstrings, + add_start_docstrings, + add_start_docstrings_to_model_forward, + is_flash_attn_2_available, + logging, +) +from ...utils.import_utils import is_triton_available +from .configuration_modernbert import ModernBertConfig + + +if is_flash_attn_2_available(): + from flash_attn.flash_attn_interface import flash_attn_varlen_qkvpacked_func + from flash_attn.layers.rotary import RotaryEmbedding + from flash_attn.ops.triton.rotary import apply_rotary +else: + RotaryEmbedding = object + +logger = logging.get_logger(__name__) + +_CHECKPOINT_FOR_DOC = "answerdotai/ModernBERT-base" +_CONFIG_FOR_DOC = "ModernBertConfig" + + +class ApplyRotaryEmbUnpad(torch.autograd.Function): + @staticmethod + def forward( + ctx, + qkv, + cos, + sin, + cu_seqlens: Optional[torch.Tensor] = None, + max_seqlen: Optional[int] = None, + ): + # (total_nnz, 3, nheads, headdim) + qkv = qkv.contiguous() + total_nnz, _three, _nheads, headdim = qkv.shape + # We need qkv to be contiguous so that when we reshape to combine (3, nheads) dimensions, + # we get the same tensor + # qk = rearrange(qkv[:, :2], "b_s t h d -> b_s (t h) d") + qk = qkv[:, :2].view(total_nnz, -1, headdim) + apply_rotary( + qk, + cos, + sin, + seqlen_offsets=0, + cu_seqlens=cu_seqlens, + max_seqlen=max_seqlen, + interleaved=False, + inplace=True, + ) + + ctx.save_for_backward(cos, sin, cu_seqlens) + ctx.max_seqlen = max_seqlen + return qkv + + @staticmethod + def backward(ctx, do): + cos, sin, cu_seqlens = ctx.saved_tensors + do = do.contiguous() + total_nnz, _three, _nheads, headdim = do.shape + # We need dqkv to be contiguous so that when we reshape to combine (3, nheads) dimensions, + # we get the same tensor + dqk = do[:, :2].view(total_nnz, -1, headdim) + apply_rotary( + dqk, + cos, + sin, + seqlen_offsets=0, + cu_seqlens=cu_seqlens, + max_seqlen=ctx.max_seqlen, + interleaved=False, + inplace=True, + conjugate=True, + ) + + return do, None, None, None, None, None, None + + +def apply_rotary_unpadded( + qkv, + cos, + sin, + cu_seqlens: Optional[torch.Tensor] = None, + max_seqlen: Optional[int] = None, +): + """ + Arguments: + qkv: (total_nnz, 3, nheads, headdim) - input tensor for packed QKV. + cos, sin: (seqlen_rotary, rotary_dim / 2) + interleaved: if True, rotate pairs of even and odd dimensions (GPT-J style) instead + of 1st half and 2nd half (GPT-NeoX style). + inplace: if True, apply rotary embedding in-place. + seqlen_offsets: (batch_size,) or int. Each sequence in x is shifted by this amount. + Most commonly used in inference when we have KV cache. + cu_seqlens: (batch + 1,) or None + max_seqlen: int + Return: + out: (total_nnz, dim) + rotary_dim must be <= headdim + Apply rotary embedding to the first rotary_dim of x. + """ + return ApplyRotaryEmbUnpad.apply(qkv, cos, sin, cu_seqlens, max_seqlen) + + +class ModernBertUnpaddedRotaryEmbedding(RotaryEmbedding): + """ + The rotary position embeddings applied directly to unpadded sequences. + """ + + def __init__( + self, + dim: int, + base: float = 10000.0, + max_seqlen: Optional[int] = None, + device: Optional[torch.device] = None, + dtype: Optional[torch.dtype] = None, + ): + """ + max_seqlen: if max_seqlen, device, and dtype are provided, we precompute the cos_sin_cache + up to max_seqlen. If the max_seqlen, device, or dtype during training/inference differ, + the cos_sin_cache wll be recomputed during the forward pass. + """ + super().__init__(dim=dim, base=base, pos_idx_in_fp32=True, device=device, interleaved=False) + self.max_seqlen = max_seqlen + + if max_seqlen is not None and device is not None and dtype is not None: + self._update_cos_sin_cache(max_seqlen, device=device, dtype=dtype) + + def forward( + self, + qkv: torch.Tensor, + cu_seqlens: torch.Tensor, + max_seqlen: Optional[int] = None, + ) -> Union[torch.Tensor, Tuple[torch.Tensor, torch.Tensor]]: + """ + Apply rotary embedding *inplace* to qkv. + qkv: (total_nnz, 3, nheads, headdim) + cu_seqlens: (batch + 1,) cumulative sequence lengths + max_seqlen: int max seq length in the batch + """ + if max_seqlen is not None: + self._update_cos_sin_cache(max_seqlen, device=qkv.device, dtype=qkv.dtype) + + qkv = apply_rotary_unpadded( + qkv, + self._cos_cached, + self._sin_cached, + cu_seqlens=cu_seqlens, + max_seqlen=max_seqlen, + ) + + return qkv + + def extra_repr(self) -> str: + return f"dim={self.dim}, base={self.base}, scale_base={self.scale_base}" + + +class ModernBertEmbeddings(nn.Module): + """ + Same as BertEmbeddings with a tiny tweak for positional embeddings indexing. + """ + + def __init__(self, config: ModernBertConfig): + super().__init__() + self.config = config + self.tok_embeddings = nn.Embedding(config.vocab_size, config.hidden_size, padding_idx=config.pad_token_id) + self.norm = nn.LayerNorm(config.hidden_size, eps=config.norm_eps, bias=config.norm_bias) + self.drop = nn.Dropout(config.embedding_dropout) + + @torch.compile(dynamic=True) + def compiled_embeddings(self, input_ids: torch.LongTensor) -> torch.Tensor: + return self.drop(self.norm(self.tok_embeddings(input_ids))) + + def forward( + self, input_ids: torch.LongTensor = None, inputs_embeds: Optional[torch.Tensor] = None + ) -> torch.Tensor: + if inputs_embeds is not None: + hidden_states = self.drop(self.norm(inputs_embeds)) + else: + hidden_states = ( + self.compiled_embeddings(input_ids) + if self.config.reference_compile + else self.drop(self.norm(self.tok_embeddings(input_ids))) + ) + return hidden_states + + +class ModernBertMLP(nn.Module): + """Applies the GLU at the end of each ModernBERT layer. + + Compared to the default BERT architecture, this block replaces :class:`~transformers.model.bert.modeling_bert.BertIntermediate` + and :class:`~transformers.model.bert.modeling_bert.SelfOutput` with a single module that has similar functionality. + """ + + def __init__(self, config: ModernBertConfig): + super().__init__() + self.config = config + self.Wi = nn.Linear(config.hidden_size, int(config.intermediate_size) * 2, bias=config.mlp_bias) + self.act = ACT2FN[config.hidden_activation] + self.drop = nn.Dropout(config.mlp_dropout) + self.Wo = nn.Linear(config.intermediate_size, config.hidden_size, bias=config.mlp_bias) + + def forward(self, hidden_states: torch.Tensor) -> torch.Tensor: + input, gate = self.Wi(hidden_states).chunk(2, dim=-1) + return self.Wo(self.drop(self.act(input) * gate)) + + +class ModernBertRotaryEmbedding(nn.Module): + def __init__(self, config: ModernBertConfig, dim: int, base: float, device: Optional[torch.device] = None): + super().__init__() + # BC: "rope_type" was originally "type" + if hasattr(config, "rope_scaling") and config.rope_scaling is not None: + self.rope_type = config.rope_scaling.get("rope_type", config.rope_scaling.get("type")) + else: + self.rope_type = "default" + self.max_seq_len_cached = config.max_position_embeddings + self.original_max_seq_len = config.max_position_embeddings + + self.config = config + self.rope_init_fn = ROPE_INIT_FUNCTIONS[self.rope_type] + inv_freq, self.attention_scaling = self.rope_init_fn(None, device, dim=dim, base=base) + self.register_buffer("inv_freq", inv_freq, persistent=False) + self.original_inv_freq = self.inv_freq + + def _dynamic_frequency_update(self, position_ids, device): + """ + dynamic RoPE layers should recompute `inv_freq` in the following situations: + 1 - growing beyond the cached sequence length (allow scaling) + 2 - the current sequence length is in the original scale (avoid losing precision with small sequences) + """ + seq_len = torch.max(position_ids) + 1 + if seq_len > self.max_seq_len_cached: # growth + inv_freq, self.attention_scaling = self.rope_init_fn(self.config, device, seq_len=seq_len) + self.register_buffer("inv_freq", inv_freq, persistent=False) # TODO joao: may break with compilation + self.max_seq_len_cached = seq_len + + if seq_len < self.original_max_seq_len and self.max_seq_len_cached > self.original_max_seq_len: # reset + # This .to() is needed if the model has been moved to a device after being initialized (because + # the buffer is automatically moved, but not the original copy) + self.original_inv_freq = self.original_inv_freq.to(device) + self.register_buffer("inv_freq", self.original_inv_freq, persistent=False) + self.max_seq_len_cached = self.original_max_seq_len + + @torch.no_grad() + def forward(self, x, position_ids): + if "dynamic" in self.rope_type: + self._dynamic_frequency_update(position_ids, device=x.device) + + # Core RoPE block + inv_freq_expanded = self.inv_freq[None, :, None].float().expand(position_ids.shape[0], -1, 1) + position_ids_expanded = position_ids[:, None, :].float() + # Force float32 (see https://github.com/huggingface/transformers/pull/29285) + device_type = x.device.type + device_type = device_type if isinstance(device_type, str) and device_type != "mps" else "cpu" + with torch.autocast(device_type=device_type, enabled=False): + freqs = (inv_freq_expanded.float().to(x.device) @ position_ids_expanded.float()).transpose(1, 2) + emb = torch.cat((freqs, freqs), dim=-1) + cos = emb.cos() + sin = emb.sin() + + # Advanced RoPE types (e.g. yarn) apply a post-processing scaling factor, equivalent to scaling attention + cos = cos * self.attention_scaling + sin = sin * self.attention_scaling + + return cos.to(dtype=x.dtype), sin.to(dtype=x.dtype) + + +def rotate_half(x): + """Rotates half the hidden dims of the input.""" + x1 = x[..., : x.shape[-1] // 2] + x2 = x[..., x.shape[-1] // 2 :] + return torch.cat((-x2, x1), dim=-1) + + +def apply_rotary_pos_emb(q, k, cos, sin, position_ids=None, unsqueeze_dim=1): + """Applies Rotary Position Embedding to the query and key tensors. + + Args: + q (`torch.Tensor`): The query tensor. + k (`torch.Tensor`): The key tensor. + cos (`torch.Tensor`): The cosine part of the rotary embedding. + sin (`torch.Tensor`): The sine part of the rotary embedding. + position_ids (`torch.Tensor`, *optional*): + Deprecated and unused. + unsqueeze_dim (`int`, *optional*, defaults to 1): + The 'unsqueeze_dim' argument specifies the dimension along which to unsqueeze cos[position_ids] and + sin[position_ids] so that they can be properly broadcasted to the dimensions of q and k. For example, note + that cos[position_ids] and sin[position_ids] have the shape [batch_size, seq_len, head_dim]. Then, if q and + k have the shape [batch_size, heads, seq_len, head_dim], then setting unsqueeze_dim=1 makes + cos[position_ids] and sin[position_ids] broadcastable to the shapes of q and k. Similarly, if q and k have + the shape [batch_size, seq_len, heads, head_dim], then set unsqueeze_dim=2. + Returns: + `tuple(torch.Tensor)` comprising of the query and key tensors rotated using the Rotary Position Embedding. + """ + cos = cos.unsqueeze(unsqueeze_dim) + sin = sin.unsqueeze(unsqueeze_dim) + q_embed = (q * cos) + (rotate_half(q) * sin) + k_embed = (k * cos) + (rotate_half(k) * sin) + return q_embed, k_embed + + +def eager_attention_forward( + module: "ModernBertAttention", + qkv: torch.Tensor, + attention_mask: torch.Tensor, + sliding_window_mask: torch.Tensor, + position_ids: Optional[torch.LongTensor], + local_attention: Tuple[int, int], + bs: int, + dim: int, + output_attentions: Optional[bool] = False, + **_kwargs, +) -> Union[Tuple[torch.Tensor, torch.Tensor], Tuple[torch.Tensor]]: + # qkv: [batch_size, seqlen, 3, nheads, headdim] + cos, sin = module.rotary_emb(qkv, position_ids=position_ids) + query, key, value = qkv.transpose(3, 1).unbind(dim=2) + # query, key, value: [batch_size, heads, seq_len, head_dim] + query, key = apply_rotary_pos_emb(query, key, cos, sin) + + scale = module.head_dim**-0.5 + attn_weights = torch.matmul(query, key.transpose(2, 3)) * scale + + if local_attention != (-1, -1): + attention_mask = sliding_window_mask + + attn_weights = attn_weights + attention_mask + + # upcast attention to fp32 + attn_weights = nn.functional.softmax(attn_weights, dim=-1, dtype=torch.float32).to(query.dtype) + attn_weights = nn.functional.dropout(attn_weights, p=module.attention_dropout, training=module.training) + attn_output = torch.matmul(attn_weights, value) + attn_output = attn_output.transpose(1, 2).contiguous() + attn_output = attn_output.view(bs, -1, dim) + if output_attentions: + return (attn_output, attn_weights) + return (attn_output,) + + +def flash_attention_forward( + module: "ModernBertAttention", + qkv: torch.Tensor, + rotary_emb: ModernBertUnpaddedRotaryEmbedding, + cu_seqlens: torch.Tensor, + max_seqlen: int, + local_attention: Tuple[int, int], + bs: int, + dim: int, + target_dtype: torch.dtype = torch.bfloat16, + **_kwargs, +) -> Tuple[torch.Tensor]: + # (total_seqlen, 3, nheads, headdim) + qkv = rotary_emb(qkv, cu_seqlens=cu_seqlens, max_seqlen=max_seqlen) + + convert_dtype = qkv.dtype not in (torch.float16, torch.bfloat16) + if convert_dtype: + # FA2 implementation only supports fp16 and bf16. If FA2 is supported, + # bfloat16 must be supported as of FA2 2.5.7. (Turing GPUs not supported) + orig_dtype = qkv.dtype + qkv = qkv.to(target_dtype) + + attn = flash_attn_varlen_qkvpacked_func( + qkv, + cu_seqlens=cu_seqlens, + max_seqlen=max_seqlen, + dropout_p=module.attention_dropout if module.training else 0.0, + deterministic=module.deterministic_flash_attn, + window_size=local_attention, + ) + attn = attn.to(orig_dtype) # type: ignore + else: + attn = flash_attn_varlen_qkvpacked_func( + qkv, + cu_seqlens=cu_seqlens, + max_seqlen=max_seqlen, + dropout_p=module.attention_dropout if module.training else 0.0, + deterministic=module.deterministic_flash_attn, + window_size=local_attention, + ) + return (attn.view(bs, dim),) + + +def sdpa_attention_forward( + module: "ModernBertAttention", + qkv: torch.Tensor, + attention_mask: torch.Tensor, + sliding_window_mask: torch.Tensor, + position_ids: Optional[torch.LongTensor], + local_attention: Tuple[int, int], + bs: int, + dim: int, + **_kwargs, +) -> Tuple[torch.Tensor]: + # qkv: [batch_size, seqlen, 3, nheads, headdim] + cos, sin = module.rotary_emb(qkv, position_ids=position_ids) + query, key, value = qkv.transpose(3, 1).unbind(dim=2) + # query, key, value: [batch_size, heads, seq_len, head_dim] + query, key = apply_rotary_pos_emb(query, key, cos, sin) + + if local_attention != (-1, -1): + attention_mask = sliding_window_mask + + attn_output = ( + F.scaled_dot_product_attention( + query, + key, + value, + dropout_p=module.attention_dropout if module.training else 0.0, + attn_mask=attention_mask, + ) + .transpose(1, 2) + .contiguous() + ) + attn_output = attn_output.view(bs, -1, dim) + return (attn_output,) + + +MODERNBERT_ATTENTION_FUNCTION = { + "flash_attention_2": flash_attention_forward, + "eager": eager_attention_forward, + "sdpa": sdpa_attention_forward, +} + + +class ModernBertAttention(nn.Module): + """Performs multi-headed self attention on a batch of unpadded sequences. + + If Flash Attention 2 is installed, this module uses Flash Attention to improve throughput. + If Flash Attention 2 is not installed, the implementation will use PyTorch's SDPA kernel, + which requires padding and unpadding inputs, adding some overhead. + + See `forward` method for additional details. + """ + + def __init__(self, config: ModernBertConfig, layer_id: Optional[int] = None): + super().