| /aosp_15_r20/external/apache-commons-math/src/main/java/org/apache/commons/math3/analysis/differentiation/ |
| H A D | SparseGradient.java | 54 private final Map<Integer, Double> derivatives; field in SparseGradient
58 * @param derivatives derivatives map, a deep copy will be performed,
60 * may be null to create an empty derivatives map, i.e. a constant value
62 private SparseGradient(final double value, final Map<Integer, Double> derivatives) { in SparseGradient() argument
64 this.derivatives = new HashMap<Integer, Double>(); in SparseGradient()
65 if (derivatives != null) { in SparseGradient()
66 this.derivatives.putAll(derivatives); in SparseGradient()
72 * @param scale scaling factor to apply to all derivatives
73 * @param derivatives derivatives map, a deep copy will be performed,
75 * may be null to create an empty derivatives map, i.e. a constant value
[all …]
|
| H A D | DerivativeStructure.java | 39 * value of a function. Dan Kalman's derivative structures hold all partial derivatives
49 * on the derivation order despite no requiring users to compute the derivatives by
72 /** Build an instance with all values and derivatives set to 0.
80 /** Build an instance with all values and derivatives set to 0.
193 /** Build an instance from all its derivatives.
196 * @param derivatives derivatives sorted according to
198 * @exception DimensionMismatchException if derivatives array does not match the
203 public DerivativeStructure(final int parameters, final int order, final double ... derivatives) in DerivativeStructure() argument
206 if (derivatives.length != data.length) { in DerivativeStructure()
207 throw new DimensionMismatchException(derivatives.length, data.length); in DerivativeStructure()
[all …]
|
| H A D | DSCompiler.java | 42 * value and partial derivatives. The class does not hold these arrays, which remains under
46 * thousands of different parameters kept together with all theur partial derivatives.
49 * The arrays on which compilers operate contain only the partial derivatives together
50 * with the 0<sup>th</sup> derivative, i.e. the value. The partial derivatives are stored in
87 * // we don't know where first derivatives are stored, so we ask the compiler
92 * // we let all higher order derivatives set to 0
109 * // first derivatives
114 * // cross derivatives (assuming order was at least 2)
137 /** Number of partial derivatives (including the single 0 order derivative element). */
140 /** Indirection array for partial derivatives. */
[all …]
|
| H A D | FiniteDifferencesDifferentiator.java | 36 * wrapper objects compute derivatives in addition to function
43 * theoretically able to compute derivatives up to order n-1, but
56 * compute accurate derivatives for any order. However, due to numerical
118 * these cases. At an extreme case, computing derivatives exactly at the lower bound
182 * Evaluate derivatives from a sample.
186 * @param t evaluation abscissa value and derivatives