__init__() + self.config = config + self.layer_id = layer_id + + if config.hidden_size % config.num_attention_heads != 0: + raise ValueError( + f"The hidden size ({config.hidden_size}) is not a multiple of the number of attention heads ({config.num_attention_heads})" + ) + + self.attention_dropout = config.attention_dropout + self.deterministic_flash_attn = config.deterministic_flash_attn + self.num_heads = config.num_attention_heads + self.head_dim = config.hidden_size // config.num_attention_heads + self.all_head_size = self.head_dim * self.num_heads + self.Wqkv = nn.Linear(config.hidden_size, 3 * self.all_head_size, bias=config.attention_bias) + + if layer_id % config.global_attn_every_n_layers != 0: + self.local_attention = (config.local_attention // 2, config.local_attention // 2) + else: + self.local_attention = (-1, -1) + + rope_theta = config.global_rope_theta + max_position_embeddings = config.max_position_embeddings + if self.local_attention != (-1, -1): + if config.local_rope_theta is not None: + rope_theta = config.local_rope_theta + max_position_embeddings = config.local_attention + + if config._attn_implementation == "flash_attention_2": + self.rotary_emb = ModernBertUnpaddedRotaryEmbedding( + dim=self.head_dim, max_seqlen=max_position_embeddings, base=rope_theta + ) + else: + self.rotary_emb = ModernBertRotaryEmbedding(config=config, dim=self.head_dim, base=rope_theta) + + self.Wo = nn.Linear(config.hidden_size, config.hidden_size, bias=config.attention_bias) + self.out_drop = nn.Dropout(config.attention_dropout) if config.attention_dropout > 0.0 else nn.Identity() + self.pruned_heads = set() + + def forward( + self, + hidden_states: torch.Tensor, + output_attentions: Optional[bool] = False, + **kwargs, + ) -> torch.Tensor: + qkv = self.Wqkv(hidden_states) + + bs = hidden_states.shape[0] + if self.config._attn_implementation == "flash_attention_2": + qkv = qkv.view(-1, 3, self.num_heads, self.head_dim) + else: + qkv = qkv.view(bs, -1, 3, self.num_heads, self.head_dim) + + attn_outputs = MODERNBERT_ATTENTION_FUNCTION[self.config._attn_implementation]( + self, + qkv=qkv, + rotary_emb=self.rotary_emb, + local_attention=self.local_attention, + bs=bs, + dim=self.all_head_size, + output_attentions=output_attentions, + **kwargs, + ) + hidden_states = attn_outputs[0] + hidden_states = self.out_drop(self.Wo(hidden_states)) + + return (hidden_states,) + attn_outputs[1:] # add attentions if outputted + + +class ModernBertEncoderLayer(nn.Module): + def __init__(self, config: ModernBertConfig, layer_id: Optional[int] = None): + super().__init__() + self.config = config + if layer_id == 0: + self.attn_norm = nn.Identity() + else: + self.attn_norm = nn.LayerNorm(config.hidden_size, eps=config.norm_eps, bias=config.norm_bias) + self.attn = ModernBertAttention(config=config, layer_id=layer_id) + self.mlp_norm = nn.LayerNorm(config.hidden_size, eps=config.norm_eps, bias=config.norm_bias) + self.mlp = ModernBertMLP(config) + + @torch.compile(dynamic=True) + def compiled_mlp(self, hidden_states: torch.Tensor) -> torch.Tensor: + return self.mlp(self.mlp_norm(hidden_states)) + + def forward( + self, + hidden_states: torch.Tensor, + attention_mask: Optional[torch.Tensor] = None, + sliding_window_mask: Optional[torch.Tensor] = None, + position_ids: Optional[torch.LongTensor] = None, + cu_seqlens: Optional[torch.Tensor] = None, + max_seqlen: Optional[int] = None, + output_attentions: Optional[bool] = False, + ) -> torch.Tensor: + attn_outputs = self.attn( + self.attn_norm(hidden_states), + attention_mask=attention_mask, + sliding_window_mask=sliding_window_mask, + position_ids=position_ids, + cu_seqlens=cu_seqlens, + max_seqlen=max_seqlen, + output_attentions=output_attentions, + ) + hidden_states = hidden_states + attn_outputs[0] + mlp_output = ( + self.compiled_mlp(hidden_states) + if self.config.reference_compile + else self.mlp(self.mlp_norm(hidden_states)) + ) + hidden_states = hidden_states + mlp_output + + return (hidden_states,) + attn_outputs[1:] # add attentions if outputted + + +MODERNBERT_START_DOCSTRING = r""" + This model inherits from [`PreTrainedModel`]. Check the superclass documentation for the generic methods the + library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads + etc.) + + This model is also a PyTorch [torch.nn.Module](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) subclass. + Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage + and behavior. + + Parameters: + config ([`ModernBertConfig`]): + Model configuration class with all the parameters of the model. Initializing with a config file does not + load the weights associated with the model, only the configuration. Check out the + [`~PreTrainedModel.from_pretrained`] method to load the model weights. +""" + + +@add_start_docstrings( + "The bare ModernBert Model outputting raw hidden-states without any specific head on top.", + MODERNBERT_START_DOCSTRING, +) +class ModernBertPreTrainedModel(PreTrainedModel): + config_class = ModernBertConfig + base_model_prefix = "model" + supports_gradient_checkpointing = True + _no_split_modules = ["ModernBertEmbeddings", "ModernBertEncoderLayer"] + _supports_flash_attn_2 = True + _supports_sdpa = True + _supports_flex_attn = False + + def _init_weights(self, module: nn.Module): + cutoff_factor = self.config.initializer_cutoff_factor + if cutoff_factor is None: + cutoff_factor = 3 + + def init_weight(module: nn.Module, std: float): + nn.init.trunc_normal_( + module.weight, + mean=0.0, + std=std, + a=-cutoff_factor * std, + b=cutoff_factor * std, + ) + + if isinstance(module, nn.Linear): + if module.bias is not None: + nn.init.zeros_(module.bias) + + stds = { + "in": self.config.initializer_range, + "out": self.config.initializer_range / math.sqrt(2.0 * self.config.num_hidden_layers), + "embedding": self.config.initializer_range, + "final_out": self.config.hidden_size**-0.5, + } + + if isinstance(module, ModernBertEmbeddings): + init_weight(module.tok_embeddings, stds["embedding"]) + elif isinstance(module, ModernBertMLP): + init_weight(module.Wi, stds["in"]) + init_weight(module.Wo, stds["out"]) + elif isinstance(module, ModernBertAttention): + init_weight(module.Wqkv, stds["in"]) + init_weight(module.Wo, stds["out"]) + elif isinstance(module, ModernBertPredictionHead): + init_weight(module.dense, stds["out"]) + elif isinstance(module, ModernBertForMaskedLM): + init_weight(module.decoder, stds["out"]) + elif isinstance(module, (ModernBertForSequenceClassification, ModernBertForTokenClassification)): + init_weight(module.classifier, stds["final_out"]) + + @classmethod + def _autoset_attn_implementation( + cls, + config, + use_flash_attention_2: bool = False, + torch_dtype: Optional[torch.dtype] = None, + device_map: Optional[Union[str, Dict[str, int]]] = None, + check_device_map: bool = True, + ): + # If the user didn't specify anything, try to use flash_attention_2 if available. + # Otherwise we fall back to the default SDPA -> Eager from the super() method. + # ModernBert's FA2 implementation correctly handles non-fp16/bf16 dtypes, we don't + # need the FA2 warning for non-fp16/bf16 dtypes so we set fp16 for the FA2 check. + if config._attn_implementation_internal is None: + config._attn_implementation_internal = "flash_attention_2" + try: + return cls._check_and_enable_flash_attn_2( + config, + torch_dtype=torch.float16, + device_map=device_map, + hard_check_only=False, + check_device_map=check_device_map, + ) + except (ValueError, ImportError): + config._attn_implementation_internal = None + return super()._autoset_attn_implementation( + config, + use_flash_attention_2=use_flash_attention_2, + torch_dtype=torch.float16, + device_map=device_map, + check_device_map=check_device_map, + ) + + def _maybe_set_compile(self): + if self.config.reference_compile is False: + return + + if hasattr(self, "hf_device_map") and len(self.hf_device_map) > 1: + if self.config.reference_compile: + logger.warning_once( + "If `accelerate` split the model across devices, `torch.compile` will not work. " + "Falling back to non-compiled mode." + ) + self.config.reference_compile = False + + if self.device.type == "mps": + if self.config.reference_compile: + logger.warning_once( + "Compiling the model with `torch.compile` and using a `torch.mps` device is not supported. " + "Falling back to non-compiled mode." + ) + self.config.reference_compile = False + + if self.device.type == "cpu": + if self.config.reference_compile: + logger.warning_once( + "Compiling the model with `torch.compile` and using a `torch.cpu` device is not supported. " + "Falling back to non-compiled mode." + ) + self.config.reference_compile = False + + if self.config.reference_compile is None: + self.config.reference_compile = is_triton_available() + + def resize_token_embeddings(self, *args, **kwargs): + model_embeds = super().resize_token_embeddings(*args, **kwargs) + + if self.config.reference_compile in {True, None}: + if self.config.reference_compile: + logger.warning_once( + "Resizing token embeddings with `torch.compile` is not supported. Falling back to non-compiled mode." + ) + self.config.reference_compile = False + + return model_embeds + + +def _unpad_modernbert_input( + inputs: torch.Tensor, + attention_mask: torch.Tensor, + position_ids: Optional[torch.Tensor] = None, + labels: Optional[torch.Tensor] = None, +) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor, int, Optional[torch.Tensor], Optional[torch.Tensor]]: + """ + Remove padding from input sequences. + + Args: + inputs: (batch, seqlen, ...) or (batch, seqlen) + attention_mask: (batch, seqlen), bool / int, 1 means valid and 0 means not valid. + position_ids: (batch, seqlen), int, position ids + labels: (batch, seqlen), int, labels + + Returns: + unpadded_inputs: (total_nnz, ...), where total_nnz = number of tokens selected in attention_mask. + indices: (total_nnz) + cu_seqlens: (batch + 1), the cumulative sequence lengths + max_seqlen_in_batch: int + unpadded_position_ids: (total_nnz) or None + unpadded_labels: (total_nnz) or None + """ + seqlens_in_batch = attention_mask.sum(dim=-1, dtype=torch.int32) + indices = torch.nonzero(attention_mask.flatten(), as_tuple=False).flatten() + max_seqlen_in_batch = int(seqlens_in_batch.max().item()) + cu_seqlens = torch.nn.functional.pad(torch.cumsum(seqlens_in_batch, dim=0, dtype=torch.int32), (1, 0)) + + if inputs.dim() == 2: + unpadded_inputs = inputs.flatten()[indices] + else: + batch, seqlen, *rest = inputs.shape + shape = batch * seqlen + unpadded_inputs = inputs.view(shape, *rest)[indices] + + unpadded_position_ids = position_ids.flatten()[indices] if position_ids is not None else None + unpadded_labels = labels.flatten()[indices] if labels is not None else None + + return unpadded_inputs, indices, cu_seqlens, max_seqlen_in_batch, unpadded_position_ids, unpadded_labels + + +def _pad_modernbert_output( + inputs: torch.Tensor, + indices: torch.Tensor, + batch: int, + seqlen: int, +) -> torch.Tensor: + """ + Add padding to sequences. + + Args: + inputs: (total_nnz, ...) or (total_nnz,), where total_nnz = number of tokens selected in attention_mask. + indices: (total_nnz) + batch: int, batch size + seqlen: int, max sequence length + + Returns: + padded_inputs: (batch, seqlen, ...) or (batch, seqlen) + """ + if inputs.dim() == 1: + output = torch.zeros(batch * seqlen, dtype=inputs.dtype, device=inputs.device) + output[indices] = inputs + padded_inputs = output.view(batch, seqlen) + else: + _, *rest = inputs.shape + output = torch.zeros(batch * seqlen, *rest, dtype=inputs.dtype, device=inputs.device) + output[indices] = inputs + padded_inputs = output.view(batch, seqlen, *rest) + + return padded_inputs + + +MODERNBERT_INPUTS_DOCSTRING = r""" + Args: + input_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`): + Indices of input sequence tokens in the vocabulary. With Flash Attention 2.0, padding will be ignored + by default should you provide it. + + Indices can be obtained using [`AutoTokenizer`]. See [`PreTrainedTokenizer.encode`] and + [`PreTrainedTokenizer.__call__`] for details. + + [What are input IDs?](../glossary#input-ids) + attention_mask (`torch.Tensor` of shape `(batch_size, sequence_length)`, *optional*): + Mask to avoid performing attention on padding token indices. Mask values selected in `[0, 1]`: + + - 1 for tokens that are **not masked**, + - 0 for tokens that are **masked**. + + [What are attention masks?](../glossary#attention-mask) + + Indices can be obtained using [`AutoTokenizer`]. See [`PreTrainedTokenizer.encode`] and + [`PreTrainedTokenizer.__call__`] for details. + + If you want to change padding behavior, you should read [`modeling_opt._prepare_decoder_attention_mask`] + and modify to your needs. See diagram 1 in [the paper](https://arxiv.org/abs/1910.13461) for more + information on the default strategy. + + - 1 indicates the head is **not masked**, + - 0 indicates the head is **masked**. + sliding_window_mask (`torch.Tensor` of shape `(batch_size, sequence_length)`, *optional*): + Mask to avoid performing attention on padding or far-away tokens. In ModernBert, only every few layers + perform global attention, while the rest perform local attention. This mask is used to avoid attending to + far-away tokens in the local attention layers when not using Flash Attention. + position_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*): + Indices of positions of each input sequence tokens in the position embeddings. Selected in the range `[0, + config.n_positions - 1]`. + + [What are position IDs?](../glossary#position-ids) + inputs_embeds (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`, *optional*): + Optionally, instead of passing `input_ids` you can choose to directly pass an embedded representation. This + is useful if you want more control over how to convert `input_ids` indices into associated vectors than the + model's internal embedding lookup matrix. + indices (`torch.Tensor` of shape `(total_unpadded_tokens,)`, *optional*): + Indices of the non-padding tokens in the input sequence. Used for unpadding the output. + cu_seqlens (`torch.Tensor` of shape `(batch + 1,)`, *optional*): + Cumulative sequence lengths of the input sequences. Used to index the unpadded tensors. + max_seqlen (`int`, *optional*): + Maximum sequence length in the batch excluding padding tokens. Used to unpad input_ids and pad output tensors. + batch_size (`int`, *optional*): + Batch size of the input sequences. Used to pad the output tensors. + seq_len (`int`, *optional*): + Sequence length of the input sequences including padding tokens. Used to pad the output tensors. + output_attentions (`bool`, *optional*): + Whether or not to return the attentions tensors of all attention layers. See `attentions` under returned + tensors for more detail. + output_hidden_states (`bool`, *optional*): + Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for + more detail. + return_dict (`bool`, *optional*): + Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple. +""" + + +@add_start_docstrings( + "The bare ModernBert Model outputting raw hidden-states without any specific head on top.", + MODERNBERT_START_DOCSTRING, +) +class ModernBertModel(ModernBertPreTrainedModel): + def __init__(self, config: ModernBertConfig): + super().__init__(config) + self.config = config + self.embeddings = ModernBertEmbeddings(config) + self.layers = nn.ModuleList( + [ModernBertEncoderLayer(config, layer_id) for layer_id in range(config.num_hidden_layers)] + ) + self.final_norm = nn.LayerNorm(config.hidden_size, eps=config.norm_eps, bias=config.norm_bias) + self.gradient_checkpointing = False + self.post_init() + + def get_input_embeddings(self): + return self.embeddings.tok_embeddings + + def set_input_embeddings(self, value): + self.embeddings.tok_embeddings = value + + @add_start_docstrings_to_model_forward(MODERNBERT_INPUTS_DOCSTRING) + @add_code_sample_docstrings( + checkpoint=_CHECKPOINT_FOR_DOC, + output_type=BaseModelOutput, + config_class=_CONFIG_FOR_DOC, + ) + def forward( + self, + input_ids: Optional[torch.LongTensor] = None, + attention_mask: Optional[torch.Tensor] = None, + sliding_window_mask: Optional[torch.Tensor] = None, + position_ids: Optional[torch.LongTensor] = None, + inputs_embeds: Optional[torch.Tensor] = None, + indices: Optional[torch.Tensor] = None, + cu_seqlens: Optional[torch.Tensor] = None, + max_seqlen: Optional[int] = None, + batch_size: Optional[int] = None, + seq_len: Optional[int] = None, + output_attentions: Optional[bool] = None, + output_hidden_states: Optional[bool] = None, + return_dict: Optional[bool] = None, + ) -> Union[Tuple[torch.Tensor, ...], BaseModelOutput]: + output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions + output_hidden_states = ( + output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states + ) + return_dict = return_dict if return_dict is not None else self.config.use_return_dict + + if (input_ids is None) ^ (inputs_embeds is not None): + raise ValueError("You must specify exactly one of input_ids or inputs_embeds") + + all_hidden_states = () if output_hidden_states else None + all_self_attentions = () if output_attentions else None + + self._maybe_set_compile() + + if input_ids is not None: + self.warn_if_padding_and_no_attention_mask(input_ids, attention_mask) + + if batch_size is None and seq_len is None: + if inputs_embeds is not None: + batch_size, seq_len = inputs_embeds.shape[:2] + else: + batch_size, seq_len = input_ids.shape[:2] + device = input_ids.device if input_ids is not None else inputs_embeds.device + + if attention_mask is None: + attention_mask = torch.ones((batch_size, seq_len), device=device, dtype=torch.bool) + + repad = False + if self.config._attn_implementation == "flash_attention_2": + if indices is None and cu_seqlens is None and max_seqlen is None: + repad = True + if inputs_embeds is None: + with torch.no_grad(): + input_ids, indices, cu_seqlens, max_seqlen, *_ = _unpad_modernbert_input( + inputs=input_ids, attention_mask=attention_mask + ) + else: + inputs_embeds, indices, cu_seqlens, max_seqlen, *_ = _unpad_modernbert_input( + inputs=inputs_embeds, attention_mask=attention_mask + ) + else: + if position_ids is None: + position_ids = torch.arange(seq_len, device=device).unsqueeze(0) + + attention_mask, sliding_window_mask = self._update_attention_mask( + attention_mask, output_attentions=output_attentions + ) + + hidden_states = self.embeddings(input_ids=input_ids, inputs_embeds=inputs_embeds) + + for encoder_layer in self.layers: + if output_hidden_states: + all_hidden_states = all_hidden_states + (hidden_states,) + + if self.gradient_checkpointing and self.training: + layer_outputs = self._gradient_checkpointing_func( + encoder_layer.__call__, + hidden_states, + attention_mask, + sliding_window_mask, + position_ids, + cu_seqlens, + max_seqlen, + output_attentions, + ) + else: + layer_outputs = encoder_layer( + hidden_states, + attention_mask=attention_mask, + sliding_window_mask=sliding_window_mask, + position_ids=position_ids, + cu_seqlens=cu_seqlens, + max_seqlen=max_seqlen, + output_attentions=output_attentions, + ) + hidden_states = layer_outputs[0] + if output_attentions and len(layer_outputs) > 1: + all_self_attentions = all_self_attentions + (layer_outputs[1],) + + if output_hidden_states: + all_hidden_states = all_hidden_states + (hidden_states,) + + hidden_states = self.final_norm(hidden_states) + + if repad: + hidden_states = _pad_modernbert_output( + inputs=hidden_states, indices=indices, batch=batch_size, seqlen=seq_len + ) + if all_hidden_states is not None: + all_hidden_states = tuple( + _pad_modernbert_output(inputs=hs, indices=indices, batch=batch_size, seqlen=seq_len) + for hs in all_hidden_states + ) + + if not return_dict: + return tuple(v for v in [hidden_states, all_hidden_states, all_self_attentions] if v is not None) + return BaseModelOutput( + last_hidden_state=hidden_states, + hidden_states=all_hidden_states, + attentions=all_self_attentions, + ) + + def _update_attention_mask(self, attention_mask: torch.Tensor, output_attentions: bool) -> torch.Tensor: + if output_attentions: + if self.config._attn_implementation == "sdpa": + logger.warning_once( + "Outputting attentions is only supported with the 'eager' attention implementation, " + 'not with "sdpa". Falling back to `attn_implementation="eager"`.' + ) + self.config._attn_implementation = "eager" + elif self.config._attn_implementation != "eager": + logger.warning_once( + "Outputting attentions is only supported with the eager attention implementation, " + f'not with {self.config._attn_implementation}. Consider setting `attn_implementation="eager"`.' + " Setting `output_attentions=False`." + ) + + global_attention_mask = _prepare_4d_attention_mask(attention_mask, self.dtype) + + # Create position indices + rows = torch.arange(global_attention_mask.shape[2]).unsqueeze(0) + # Calculate distance between positions + distance = torch.abs(rows - rows.T) + + # Create sliding window mask (1 for positions within window, 0 outside) + window_mask = ( + (distance <= self.config.local_attention // 2).unsqueeze(0).unsqueeze(0).to(attention_mask.device) + ) + # Combine with existing mask + sliding_window_mask = global_attention_mask.masked_fill(window_mask.logical_not(), torch.finfo(self.dtype).min) + + return global_attention_mask, sliding_window_mask + + +class ModernBertPredictionHead(nn.Module): + def __init__(self, config: ModernBertConfig): + super().__init__() + self.config = config + self.dense = nn.Linear(config.hidden_size, config.hidden_size, config.classifier_bias) + self.act = ACT2FN[config.classifier_activation] + self.norm = nn.LayerNorm(config.hidden_size, eps=config.norm_eps, bias=config.norm_bias) + + def forward(self, hidden_states: torch.Tensor) -> torch.Tensor: + return self.norm(self.act(self.dense(hidden_states))) + + +@add_start_docstrings( + "The ModernBert Model with a decoder head on top that is used for masked language modeling.", + MODERNBERT_START_DOCSTRING, +) +class ModernBertForMaskedLM(ModernBertPreTrainedModel): + _tied_weights_keys = ["decoder.weight"] + + def __init__(self, config: ModernBertConfig): + super().__init__(config) + self.config = config + self.model = ModernBertModel(config) + self.head = ModernBertPredictionHead(config) + self.decoder = nn.Linear(config.hidden_size, config.vocab_size, bias=config.decoder_bias) + + self.sparse_prediction = self.config.sparse_prediction + self.sparse_pred_ignore_index = self.config.sparse_pred_ignore_index + + # Initialize weights and apply final processing + self.post_init() + + def get_output_embeddings(self): + return self.decoder + + def set_output_embeddings(self, new_embeddings: nn.Linear): + self.decoder = new_embeddings + + @torch.compile(dynamic=True) + def compiled_head(self, output: torch.Tensor) -> torch.Tensor: + return self.decoder(self.head(output)) + + @add_start_docstrings_to_model_forward(MODERNBERT_INPUTS_DOCSTRING) + @add_code_sample_docstrings( + checkpoint=_CHECKPOINT_FOR_DOC, + output_type=MaskedLMOutput, + config_class=_CONFIG_FOR_DOC, + ) + def forward( + self, + input_ids: Optional[torch.LongTensor] = None, + attention_mask: Optional[torch.Tensor] = None, + sliding_window_mask: Optional[torch.Tensor] = None, + position_ids: Optional[torch.Tensor] = None, + inputs_embeds: Optional[torch.Tensor] = None, + labels: Optional[torch.Tensor] = None, + indices: Optional[torch.Tensor] = None, + cu_seqlens: Optional[torch.Tensor] = None, + max_seqlen: Optional[int] = None, + batch_size: Optional[int] = None, + seq_len: Optional[int] = None, + output_attentions: Optional[bool] = None, + output_hidden_states: Optional[bool] = None, + return_dict: Optional[bool] = None, + **kwargs, + ) -> Union[Tuple[torch.Tensor], MaskedLMOutput]: + return_dict = return_dict if return_dict is not None else self.config.use_return_dict + self._maybe_set_compile() + + if self.config._attn_implementation == "flash_attention_2": + if indices is None and cu_seqlens is None and max_seqlen is None: + if batch_size is None and seq_len is None: + if inputs_embeds is not None: + batch_size, seq_len = inputs_embeds.shape[:2] + else: + batch_size, seq_len = input_ids.shape[:2] + device = input_ids.device if input_ids is not None else inputs_embeds.device + + if attention_mask is None: + attention_mask = torch.ones((batch_size, seq_len), device=device, dtype=torch.bool) + + if inputs_embeds is None: + with torch.no_grad(): + input_ids, indices, cu_seqlens, max_seqlen, position_ids, labels = _unpad_modernbert_input( + inputs=input_ids, attention_mask=attention_mask, position_ids=position_ids, labels=labels + ) + else: + inputs_embeds, indices, cu_seqlens, max_seqlen, position_ids, labels = _unpad_modernbert_input( + inputs=inputs_embeds, attention_mask=attention_mask, position_ids=position_ids, labels=labels + ) + + outputs = self.model( + input_ids=input_ids, + attention_mask=attention_mask, + sliding_window_mask=sliding_window_mask, + position_ids=position_ids, + inputs_embeds=inputs_embeds, + indices=indices, + cu_seqlens=cu_seqlens, + max_seqlen=max_seqlen, + batch_size=batch_size, + seq_len=seq_len, + output_attentions=output_attentions, + output_hidden_states=output_hidden_states, + return_dict=return_dict, + ) + last_hidden_state = outputs[0] + + if self.sparse_prediction and labels is not None: + # flatten labels and output first + labels = labels.view(-1) + last_hidden_state = last_hidden_state.view(labels.shape[0], -1) + + # then filter out the non-masked tokens + mask_tokens = labels != self.sparse_pred_ignore_index + last_hidden_state = last_hidden_state[mask_tokens] + labels = labels[mask_tokens] + + logits = ( + self.compiled_head(last_hidden_state) + if self.config.reference_compile + else self.decoder(self.head(last_hidden_state)) + ) + + loss = None + if labels is not None: + loss = self.loss_function(logits, labels, vocab_size=self.config.vocab_size) + + if self.config._attn_implementation == "flash_attention_2": + with nullcontext() if self.config.repad_logits_with_grad or labels is None else torch.no_grad(): + logits = _pad_modernbert_output(inputs=logits, indices=indices, batch=batch_size, seqlen=seq_len) + + if not return_dict: + output = (logits,) + return ((loss,) + output) if loss is not None else output + + return MaskedLMOutput( + loss=loss, + logits=logits, + hidden_states=outputs.hidden_states, + attentions=outputs.attentions, + ) + + +@add_start_docstrings( + "The ModernBert Model with a sequence classification head on top that performs pooling.", + MODERNBERT_START_DOCSTRING, +) +class ModernBertForSequenceClassification(ModernBertPreTrainedModel): + def __init__(self, config: ModernBertConfig): + super().__init__(config) + self.num_labels = config.num_labels + self.config = config + + self.model = ModernBertModel(config) + self.head = ModernBertPredictionHead(config) + self.drop = torch.nn.Dropout(config.classifier_dropout) + self.classifier = nn.Linear(config.hidden_size, config.num_labels) + + # Initialize weights and apply final processing + self.post_init() + + @add_start_docstrings_to_model_forward(MODERNBERT_INPUTS_DOCSTRING) + @add_code_sample_docstrings( + checkpoint=_CHECKPOINT_FOR_DOC, + output_type=SequenceClassifierOutput, + config_class=_CONFIG_FOR_DOC, + ) + def forward( + self, + input_ids: Optional[torch.LongTensor] = None, + attention_mask: Optional[torch.Tensor] = None, + sliding_window_mask: Optional[torch.Tensor] = None, + position_ids: Optional[torch.Tensor] = None, + inputs_embeds: Optional[torch.Tensor] = None, + labels: Optional[torch.Tensor] = None, + indices: Optional[torch.Tensor] = None, + cu_seqlens: Optional[torch.Tensor] = None, + max_seqlen: Optional[int] = None, + batch_size: Optional[int] = None, + seq_len: Optional[int] = None, + output_attentions: Optional[bool] = None, + output_hidden_states: Optional[bool] = None, + return_dict: Optional[bool] = None, + **kwargs, + ) -> Union[Tuple[torch.Tensor], SequenceClassifierOutput]: + r""" + labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*): + Labels for computing the sequence classification/regression loss. Indices should be in `[0, ..., + config.num_labels - 1]`. If `config.num_labels == 1` a regression loss is computed (Mean-Square loss), If + `config.num_labels > 1` a classification loss is computed (Cross-Entropy). + """ + return_dict = return_dict if return_dict is not None else self.config.use_return_dict + self._maybe_set_compile() + + outputs = self.model( + input_ids=input_ids, + attention_mask=attention_mask, + sliding_window_mask=sliding_window_mask, + position_ids=position_ids, + inputs_embeds=inputs_embeds, + indices=indices, + cu_seqlens=cu_seqlens, + max_seqlen=max_seqlen, + batch_size=batch_size, + seq_len=seq_len, + output_attentions=output_attentions, + output_hidden_states=output_hidden_states, + return_dict=return_dict, + ) + last_hidden_state = outputs[0] + + if self.config.classifier_pooling == "cls": + last_hidden_state = last_hidden_state[:, 0] + elif self.config.classifier_pooling == "mean": + last_hidden_state = (last_hidden_state * attention_mask.unsqueeze(-1)).sum(dim=1) / attention_mask.sum( + dim=1, keepdim=True + ) + + pooled_output = self.head(last_hidden_state) + pooled_output = self.drop(pooled_output) + logits = self.classifier(pooled_output) + + loss = None + if labels is not None: + if self.config.problem_type is None: + if self.num_labels == 1: + self.config.problem_type = "regression" + elif self.num_labels > 1 and (labels.dtype == torch.long or labels.dtype == torch.int): + self.config.problem_type = "single_label_classification" + else: + self.config.problem_type = "multi_label_classification" + + if self.config.problem_type == "regression": + loss_fct = MSELoss() + if self.num_labels == 1: + loss = loss_fct(logits.squeeze(), labels.squeeze()) + else: + loss = loss_fct(logits, labels) + elif self.config.problem_type == "single_label_classification": + loss_fct = CrossEntropyLoss() + loss = loss_fct(logits.view(-1, self.num_labels), labels.view(-1)) + elif self.config.problem_type == "multi_label_classification": + loss_fct = BCEWithLogitsLoss() + loss = loss_fct(logits, labels) + + if not return_dict: + output = (logits,) + return ((loss,) + output) if loss is not None else output + + return SequenceClassifierOutput( + loss=loss, + logits=logits, + hidden_states=outputs.hidden_states, + attentions=outputs.attentions, + ) + + +@add_start_docstrings( + "The ModernBert Model with a token classification head on top, e.g. for Named Entity Recognition (NER) tasks.", + MODERNBERT_START_DOCSTRING, +) +class ModernBertForTokenClassification(ModernBertPreTrainedModel): + def __init__(self, config: ModernBertConfig): + super().__init__(config) + self.num_labels = config.num_labels + + self.model = ModernBertModel(config) + self.head = ModernBertPredictionHead(config) + self.drop = torch.nn.Dropout(config.classifier_dropout) + self.classifier = nn.Linear(config.hidden_size, config.num_labels) + + # Initialize weights and apply final processing + self.post_init() + + @add_start_docstrings_to_model_forward(MODERNBERT_INPUTS_DOCSTRING) + @add_code_sample_docstrings( + checkpoint=_CHECKPOINT_FOR_DOC, + output_type=TokenClassifierOutput, + config_class=_CONFIG_FOR_DOC, + ) + def forward( + self, + input_ids: Optional[torch.LongTensor] = None, + attention_mask: Optional[torch.Tensor] = None, + sliding_window_mask: Optional[torch.Tensor] = None, + position_ids: Optional[torch.Tensor] = None, + inputs_embeds: Optional[torch.Tensor] = None, + labels: Optional[torch.Tensor] = None, + indices: Optional[torch.Tensor] = None, + cu_seqlens: Optional[torch.Tensor] = None, + max_seqlen: Optional[int] = None, + batch_size: Optional[int] = None, + seq_len: Optional[int] = None, + output_attentions: Optional[bool] = None, + output_hidden_states: Optional[bool] = None, + return_dict: Optional[bool] = None, + ) -> Union[Tuple[torch.Tensor], TokenClassifierOutput]: + r""" + labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*): + Labels for computing the token classification loss. Indices should be in `[0, ..., config.num_labels - 1]`. + """ + return_dict = return_dict if return_dict is not None else self.config.use_return_dict + self._maybe_set_compile() + + outputs = self.model( + input_ids=input_ids, + attention_mask=attention_mask, + sliding_window_mask=sliding_window_mask, + position_ids=position_ids, + inputs_embeds=inputs_embeds, + indices=indices, + cu_seqlens=cu_seqlens, + max_seqlen=max_seqlen, + batch_size=batch_size, + seq_len=seq_len, + output_attentions=output_attentions, + output_hidden_states=output_hidden_states, + return_dict=return_dict, + ) + last_hidden_state = outputs[0] + + last_hidden_state = self.head(last_hidden_state) + last_hidden_state = self.drop(last_hidden_state) + logits = self.classifier(last_hidden_state) + + loss = None + if labels is not None: + loss_fct = CrossEntropyLoss() + loss = loss_fct(logits.view(-1, self.num_labels), labels.view(-1)) + + if not return_dict: + output = (logits,) + outputs[1:] + return ((loss,) + output) if loss is not None else output + + return TokenClassifierOutput( + loss=loss, + logits=logits, + hidden_states=outputs.hidden_states, + attentions=outputs.attentions, + ) + + +__all__ = [ + "ModernBertModel", + "ModernBertPreTrainedModel", + "ModernBertForMaskedLM", + "ModernBertForSequenceClassification", + "ModernBertForTokenClassification", +] diff --git a/vllm/lib/python3.10/site-packages/transformers/models/modernbert/modular_modernbert.py b/vllm/lib/python3.10/site-packages/transformers/models/modernbert/modular_modernbert.py new file mode 100644 index 0000000000000000000000000000000000000000..edfdc94346bf122b71f25f981dc30af89bfc5c7a --- /dev/null +++ b/vllm/lib/python3.10/site-packages/transformers/models/modernbert/modular_modernbert.py @@ -0,0 +1,1497 @@ +# Copyright 2024 Answer.AI, LightOn, and contributors, and the HuggingFace Inc. team. All rights reserved. +# +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +import math +from contextlib import nullcontext +from typing import Dict, Literal, Optional, Tuple, Union + +import torch +import torch.nn.functional as F +import torch.utils.checkpoint +from torch import nn +from torch.nn import BCEWithLogitsLoss, CrossEntropyLoss, MSELoss + +from ...activations import ACT2FN +from ...configuration_utils import PretrainedConfig +from ...modeling_attn_mask_utils import _prepare_4d_attention_mask +from ...modeling_outputs import ( + BaseModelOutput, + MaskedLMOutput, + SequenceClassifierOutput, + TokenClassifierOutput, +) +from ...modeling_utils import PreTrainedModel +from ...utils import ( + add_code_sample_docstrings, + add_start_docstrings, + add_start_docstrings_to_model_forward, + is_flash_attn_2_available, + logging, +) +from ...utils.import_utils import is_triton_available +from ..gemma.modeling_gemma import GemmaRotaryEmbedding, apply_rotary_pos_emb + + +if is_flash_attn_2_available(): + from flash_attn.flash_attn_interface import flash_attn_varlen_qkvpacked_func + from flash_attn.layers.rotary import RotaryEmbedding + from flash_attn.ops.triton.rotary import apply_rotary +else: + RotaryEmbedding = object + +_CHECKPOINT_FOR_DOC = "answerdotai/ModernBERT-base" +_CONFIG_FOR_DOC = "ModernBertConfig" + +logger = logging.get_logger(__name__) + + +class ModernBertConfig(PretrainedConfig): + r""" + This is the configuration class to store the configuration of a [`ModernBertModel`]. It is used to instantiate an ModernBert + model according to the specified arguments, defining the model architecture. Instantiating a configuration with the + defaults will yield a similar configuration to that of the ModernBERT-base. + e.g. [answerdotai/ModernBERT-base](https://huggingface.co/answerdotai/ModernBERT-base) + + Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the + documentation from [`PretrainedConfig`] for more information. + + Args: + vocab_size (`int`, *optional*, defaults to 50368): + Vocabulary size of the ModernBert model. Defines the number of different tokens that can be represented by the + `inputs_ids` passed when calling [`ModernBertModel`] + hidden_size (`int`, *optional*, defaults to 768): + Dimension of the hidden representations. + intermediate_size (`int`, *optional*, defaults to 1152): + Dimension of the MLP representations. + num_hidden_layers (`int`, *optional*, defaults to 22): + Number of hidden layers in the Transformer decoder. + num_attention_heads (`int`, *optional*, defaults to 12): + Number of attention heads for each attention layer in the Transformer decoder. + hidden_activation (`str` or `function`, *optional*, defaults to `"gelu"`): + The non-linear activation function (function or string) in the decoder. Will default to `"gelu"` + if not specified. + max_position_embeddings (`int`, *optional*, defaults to 8192): + The maximum sequence length that this model might ever be used with. + initializer_range (`float`, *optional*, defaults to 0.02): + The standard deviation of the truncated_normal_initializer for initializing all weight matrices. + initializer_cutoff_factor (`float`, *optional*, defaults to 2.0): + The cutoff factor for the truncated_normal_initializer for initializing all weight matrices. + norm_eps (`float`, *optional*, defaults to 1e-05): + The epsilon used by the rms normalization layers. + norm_bias (`bool`, *optional*, defaults to `False`): + Whether to use bias in the normalization layers. + pad_token_id (`int`, *optional*, defaults to 50283): + Padding token id. + eos_token_id (`int`, *optional*, defaults to 50282): + End of stream token id. + bos_token_id (`int`, *optional*, defaults to 50281): + Beginning of stream token id. + cls_token_id (`int`, *optional*, defaults to 50281): + Classification token id. + sep_token_id (`int`, *optional*, defaults to 50282): + Separation token id. + global_rope_theta (`float`, *optional*, defaults to 160000.0): + The base period of the global RoPE embeddings. + attention_bias (`bool`, *optional*, defaults to `False`): + Whether to use a bias in the query, key, value and output projection layers during self-attention. + attention_dropout (`float`, *optional*, defaults to 0.0): + The dropout ratio for the attention probabilities. + global_attn_every_n_layers (`int`, *optional*, defaults to 3): + The number of layers between global attention layers. + local_attention (`int`, *optional*, defaults to 128): + The window size for local attention. + local_rope_theta (`float`, *optional*, defaults to 10000.0): + The base period of the local RoPE embeddings. + embedding_dropout (`float`, *optional*, defaults to 0.0): + The dropout ratio for the embeddings. + mlp_bias (`bool`, *optional*, defaults to `False`): + Whether to use bias in the MLP layers. + mlp_dropout (`float`, *optional*, defaults to 0.0): + The dropout ratio for the MLP layers. + decoder_bias (`bool`, *optional*, defaults to `True`): + Whether to use bias in the decoder layers. + classifier_pooling (`str`, *optional*, defaults to `"cls"`): + The pooling method for the classifier. Should be either `"cls"` or `"mean"`. In local attention layers, the + CLS token doesn't attend to all tokens on long sequences. + classifier_dropout (`float`, *optional*, defaults to 0.0): + The dropout ratio for the classifier. + classifier_bias (`bool`, *optional*, defaults to `False`): + Whether to use bias in the classifier. + classifier_activation (`str`, *optional*, defaults to `"gelu"`): + The activation function for the classifier. + deterministic_flash_attn (`bool`, *optional*, defaults to `False`): + Whether to use deterministic flash attention. If `False`, inference will be faster but not deterministic. + sparse_prediction (`bool`, *optional*, defaults to `False`): + Whether to use sparse prediction for the masked language model instead of returning the full dense logits. + sparse_pred_ignore_index (`int`, *optional*, defaults to -100): + The index to ignore for the sparse prediction. + reference_compile (`bool`, *optional*): + Whether to compile the layers of the model which were compiled during pretraining. If `None`, then parts of + the model will be compiled if 1) `triton` is installed, 2) the model is not on MPS, 3) the model is not + shared between devices, and 4) the model is not resized after initialization. If `True`, then the model may + be faster in some scenarios. + repad_logits_with_grad (`bool`, *optional*, defaults to `False`): + When True, ModernBertForMaskedLM keeps track of the logits' gradient when repadding for output. This only + applies when using Flash Attention 2 with passed labels. Otherwise output logits always have a gradient. + + Examples: + + ```python + >>> from transformers import ModernBertModel, ModernBertConfig + + >>> # Initializing a ModernBert style configuration + >>> configuration = ModernBertConfig() + + >>> # Initializing a model from the modernbert-base style configuration + >>> model = ModernBertModel(configuration) + + >>> # Accessing the model configuration + >>> configuration = model.config + ```""" + + model_type = "modernbert" + keys_to_ignore_at_inference = ["past_key_values"] + + def __init__( + self, + vocab_size=50368, + hidden_size=768, + intermediate_size=1152, + num_hidden_layers=22, + num_attention_heads=12, + hidden_activation="gelu", + max_position_embeddings=8192, + initializer_range=0.02, + initializer_cutoff_factor=2.0, + norm_eps=1e-5, + norm_bias=False, + pad_token_id=50283, + eos_token_id=50282, + bos_token_id=50281, + cls_token_id=50281, + sep_token_id=50282, + global_rope_theta=160000.0, + attention_bias=False, + attention_dropout=0.0, + global_attn_every_n_layers=3, + local_attention=128, + local_rope_theta=10000.0, + embedding_dropout=0.0, + mlp_bias=False, + mlp_dropout=0.0, + decoder_bias=True, + classifier_pooling: Literal["cls", "mean"] = "cls", + classifier_dropout=0.0, + classifier_bias=False, + classifier_activation="gelu", + deterministic_flash_attn=False, + sparse_prediction=False, + sparse_pred_ignore_index=-100, + reference_compile=None, + repad_logits_with_grad=False, + **kwargs, + ): + super().__init__( + pad_token_id=pad_token_id, + bos_token_id=bos_token_id, + eos_token_id=eos_token_id, + cls_token_id=cls_token_id, + sep_token_id=sep_token_id, + **kwargs, + ) + self.vocab_size = vocab_size + self.max_position_embeddings = max_position_embeddings + self.hidden_size = hidden_size + self.intermediate_size = intermediate_size + self.num_hidden_layers = num_hidden_layers + self.num_attention_heads = num_attention_heads + self.initializer_range = initializer_range + self.initializer_cutoff_factor = initializer_cutoff_factor + self.norm_eps = norm_eps + self.norm_bias = norm_bias + self.global_rope_theta = global_rope_theta + self.attention_bias = attention_bias + self.attention_dropout = attention_dropout + self.hidden_activation = hidden_activation + self.global_attn_every_n_layers = global_attn_every_n_layers + self.local_attention = local_attention + self.local_rope_theta = local_rope_theta + self.embedding_dropout = embedding_dropout + self.mlp_bias = mlp_bias + self.mlp_dropout = mlp_dropout + self.decoder_bias = decoder_bias + self.classifier_pooling = classifier_pooling + self.classifier_dropout = classifier_dropout + self.classifier_bias = classifier_bias + self.classifier_activation = classifier_activation + self.deterministic_flash_attn = deterministic_flash_attn + self.sparse_prediction = sparse_prediction + self.sparse_pred_ignore_index = sparse_pred_ignore_index + self.reference_compile = reference_compile + self.repad_logits_with_grad = repad_logits_with_grad + + if self.classifier_pooling not in ["cls", "mean"]: + raise ValueError( + f'Invalid value for `classifier_pooling`, should be either "cls" or "mean", but is {self.classifier_pooling}.' + ) + + +def _unpad_modernbert_input( + inputs: torch.Tensor, + attention_mask: torch.Tensor, + position_ids: Optional[torch.Tensor] = None, + labels: Optional[torch.Tensor] = None, +) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor, int, Optional[torch.Tensor], Optional[torch.Tensor]]: + """ + Remove padding from input sequences. + + Args: + inputs: (batch, seqlen, ...) or (batch, seqlen) + attention_mask: (batch, seqlen), bool / int, 1 means valid and 0 means not valid. + position_ids: (batch, seqlen), int, position ids + labels: (batch, seqlen), int, labels + + Returns: + unpadded_inputs: (total_nnz, ...), where total_nnz = number of tokens selected in attention_mask. + indices: (total_nnz) + cu_seqlens: (batch + 1), the cumulative sequence lengths + max_seqlen_in_batch: int + unpadded_position_ids: (total_nnz) or None + unpadded_labels: (total_nnz) or None + """ + seqlens_in_batch = attention_mask.sum(dim=-1, dtype=torch.int32) + indices = torch.nonzero(attention_mask.flatten(), as_tuple=False).flatten() + max_seqlen_in_batch = int(seqlens_in_batch.max().item()) + cu_seqlens = torch.nn.functional.pad(torch.cumsum(seqlens_in_batch, dim=0, dtype=torch.int32), (1, 0)) + + if inputs.dim() == 2: + unpadded_inputs = inputs.flatten()[indices] + else: + batch, seqlen, *rest = inputs.shape + shape = batch * seqlen + unpadded_inputs = inputs.view(shape, *rest)[indices] + + unpadded_position_ids = position_ids.flatten()[indices] if position_ids is not None else None + unpadded_labels = labels.flatten()[indices] if labels is not None else None + + return unpadded_inputs, indices, cu_seqlens, max_seqlen_in_batch, unpadded_position_ids, unpadded_labels + + +def _pad_modernbert_output( + inputs: torch.Tensor, + indices: torch.Tensor, + batch: int, + seqlen: int, +) -> torch.Tensor: + """ + Add padding to sequences. + + Args: + inputs: (total_nnz, ...) or (total_nnz,), where total_nnz = number of tokens selected in attention_mask. + indices: (total_nnz) + batch: int, batch size + seqlen: int, max sequence length + + Returns: + padded_inputs: (batch, seqlen, ...) or (batch, seqlen) + """ + if inputs.dim() == 1: + output = torch.zeros(batch * seqlen, dtype=inputs.dtype, device=inputs.device) + output[indices] = inputs + padded_inputs = output.view(batch, seqlen) + else: + _, *rest = inputs.shape + output = torch.zeros(batch * seqlen, *rest, dtype=inputs.dtype, device=inputs.device) + output[indices] = inputs + padded_inputs = output.view(batch, seqlen, *rest) + + return padded_inputs + + +class ApplyRotaryEmbUnpad(torch.autograd.Function): + @staticmethod + def forward( + ctx, + qkv, + cos, + sin, + cu_seqlens: Optional[torch.Tensor] = None, + max_seqlen: Optional[int] = None, + ): + # (total_nnz, 3, nheads, headdim) + qkv = qkv.contiguous() + total_nnz, _three, _nheads, headdim = qkv.shape + # We need qkv to be contiguous so that when we reshape to combine (3, nheads) dimensions, + # we get the same tensor + # qk = rearrange(qkv[:, :2], "b_s t h d -> b_s (t h) d") + qk = qkv[:, :2].view(total_nnz, -1, headdim) + apply_rotary( + qk, + cos, + sin, + seqlen_offsets=0, + cu_seqlens=cu_seqlens, + max_seqlen=max_seqlen, + interleaved=False, + inplace=True, + ) + + ctx.save_for_backward(cos, sin, cu_seqlens) + ctx.max_seqlen = max_seqlen + return qkv + + @staticmethod + def backward(ctx, do): + cos, sin, cu_seqlens = ctx.saved_tensors + do = do.contiguous() + total_nnz, _three, _nheads, headdim = do.shape + # We need dqkv to be contiguous so that when we reshape to combine (3, nheads) dimensions, + # we get the same tensor + dqk = do[:, :2].view(total_nnz, -1, headdim) + apply_rotary( + dqk, + cos, + sin, + seqlen_offsets=0, + cu_seqlens=cu_seqlens, + max_seqlen=ctx.max_seqlen, + interleaved=False, + inplace=True, + conjugate=True, + ) + + return do, None, None, None, None, None, None + + +def apply_rotary_unpadded( + qkv, + cos, + sin, + cu_seqlens: Optional[torch.Tensor] = None, + max_seqlen: Optional[int] = None, +): + """ + Arguments: + qkv: (total_nnz, 3, nheads, headdim) - input tensor for packed QKV. + cos, sin: (seqlen_rotary, rotary_dim / 2) + interleaved: if True, rotate pairs of even and odd dimensions (GPT-J style) instead + of 1st half and 2nd half (GPT-NeoX style). + inplace: if True, apply rotary embedding in-place. + seqlen_offsets: (batch_size,) or int. Each sequence in x is shifted by this amount. + Most commonly used in inference when we have KV cache. + cu_seqlens: (batch + 1,) or None + max_seqlen: int + Return: + out: (total_nnz, dim) + rotary_dim must be <= headdim + Apply rotary embedding to the first rotary_dim of x. + """ + return ApplyRotaryEmbUnpad.apply(qkv, cos, sin, cu_seqlens, max_seqlen) + + +class ModernBertUnpaddedRotaryEmbedding(RotaryEmbedding): + """ + The rotary position embeddings applied directly to unpadded sequences. + """ + + def __init__( + self, + dim: int, + base: float = 10000.0, + max_seqlen: Optional[int] = None, + device: Optional[torch.device] = None, + dtype: Optional[torch.dtype] = None, + ): + """ + max_seqlen: if max_seqlen, device, and dtype are provided, we precompute the cos_sin_cache + up to max_seqlen. If the max_seqlen, device, or dtype during training/inference differ, + the cos_sin_cache wll be recomputed during the forward pass. + """ + super().__init__(dim=dim, base=base, pos_idx_in_fp32=True, device=device, interleaved=False) + self.max_seqlen = max_seqlen + + if max_seqlen is not None and device is not None and dtype is not None: + self._update_cos_sin_cache(max_seqlen, device=device, dtype=dtype) + + def forward( + self, + qkv: torch.Tensor, + cu_seqlens: torch.Tensor, + max_seqlen: Optional[int] = None, + ) -> Union[torch.Tensor, Tuple[torch.Tensor, torch.Tensor]]: + """ + Apply rotary embedding *inplace* to qkv. + qkv: (total_nnz, 3, nheads, headdim) + cu_seqlens: (batch + 1,) cumulative sequence lengths + max_seqlen: int max seq length in the batch + """ + if max_seqlen is not None: + self._update_cos_sin_cache(max_seqlen, device=qkv.device, dtype=qkv.dtype) + + qkv = apply_rotary_unpadded( + qkv, + self._cos_cached, + self._sin_cached, + cu_seqlens=cu_seqlens, + max_seqlen=max_seqlen, + ) + + return qkv + + def extra_repr(self) -> str: + return f"dim={self.dim}, base={self.base}, scale_base={self.scale_base}" + + +class ModernBertEmbeddings(nn.Module): + """ + Same as BertEmbeddings with a tiny tweak for positional embeddings indexing. + """ + + def __init__(self, config: ModernBertConfig): + super().__init__() + self.config = config + self.tok_embeddings = nn.Embedding(config.vocab_size, config.hidden_size, padding_idx=config.pad_token_id) + self.norm = nn.LayerNorm(config.hidden_size, eps=config.norm_eps, bias=config.norm_bias) + self.drop = nn.Dropout(config.embedding_dropout) + + @torch.compile(dynamic=True) + def compiled_embeddings(self, input_ids: torch.LongTensor) -> torch.Tensor: + return self.drop(self.norm(self.tok_embeddings(input_ids))) + + def forward( + self, input_ids: torch.LongTensor = None, inputs_embeds: Optional[torch.Tensor] = None + ) -> torch.Tensor: + if inputs_embeds is not None: + hidden_states = self.drop(self.norm(inputs_embeds)) + else: + hidden_states = ( + self.compiled_embeddings(input_ids) + if self.config.reference_compile + else self.drop(self.norm(self.tok_embeddings(input_ids))) + ) + return hidden_states + + +class ModernBertMLP(nn.Module): + """Applies the GLU at the end of each ModernBERT layer. + + Compared to the default BERT architecture, this block replaces :class:`~transformers.model.bert.modeling_bert.BertIntermediate` + and :class:`~transformers.model.bert.modeling_bert.SelfOutput` with a single module that has similar functionality. + """ + + def __init__(self, config: ModernBertConfig): + super().__init__() + self.config = config + self.Wi = nn.Linear(config.hidden_size, int(config.intermediate_size) * 2, bias=config.mlp_bias) + self.act = ACT2FN[config.hidden_activation] + self.drop = nn.Dropout(config.mlp_dropout) + self.Wo = nn.Linear(config.intermediate_size, config.hidden_size, bias=config.mlp_bias) + + def forward(self, hidden_states: torch.Tensor) -> torch.Tensor: + input, gate = self.Wi(hidden_states).chunk(2, dim=-1) + return self.Wo(self.drop(self.act(input) * gate)) + + +class ModernBertRotaryEmbedding(GemmaRotaryEmbedding): + def __init__(self, config: ModernBertConfig, dim: int, base: float, device: Optional[torch.device] = None): + super().__init__(self, config=config, device=device) + inv_freq, self.attention_scaling = self.rope_init_fn(None, device, dim=dim, base=base) + + +def eager_attention_forward( + module: "ModernBertAttention", + qkv: torch.Tensor, + attention_mask: torch.Tensor, + sliding_window_mask: torch.Tensor, + position_ids: Optional[torch.LongTensor], + local_attention: Tuple[int, int], + bs: int, + dim: int, + output_attentions: Optional[bool] = False, + **_kwargs, +) -> Union[Tuple[torch.Tensor, torch.Tensor], Tuple[torch.Tensor]]: + # qkv: [batch_size, seqlen, 3, nheads, headdim] + cos, sin = module.rotary_emb(qkv, position_ids=position_ids) + query, key, value = qkv.transpose(3, 1).unbind(dim=2) + # query, key, value: [batch_size, heads, seq_len, head_dim] + query, key = apply_rotary_pos_emb(query, key, cos, sin) + + scale = module.head_dim**-0.5 + attn_weights = torch.matmul(query, key.transpose(2, 3)) * scale + + if local_attention != (-1, -1): + attention_mask = sliding_window_mask + + attn_weights = attn_weights + attention_mask + + # upcast attention to fp32 + attn_weights = nn.functional.softmax(attn_weights, dim=-1, dtype=torch.float32).to(query.dtype) + attn_weights = nn.functional.dropout(attn_weights, p=module.attention_dropout, training=module.training) + attn_output = torch.matmul(attn_weights, value) + attn_output = attn_output.transpose(1, 2).contiguous() + attn_output = attn_output.view(bs, -1, dim) + if output_attentions: + return (attn_output, attn_weights) + return (attn_output,) + + +def flash_attention_forward( + module: "ModernBertAttention", + qkv: torch.Tensor, + rotary_emb: ModernBertUnpaddedRotaryEmbedding, + cu_seqlens: torch.Tensor, + max_seqlen: int, + local_attention: Tuple[int, int], + bs: int, + dim: int, + target_dtype: torch.dtype = torch.bfloat16, + **_kwargs, +) -> Tuple[torch.Tensor]: + # (total_seqlen, 3, nheads, headdim) + qkv = rotary_emb(qkv, cu_seqlens=cu_seqlens, max_seqlen=max_seqlen) + + convert_dtype = qkv.dtype not in (torch.float16, torch.bfloat16) + if convert_dtype: + # FA2 implementation only supports fp16 and bf16. If FA2 is supported, + # bfloat16 must be supported as of FA2 2.5.7. (Turing GPUs not supported) + orig_dtype = qkv.dtype + qkv = qkv.to(target_dtype) + + attn = flash_attn_varlen_qkvpacked_func( + qkv, + cu_seqlens=cu_seqlens, + max_seqlen=max_seqlen, + dropout_p=module.attention_dropout if module.training else 0.0, + deterministic=module.deterministic_flash_attn, + window_size=local_attention, + ) + attn = attn.to(orig_dtype) # type: ignore + else: + attn = flash_attn_varlen_qkvpacked_func( + qkv, + cu_seqlens=cu_seqlens, + max_seqlen=max_seqlen, + dropout_p=module.attention_dropout if module.training else 0.0, + deterministic=module.deterministic_flash_attn, + window_size=local_attention, + ) + return (attn.view(bs, dim),) + + +def sdpa_attention_forward( + module: "ModernBertAttention", + qkv: torch.Tensor, + attention_mask: torch.Tensor, + sliding_window_mask: torch.Tensor, + position_ids: Optional[torch.LongTensor], + local_attention: Tuple[int, int], + bs: int, + dim: int, + **_kwargs, +) -> Tuple[torch.Tensor]: + # qkv: [batch_size, seqlen, 3, nheads, headdim] + cos, sin = module.rotary_emb(qkv, position_ids=position_ids) + query, key, value = qkv.transpose(3, 1).unbind(dim=2) + # query, key, value: [batch_size, heads, seq_len, head_dim] + query, key = apply_rotary_pos_emb(query, key, cos, sin) + + if local_attention != (-1, -1): + attention_mask = sliding_window_mask + + attn_output = ( + F.scaled_dot_product_attention( + query, + key, + value, + dropout_p=module.attention_dropout if module.training else 0.0, + attn_mask=attention_mask, + ) + .transpose(1, 2) + .contiguous() + ) + attn_output = attn_output.view(bs, -1, dim) + return (attn_output,) + + +MODERNBERT_ATTENTION_FUNCTION = { + "flash_attention_2": flash_attention_forward, + "eager": eager_attention_forward, + "sdpa": sdpa_attention_forward, +} + + +class ModernBertAttention(nn.Module): + """Performs multi-headed self attention on a batch of unpadded sequences. + + If Flash Attention 2 is installed, this module uses Flash Attention to improve throughput. + If Flash Attention 2 is not installed, the implementation will use PyTorch's SDPA kernel, + which requires padding and unpadding inputs, adding some overhead. + + See `forward` method for additional details. + """ + + def __init__(self, config: ModernBertConfig, layer_id: Optional[int] = None): + super().__init__() + self.config = config + self.layer_id = layer_id + + if config.hidden_size % config.num_attention_heads != 0: + raise ValueError( + f"The hidden size ({config.hidden_size}) is not a multiple of the number of attention heads ({config.num_attention_heads})" + ) + + self.attention_dropout = config.attention_dropout + self.deterministic_flash_attn = config.deterministic_flash_attn + self.num_heads = config.num_attention_heads + self.head_dim = config.hidden_size // config.num_attention_heads + self.all_head_size = self.head_dim * self.num_heads + self.Wqkv = nn.Linear(config.hidden_size, 3 * self.all_head_size, bias=config.attention_bias) + + if layer_id % config.global_attn_every_n_layers != 0: + self.local_attention = (config.local_attention // 2, config.local_attention // 2) + else: + self.local_attention = (-1, -1) + + rope_theta = config.global_rope_theta + max_position_embeddings = config.max_position_embeddings + if self.local_attention != (-1, -1): + if config.local_rope_theta is not None: + rope_theta = config.local_rope_theta + max_position_embeddings = config.local_attention + + if config._attn_implementation == "flash_attention_2": + self.rotary_emb = ModernBertUnpaddedRotaryEmbedding( + dim=self.head_dim, max_seqlen=max_position_embeddings, base=rope_theta + ) + else: + self.rotary_emb = ModernBertRotaryEmbedding(config=config, dim=self.head_dim, base=rope_theta) + + self.Wo = nn.Linear(config.hidden_size, config.hidden_size, bias=config.attention_bias) + self.out_drop = nn.Dropout(config.attention_dropout) if config.attention_dropout > 0.0 else nn.Identity() + self.pruned_heads = set() + + def forward( + self, + hidden_states: torch.Tensor, + output_attentions: Optional[bool] = False, + **kwargs, + ) -> torch.Tensor: + qkv = self.Wqkv(hidden_states) + + bs = hidden_states.shape[0] + if self.config._attn_implementation == "flash_attention_2": + qkv = qkv.view(-1, 3, self.num_heads, self.head_dim) + else: + qkv = qkv.view(bs, -1, 3, self.num_heads, self.head_dim) + + attn_outputs = MODERNBERT_ATTENTION_FUNCTION[self.config._attn_implementation]( + self, + qkv=qkv, + rotary_emb=self.rotary_emb, + local_attention=self.local_attention, + bs=bs, + dim=self.all_head_size, + output_attentions=output_attentions, + **kwargs, + ) + hidden_states = attn_outputs[0] + hidden_states = self.out_drop(self.Wo(hidden_states)) + + return (hidden_states,) + attn_outputs[1:] # add attentions if outputted + + +class ModernBertEncoderLayer(nn.Module): + def __init__(self, config: ModernBertConfig, layer_id: Optional[int] = None): + super().__init__() + self.config = config + if layer_id == 0: + self.attn_norm = nn.Identity() + else: + self.attn_norm = nn.LayerNorm(config.hidden_size, eps=config.norm_eps, bias=config.norm_bias) + self.attn = ModernBertAttention(config=config, layer_id=layer_id) + self.mlp_norm = nn.LayerNorm(config.hidden_size, eps=config.norm_eps, bias=config.norm_bias) + self.mlp = ModernBertMLP(config) + + @torch.compile(dynamic=True) + def compiled_mlp(self, hidden_states: torch.Tensor) -> torch.Tensor: + return self.mlp(self.mlp_norm(hidden_states)) + + def forward( + self, + hidden_states: torch.Tensor, + attention_mask: Optional[torch.Tensor] = None, + sliding_window_mask: Optional[torch.Tensor] = None, + position_ids: Optional[torch.LongTensor] = None, + cu_seqlens: Optional[torch.Tensor] = None, + max_seqlen: Optional[int] = None, + output_attentions: Optional[bool] = False, + ) -> torch.Tensor: + attn_outputs = self.attn( + self.attn_norm(hidden_states), + attention_mask=attention_mask, + sliding_window_mask=sliding_window_mask, + position_ids=position_ids, + cu_seqlens=cu_seqlens, + max_seqlen=max_seqlen, + output_attentions=output_attentions, + ) + hidden_states = hidden_states + attn_outputs[0] + mlp_output = ( + self.compiled_mlp(hidden_states) + if self.config.reference_compile + else self.mlp(self.mlp_norm(hidden_states)) + ) + hidden_states = hidden_states + mlp_output + + return (hidden_states,) + attn_outputs[1:] # add attentions if outputted + + +MODERNBERT_START_DOCSTRING = r""" + This model inherits from [`PreTrainedModel`]. Check the superclass documentation for the generic methods the + library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads + etc.) + + This model is also a PyTorch [torch.nn.Module](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) subclass. + Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage + and behavior. + + Parameters: + config ([`ModernBertConfig`]): + Model configuration class with all the parameters of the model. Initializing with a config file does not + load the weights associated with the model, only the configuration. Check out the + [`~PreTrainedModel.from_pretrained`] method to load the model weights. +""" + + +@add_start_docstrings( + "The bare ModernBert Model outputting raw hidden-states without any specific head on top.", + MODERNBERT_START_DOCSTRING, +) +class ModernBertPreTrainedModel(PreTrainedModel): + config_class = ModernBertConfig + base_model_prefix = "model" + supports_gradient_checkpointing = True + _no_split_modules = ["ModernBertEmbeddings", "ModernBertEncoderLayer"] + _supports_flash_attn_2 = True + _supports_sdpa = True + _supports_flex_attn = False + + def _init_weights(self, module: nn.Module): + cutoff_factor = self.config.initializer_cutoff_factor + if cutoff_factor is None: + cutoff_factor = 3 + + def init_weight(module: nn.Module, std: float): + nn.init.trunc_normal_( + module.weight, + mean=0.0, + std=std, + a=-cutoff_factor * std, + b=cutoff_factor * std, + ) + + if isinstance(module, nn.Linear): + if module.bias is not None: + nn.init.zeros_(module.bias) + + stds = { + "in": self.config.initializer_range, + "out": self.config.initializer_range / math.sqrt(2.0 * self.config.num_hidden_layers), + "embedding": self.config.initializer_range, + "final_out": self.config.hidden_size**-0.5, + } + + if isinstance(module, ModernBertEmbeddings): + init_weight(module.tok_embeddings, stds["embedding"]) + elif isinstance(module, ModernBertMLP): + init_weight(module.Wi, stds["in"]) + init_weight(module.Wo, stds["out"]) + elif isinstance(module, ModernBertAttention): + init_weight(module.Wqkv, stds["in"]) + init_weight(module.Wo, stds["out"]) + elif isinstance(module, ModernBertPredictionHead): + init_weight(module.dense, stds["out"]) + elif isinstance(module, ModernBertForMaskedLM): + init_weight(module.decoder, stds["out"]) + elif isinstance(module, (ModernBertForSequenceClassification, ModernBertForTokenClassification)): + init_weight(module.classifier, stds["final_out"]) + + @classmethod + def _autoset_attn_implementation( + cls, + config, + use_flash_attention_2: bool = False, + torch_dtype: Optional[torch.dtype] = None, + device_map: Optional[Union[str, Dict[str, int]]] = None, + check_device_map: bool = True, + ): + # If the user didn't specify anything, try to use flash_attention_2 if available. + # Otherwise we fall back to the default SDPA -> Eager from the super() method. + # ModernBert's FA2 implementation correctly handles non-fp16/bf16 dtypes, we don't + # need the FA2 warning for non-fp16/bf16 dtypes so we set fp16 for the FA2 check. + if config._attn_implementation_internal is None: + config._attn_implementation_internal = "flash_attention_2" + try: + return cls._check_and_enable_flash_attn_2( + config, + torch_dtype=torch.float16, + device_map=device_map, + hard_check_only=False, + check_device_map=check_device_map, + ) + except (ValueError, ImportError): + config._attn_implementation_internal = None + return super()._autoset_attn_implementation( + config, + use_flash_attention_2=use_flash_attention_2, + torch_dtype=torch.float16, + device_map=device_map, + check_device_map=check_device_map, + ) + + def _maybe_set_compile(self): + if self.config.reference_compile is False: + return + + if hasattr(self, "hf_device_map") and len(self.hf_device_map) > 1: + if self.config.reference_compile: + logger.warning_once( + "If `accelerate` split the model across devices, `torch.compile` will not work. " + "Falling back to non-compiled mode." + ) + self.config.reference_compile = False + + if self.device.type == "mps": + if self.config.reference_compile: + logger.warning_once( + "Compiling the model with `torch.compile` and using a `torch.mps` device is not supported. " + "Falling back to non-compiled mode." + ) + self.config.reference_compile = False + + if self.device.type == "cpu": + if self.config.reference_compile: + logger.warning_once( + "Compiling the model with `torch.compile` and using a `torch.cpu` device is not supported. " + "Falling back to non-compiled mode." + ) + self.config.reference_compile = False + + if self.config.reference_compile is None: + self.config.reference_compile = is_triton_available() + + def resize_token_embeddings(self, *args, **kwargs): + model_embeds = super().resize_token_embeddings(*args, **kwargs) + + if self.config.reference_compile in {True, None}: + if self.config.reference_compile: + logger.warning_once( + "Resizing token embeddings with `torch.compile` is not supported. Falling back to non-compiled mode." + ) + self.config.reference_compile = False + + return model_embeds + + +MODERNBERT_INPUTS_DOCSTRING = r""" + Args: + input_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`): + Indices of input sequence tokens in the vocabulary. With Flash Attention 2.0, padding will be ignored + by default should you provide it. + + Indices can be obtained using [`AutoTokenizer`]. See [`PreTrainedTokenizer.encode`] and + [`PreTrainedTokenizer.__call__`] for details. + + [What are input IDs?](../glossary#input-ids) + attention_mask (`torch.Tensor` of shape `(batch_size, sequence_length)`, *optional*): + Mask to avoid performing attention on padding token indices. Mask values selected in `[0, 1]`: + + - 1 for tokens that are **not masked**, + - 0 for tokens that are **masked**. + + [What are attention masks?](../glossary#attention-mask) + + Indices can be obtained using [`AutoTokenizer`]. See [`PreTrainedTokenizer.encode`] and + [`PreTrainedTokenizer.__call__`] for details. + + If you want to change padding behavior, you should read [`modeling_opt._prepare_decoder_attention_mask`] + and modify to your needs. See diagram 1 in [the paper](https://arxiv.org/abs/1910.13461) for more + information on the default strategy. + + - 1 indicates the head is **not masked**, + - 0 indicates the head is **masked**. + sliding_window_mask (`torch.Tensor` of shape `(batch_size, sequence_length)`, *optional*): + Mask to avoid performing attention on padding or far-away tokens. In ModernBert, only every few layers + perform global attention, while the rest perform local attention. This mask is used to avoid attending to + far-away tokens in the local attention layers when not using Flash Attention. + position_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*): + Indices of positions of each input sequence tokens in the position embeddings. Selected in the range `[0, + config.n_positions - 1]`. + + [What are position IDs?](../glossary#position-ids) + inputs_embeds (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`, *optional*): + Optionally, instead of passing `input_ids` you can choose to directly pass an embedded representation. This + is useful if you want more control over how to convert `input_ids` indices into associated vectors than the + model's internal embedding lookup matrix. + indices (`torch.Tensor` of shape `(total_unpadded_tokens,)`, *optional*): + Indices of the non-padding tokens in the input sequence. Used for unpadding the output. + cu_seqlens (`torch.Tensor` of shape `(batch + 1,)`, *optional*): + Cumulative sequence lengths of the input sequences. Used to index the unpadded tensors. + max_seqlen (`int`, *optional*): + Maximum sequence length in the batch excluding padding tokens. Used to unpad input_ids and pad output tensors. + batch_size (`int`, *optional*): + Batch size of the input sequences. Used to pad the output tensors. + seq_len (`int`, *optional*): + Sequence length of the input sequences including padding tokens. Used to pad the output tensors. + output_attentions (`bool`, *optional*): + Whether or not to return the attentions tensors of all attention layers. See `attentions` under returned + tensors for more detail. + output_hidden_states (`bool`, *optional*): + Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for + more detail. + return_dict (`bool`, *optional*): + Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple. +""" + + +@add_start_docstrings( + "The bare ModernBert Model outputting raw hidden-states without any specific head on top.", + MODERNBERT_START_DOCSTRING, +) +class ModernBertModel(ModernBertPreTrainedModel): + def __init__(self, config: ModernBertConfig): + super().__init__(config) + self.config = config + self.embeddings = ModernBertEmbeddings(config) + self.layers = nn.ModuleList( + [ModernBertEncoderLayer(config, layer_id) for layer_id in range(config.num_hidden_layers)] + ) + self.final_norm = nn.LayerNorm(config.hidden_size, eps=config.norm_eps, bias=config.norm_bias) + self.gradient_checkpointing = False + self.post_init() + + def get_input_embeddings(self): + return self.embeddings.tok_embeddings + + def set_input_embeddings(self, value): + self.embeddings.tok_embeddings = value + + @add_start_docstrings_to_model_forward(MODERNBERT_INPUTS_DOCSTRING) + @add_code_sample_docstrings( + checkpoint=_CHECKPOINT_FOR_DOC, + output_type=BaseModelOutput, + config_class=_CONFIG_FOR_DOC, + ) + def forward( + self, + input_ids: Optional[torch.LongTensor] = None, + attention_mask: Optional[torch.Tensor] = None, + sliding_window_mask: Optional[torch.Tensor] = None, + position_ids: Optional[torch.LongTensor] = None, + inputs_embeds: Optional[torch.Tensor] = None, + indices: Optional[torch.Tensor] = None, + cu_seqlens: Optional[torch.Tensor] = None, + max_seqlen: Optional[int] = None, + batch_size: Optional[int] = None, + seq_len: Optional[int] = None, + output_attentions: Optional[bool] = None, + output_hidden_states: Optional[bool] = None, + return_dict: Optional[bool] = None, + ) -> Union[Tuple[torch.Tensor, ...], BaseModelOutput]: + output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions + output_hidden_states = ( + output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states + ) + return_dict = return_dict if return_dict is not None else self.config.use_return_dict + + if (input_ids is None) ^ (inputs_embeds is not None): + raise ValueError("You must specify exactly one of input_ids or inputs_embeds") + + all_hidden_states = () if output_hidden_states else None + all_self_attentions = () if output_attentions else None + + self._maybe_set_compile() + + if input_ids is not None: + self.warn_if_padding_and_no_attention_mask(input_ids, attention_mask) + + if batch_size is None and seq_len is None: + if inputs_embeds is not None: + batch_size, seq_len = inputs_embeds.shape[:2] + else: + batch_size, seq_len = input_ids.shape[:2] + device = input_ids.device if input_ids is not None else inputs_embeds.device + + if attention_mask is None: + attention_mask = torch.ones((batch_size, seq_len), device=device, dtype=torch.bool) + + repad = False + if self.config._attn_implementation == "flash_attention_2": + if indices is None and cu_seqlens is None and max_seqlen is None: + repad = True + if inputs_embeds is None: + with torch.no_grad(): + input_ids, indices, cu_seqlens, max_seqlen, *_ = _unpad_modernbert_input( + inputs=input_ids, attention_mask=attention_mask + ) + else: + inputs_embeds, indices, cu_seqlens, max_seqlen, *_ = _unpad_modernbert_input( + inputs=inputs_embeds, attention_mask=attention_mask + ) + else: + if position_ids is None: + position_ids = torch.arange(seq_len, device=device).unsqueeze(0) + + attention_mask, sliding_window_mask = self._update_attention_mask( + attention_mask, output_attentions=output_attentions + ) + + hidden_states = self.embeddings(input_ids=input_ids, inputs_embeds=inputs_embeds) + + for encoder_layer in self.layers: + if output_hidden_states: + all_hidden_states = all_hidden_states + (hidden_states,) + + if self.gradient_checkpointing and self.training: + layer_outputs = self._gradient_checkpointing_func( + encoder_layer.__call__, + hidden_states, + attention_mask, + sliding_window_mask, + position_ids, + cu_seqlens, + max_seqlen, + output_attentions, + ) + else: + layer_outputs = encoder_layer( + hidden_states, + attention_mask=attention_mask, + sliding_window_mask=sliding_window_mask, + position_ids=position_ids, + cu_seqlens=cu_seqlens, + max_seqlen=max_seqlen, + output_attentions=output_attentions, + ) + hidden_states = layer_outputs[0] + if output_attentions and len(layer_outputs) > 1: + all_self_attentions = all_self_attentions + (layer_outputs[1],) + + if output_hidden_states: + all_hidden_states = all_hidden_states + (hidden_states,) + + hidden_states = self.final_norm(hidden_states) + + if repad: + hidden_states = _pad_modernbert_output( + inputs=hidden_states, indices=indices, batch=batch_size, seqlen=seq_len + ) + if all_hidden_states is not None: + all_hidden_states = tuple( + _pad_modernbert_output(inputs=hs, indices=indices, batch=batch_size, seqlen=seq_len) + for hs in all_hidden_states + ) + + if not return_dict: + return tuple(v for v in [hidden_states, all_hidden_states, all_self_attentions] if v is not None) + return BaseModelOutput( + last_hidden_state=hidden_states, + hidden_states=all_hidden_states, + attentions=all_self_attentions, + ) + + def _update_attention_mask(self, attention_mask: torch.Tensor, output_attentions: bool) -> torch.Tensor: + if output_attentions: + if self.config._attn_implementation == "sdpa": + logger.warning_once( + "Outputting attentions is only supported with the 'eager' attention implementation, " + 'not with "sdpa". Falling back to `attn_implementation="eager"`.' + ) + self.config._attn_implementation = "eager" + elif self.config._attn_implementation != "eager": + logger.warning_once( + "Outputting attentions is only supported with the eager attention implementation, " + f'not with {self.config._attn_implementation}. Consider setting `attn_implementation="eager"`.' + " Setting `output_attentions=False`." + ) + + global_attention_mask = _prepare_4d_attention_mask(attention_mask, self.dtype) + + # Create position indices + rows = torch.arange(global_attention_mask.shape[2]).unsqueeze(0) + # Calculate distance between positions + distance = torch.abs(rows - rows.T) + + # Create sliding window mask (1 for positions within window, 0 outside) + window_mask = ( + (distance <= self.config.local_attention // 2).unsqueeze(0).unsqueeze(0).to(attention_mask.device) + ) + # Combine with existing mask + sliding_window_mask = global_attention_mask.masked_fill(window_mask.logical_not(), torch.finfo(self.dtype).min) + + return global_attention_mask, sliding_window_mask + + +class ModernBertPredictionHead(nn.Module): + def __init__(self, config: ModernBertConfig): + super().__init__() + self.config = config + self.dense = nn.Linear(config.hidden_size, config.hidden_size, config.classifier_bias) + self.act = ACT2FN[config.classifier_activation] + self.norm = nn.LayerNorm(config.hidden_size, eps=config.norm_eps, bias=config.norm_bias) + + def forward(self, hidden_states: torch.Tensor) -> torch.Tensor: + return self.norm(self.act(self.dense(hidden_states))) + + +@add_start_docstrings( + "The ModernBert Model with a decoder head on top that is used for masked language modeling.", + MODERNBERT_START_DOCSTRING, +) +class ModernBertForMaskedLM(ModernBertPreTrainedModel): + _tied_weights_keys = ["decoder.weight"] + + def __init__(self, config: ModernBertConfig): + super().__init__(config) + self.config = config + self.model = ModernBertModel(config) + self.head = ModernBertPredictionHead(config) + self.decoder = nn.Linear(config.hidden_size, config.vocab_size, bias=config.decoder_bias) + + self.sparse_prediction = self.config.sparse_prediction + self.sparse_pred_ignore_index = self.config.sparse_pred_ignore_index + + # Initialize weights and apply final processing + self.post_init() + + def get_output_embeddings(self): + return self.decoder + + def set_output_embeddings(self, new_embeddings: nn.Linear): + self.decoder = new_embeddings + + @torch.compile(dynamic=True) + def compiled_head(self, output: torch.Tensor) -> torch.Tensor: + return self.decoder(self.head(output)) + + @add_start_docstrings_to_model_forward(MODERNBERT_INPUTS_DOCSTRING) + @add_code_sample_docstrings( + checkpoint=_CHECKPOINT_FOR_DOC, + output_type=MaskedLMOutput, + config_class=_CONFIG_FOR_DOC, + ) + def forward( + self, + input_ids: Optional[torch.LongTensor] = None, + attention_mask: Optional[torch.Tensor] = None, + sliding_window_mask: Optional[torch.Tensor] = None, + position_ids: Optional[torch.Tensor] = None, + inputs_embeds: Optional[torch.Tensor] = None, + labels: Optional[torch.Tensor] = None, + indices: Optional[torch.Tensor] = None, + cu_seqlens: Optional[torch.Tensor] = None, + max_seqlen: Optional[int] = None, + batch_size: Optional[int] = None, + seq_len: Optional[int] = None, + output_attentions: Optional[bool] = None, + output_hidden_states: Optional[bool] = None, + return_dict: Optional[bool] = None, + **kwargs, + ) -> Union[Tuple[torch.Tensor], MaskedLMOutput]: + return_dict = return_dict if return_dict is not None else self.config.use_return_dict + self._maybe_set_compile() + + if self.config._attn_implementation == "flash_attention_2": + if indices is None and cu_seqlens is None and max_seqlen is None: + if batch_size is None and seq_len is None: + if inputs_embeds is not None: + batch_size, seq_len = inputs_embeds.shape[:2] + else: + batch_size, seq_len = input_ids.shape[:2] + device = input_ids.device if input_ids is not None else inputs_embeds.device + + if attention_mask is None: + attention_mask = torch.ones((batch_size, seq_len), device=device, dtype=torch.bool) + + if inputs_embeds is None: + with torch.no_grad(): + input_ids, indices, cu_seqlens, max_seqlen, position_ids, labels = _unpad_modernbert_input( + inputs=input_ids, attention_mask=attention_mask, position_ids=position_ids, labels=labels + ) + else: + inputs_embeds, indices, cu_seqlens, max_seqlen, position_ids, labels = _unpad_modernbert_input( + inputs=inputs_embeds, attention_mask=attention_mask, position_ids=position_ids, labels=labels + ) + + outputs = self.model( + input_ids=input_ids, + attention_mask=attention_mask, + sliding_window_mask=sliding_window_mask, + position_ids=position_ids, + inputs_embeds=inputs_embeds, + indices=indices, + cu_seqlens=cu_seqlens, + max_seqlen=max_seqlen, + batch_size=batch_size, + seq_len=seq_len, + output_attentions=output_attentions, + output_hidden_states=output_hidden_states, + return_dict=return_dict, + ) + last_hidden_state = outputs[0] + + if self.sparse_prediction and labels is not None: + # flatten labels and output first + labels = labels.view(-1) + last_hidden_state = last_hidden_state.view(labels.shape[0], -1) + + # then filter out the non-masked tokens + mask_tokens = labels != self.sparse_pred_ignore_index + last_hidden_state = last_hidden_state[mask_tokens] + labels = labels[mask_tokens] + + logits = ( + self.compiled_head(last_hidden_state) + if self.config.reference_compile + else self.decoder(self.head(last_hidden_state)) + ) + + loss = None + if labels is not None: + loss = self.loss_function(logits, labels, vocab_size=self.config.vocab_size) + + if self.config._attn_implementation == "flash_attention_2": + with nullcontext() if self.config.repad_logits_with_grad or labels is None else torch.no_grad(): + logits = _pad_modernbert_output(inputs=logits, indices=indices, batch=batch_size, seqlen=seq_len) + + if not return_dict: + output = (logits,) + return ((loss,) + output) if loss is not None else output + + return MaskedLMOutput( + loss=loss, + logits=logits, + hidden_states=outputs.hidden_states, + attentions=outputs.attentions, + ) + + +@add_start_docstrings( + "The ModernBert Model with a sequence classification head on top that performs pooling.", + MODERNBERT_START_DOCSTRING, +) +class ModernBertForSequenceClassification(ModernBertPreTrainedModel): + def __init__(self, config: ModernBertConfig): + super().__init__(config) + self.num_labels = config.num_labels + self.config = config + + self.model = ModernBertModel(config) + self.head = ModernBertPredictionHead(config) + self.drop = torch.nn.Dropout(config.classifier_dropout) + self.classifier = nn.Linear(config.hidden_size, config.num_labels) + + # Initialize weights and apply final processing + self.post_init() + + @add_start_docstrings_to_model_forward(MODERNBERT_INPUTS_DOCSTRING) + @add_code_sample_docstrings( + checkpoint=_CHECKPOINT_FOR_DOC, + output_type=SequenceClassifierOutput, + config_class=_CONFIG_FOR_DOC, + ) + def forward( + self, + input_ids: Optional[torch.LongTensor] = None, + attention_mask: Optional[torch.Tensor] = None, + sliding_window_mask: Optional[torch.Tensor] = None, + position_ids: Optional[torch.Tensor] = None, + inputs_embeds: Optional[torch.Tensor] = None, + labels: Optional[torch.Tensor] = None, + indices: Optional[torch.Tensor] = None, + cu_seqlens: Optional[torch.Tensor] = None, + max_seqlen: Optional[int] = None, + batch_size: Optional[int] = None, + seq_len: Optional[int] = None, + output_attentions: Optional[bool] = None, + output_hidden_states: Optional[bool] = None, + return_dict: Optional[bool] = None, + **kwargs, + ) -> Union[Tuple[torch.Tensor], SequenceClassifierOutput]: + r""" + labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*): + Labels for computing the sequence classification/regression loss. Indices should be in `[0, ..., + config.num_labels - 1]`. If `config.num_labels == 1` a regression loss is computed (Mean-Square loss), If + `config.num_labels > 1` a classification loss is computed (Cross-Entropy). + """ + return_dict = return_dict if return_dict is not None else self.config.use_return_dict + self._maybe_set_compile() + + outputs = self.model( + input_ids=input_ids, + attention_mask=attention_mask, + sliding_window_mask=sliding_window_mask, + position_ids=position_ids, + inputs_embeds=inputs_embeds, + indices=indices, + cu_seqlens=cu_seqlens, + max_seqlen=max_seqlen, + batch_size=batch_size, + seq_len=seq_len, + output_attentions=output_attentions, + output_hidden_states=output_hidden_states, + return_dict=return_dict, + ) + last_hidden_state = outputs[0] + + if self.config.classifier_pooling == "cls": + last_hidden_state = last_hidden_state[:, 0] + elif self.config.classifier_pooling == "mean": + last_hidden_state = (last_hidden_state * attention_mask.unsqueeze(-1)).sum(dim=1) / attention_mask.sum( + dim=1, keepdim=True + ) + + pooled_output = self.head(last_hidden_state) + pooled_output = self.drop(pooled_output) + logits = self.classifier(pooled_output) + + loss = None + if labels is not None: + if self.config.problem_type is None: + if self.num_labels == 1: + self.config.problem_type = "regression" + elif self.num_labels > 1 and (labels.dtype == torch.long or labels.dtype == torch.int): + self.config.problem_type = "single_label_classification" + else: + self.config.problem_type = "multi_label_classification" + + if self.config.problem_type == "regression": + loss_fct = MSELoss() + if self.num_labels == 1: + loss = loss_fct(logits.squeeze(), labels.squeeze()) + else: + loss = loss_fct(logits, labels) + elif self.config.problem_type == "single_label_classification": + loss_fct = CrossEntropyLoss() + loss = loss_fct(logits.view(-1, self.num_labels), labels.view(-1)) + elif self.config.problem_type == "multi_label_classification": + loss_fct = BCEWithLogitsLoss() + loss = loss_fct(logits, labels) + + if not return_dict: + output = (logits,) + return ((loss,) + output) if loss is not None else output + + return SequenceClassifierOutput( + loss=loss, + logits=logits, + hidden_states=outputs.hidden_states, + attentions=outputs.attentions, + ) + + +@add_start_docstrings( + "The ModernBert Model with a token classification head on top, e.g. for Named Entity Recognition (NER) tasks.", + MODERNBERT_START_DOCSTRING, +) +class ModernBertForTokenClassification(ModernBertPreTrainedModel): + def __init__(self, config: ModernBertConfig): + super().__init__(config) + self.num_labels = config.num_labels + + self.model = ModernBertModel(config) + self.head = ModernBertPredictionHead(config) + self.drop = torch.nn.Dropout(config.classifier_dropout) + self.classifier = nn.Linear(config.hidden_size, config.num_labels) + + # Initialize weights and apply final processing + self.post_init() + + @add_start_docstrings_to_model_forward(MODERNBERT_INPUTS_DOCSTRING) + @add_code_sample_docstrings( + checkpoint=_CHECKPOINT_FOR_DOC, + output_type=TokenClassifierOutput, + config_class=_CONFIG_FOR_DOC, + ) + def forward( + self, + input_ids: Optional[torch.LongTensor] = None, + attention_mask: Optional[torch.Tensor] = None, + sliding_window_mask: Optional[torch.Tensor] = None, + position_ids: Optional[torch.Tensor] = None, + inputs_embeds: Optional[torch.Tensor] = None, + labels: Optional[torch.Tensor] = None, + indices: Optional[torch.Tensor] = None, + cu_seqlens: Optional[torch.Tensor] = None, + max_seqlen: Optional[int] = None, + batch_size: Optional[int] = None, + seq_len: Optional[int] = None, + output_attentions: Optional[bool] = None, + output_hidden_states: Optional[bool] = None, + return_dict: Optional[bool] = None, + ) -> Union[Tuple[torch.Tensor], TokenClassifierOutput]: + r""" + labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*): + Labels for computing the token classification loss. Indices should be in `[0, ..., config.num_labels - 1]`. + """ + return_dict = return_dict if return_dict is not None else self.config.use_return_dict + self._maybe_set_compile() + + outputs = self.model( + input_ids=input_ids, + attention_mask=attention_mask, + sliding_window_mask=sliding_window_mask, + position_ids=position_ids, + inputs_embeds=inputs_embeds, + indices=indices, + cu_seqlens=cu_seqlens, + max_seqlen=max_seqlen, + batch_size=batch_size, + seq_len=seq_len, + output_attentions=output_attentions, + output_hidden_states=output_hidden_states, + return_dict=return_dict, + ) + last_hidden_state = outputs[0] + + last_hidden_state = self.head(last_hidden_state) + last_hidden_state = self.drop(last_hidden_state) + logits = self.classifier(last_hidden_state) + + loss = None + if labels is not None: + loss_fct = CrossEntropyLoss() + loss = loss_fct(logits.view(-1, self.num_labels), labels.view(-1)) + + if not return_dict: + output = (logits,) + outputs[1:] + return ((loss,) + output) if loss is not None else output + + return TokenClassifierOutput( + loss=loss, + logits=logits, + hidden_states=outputs.hidden_states, + attentions=outputs.attentions, + ) + + +__all__ = [ + "ModernBertConfig", + "ModernBertModel", + "ModernBertPreTrainedModel", + "ModernBertForMaskedLM", + "ModernBertForSequenceClassification", + "ModernBertForTokenClassification", +] diff --git a/vllm/lib/python3.10/site-packages/transformers/models/xglm/__init__.py b/vllm/lib/python3.10/site-packages/transformers/models/xglm/__init__.py new file mode 100644 index 0000000000000000000000000000000000000000..1eefd79d4cf7c3e9a24aa2f06d40f354a8fc28b2 --- /dev/null +++ b/vllm/lib/python3.10/site-packages/transformers/models/xglm/__init__.py @@ -0,0 +1,31 @@ +# Copyright 2024 The HuggingFace Team. All rights reserved. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +from typing import TYPE_CHECKING + +from ...utils import _LazyModule +from ...utils.import_utils import define_import_structure + + +if TYPE_CHECKING: + from .configuration_xglm import * + from .modeling_flax_xglm import * + from .modeling_tf_xglm import * + from .modeling_xglm import * + from .tokenization_xglm import * + from .tokenization_xglm_fast import * +else: + import sys + + _file = globals()["__file__"] + sys.modules[__name__] = _LazyModule(__name__, _file, define_import_structure(_file), module_spec=__spec__) diff --git a/vllm/lib/python3.10/site-packages/transformers/models/xglm/configuration_xglm.py b/vllm/lib/python3.10/site-packages/transformers/models/xglm/configuration_xglm.py new file mode 100644 index 0000000000000000000000000000000000000000..da5ded49162306551984b7ae3d7cc62319970e16 --- /dev/null +++ b/vllm/lib/python3.10/site-packages/transformers/models/xglm/configuration_xglm.py @@ -0,0 +1,139 @@ +# coding=utf-8 +# Copyright The HuggingFace Inc. team. All rights reserved. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +"""XGLM model configuration""" + +from ...configuration_utils import PretrainedConfig +from ...utils import logging + + +logger = logging.get_logger(__name__) + + +class XGLMConfig(PretrainedConfig): + r""" + This is the configuration class to store the configuration of a [`XGLMModel`]. It is used to instantiate an XGLM + model according to the specified arguments, defining the model architecture. Instantiating a configuration with the + defaults will yield a similar configuration to that of the XGLM + [facebook/xglm-564M](https://huggingface.co/facebook/xglm-564M) architecture. + + Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the + documentation from [`PretrainedConfig`] for more information. + + + Args: + vocab_size (`int`, *optional*, defaults to 256008): + Vocabulary size of the XGLM model. Defines the number of different tokens that can be represented by the + `inputs_ids` passed when calling [`XGLMModel`] or [`FlaxXGLMModel`]. + max_position_embeddings (`int`, *optional*, defaults to 2048): + The maximum sequence length that this model might ever be used with. Typically set this to something large + just in case (e.g., 512 or 1024 or 2048). + d_model (`int`, *optional*, defaults to 1024): + Dimension of the layers and the pooler layer. + ffn_dim (`int`, *optional*, defaults to 4096): + Dimension of the "intermediate" (often named feed-forward) layer in decoder. + num_layers (`int`, *optional*, defaults to 24): + Number of hidden layers Transformer decoder. + attention_heads (`int`, *optional*, defaults to 16): + Number of attention heads for each attention layer in the Transformer decoder. + activation_function (`str` or `function`, *optional*, defaults to `"gelu"`): + The non-linear activation function (function or string) in the encoder and pooler. If string, `"gelu"`, + `"relu"`, `"silu"` and `"gelu_new"` are supported. + dropout (`float`, *optional*, defaults to 0.1): + The dropout probability for all fully connected layers in the embeddings, dencoder, and pooler. + attention_dropout (`float`, *optional*, defaults to 0.1): + The dropout ratio for the attention probabilities. + activation_dropout (`float`, *optional*, defaults to 0.0): + The dropout ratio for activations inside the fully connected layer. + layerdrop (`float`, *optional*, defaults to 0.0): + The LayerDrop probability for the encoder. See the [LayerDrop paper](see https://arxiv.org/abs/1909.11556) + for more details. + init_std (`float`, *optional*, defaults to 0.02): + The standard deviation of the truncated_normal_initializer for initializing all weight matrices. + scale_embedding (`bool`, *optional*, defaults to `True`): + Scale embeddings by diving by sqrt(d_model). + use_cache (`bool`, *optional*, defaults to `True`): + Whether or not the model should return the last key/values attentions (not used by all models). + + Example: + + ```python + >>> from transformers import XGLMModel, XGLMConfig + + >>> # Initializing a XGLM facebook/xglm-564M style configuration + >>> configuration = XGLMConfig() + + >>> # Initializing a model from the facebook/xglm-564M style configuration + >>> model = XGLMModel(configuration) + + >>> # Accessing the model configuration + >>> configuration = model.config + ```""" + + model_type = "xglm" + keys_to_ignore_at_inference = ["past_key_values"] + + attribute_map = { + "num_attention_heads": "attention_heads", + "hidden_size": "d_model", + "num_hidden_layers": "num_layers", + } + + def __init__( + self, + vocab_size=256008, + max_position_embeddings=2048, + d_model=1024, + ffn_dim=4096, + num_layers=24, + attention_heads=16, + activation_function="gelu", + dropout=0.1, + attention_dropout=0.1, + activation_dropout=0.0, + layerdrop=0.0, + init_std=0.02, + scale_embedding=True, + use_cache=True, + decoder_start_token_id=2, + pad_token_id=1, + bos_token_id=0, + eos_token_id=2, + **kwargs, + ): + self.vocab_size = vocab_size + self.max_position_embeddings = max_position_embeddings + self.d_model = d_model + self.ffn_dim = ffn_dim + self.num_layers = num_layers + self.attention_heads = attention_heads + self.activation_function = activation_function + self.dropout = dropout + self.attention_dropout = attention_dropout + self.activation_dropout = activation_dropout + self.layerdrop = layerdrop + self.init_std = init_std + self.scale_embedding = scale_embedding # scale factor will be sqrt(d_model) if True + self.use_cache = use_cache + + super().__init__( + pad_token_id=pad_token_id, + bos_token_id=bos_token_id, + eos_token_id=eos_token_id, + decoder_start_token_id=decoder_start_token_id, + **kwargs, + ) + + +__all__ = ["XGLMConfig"]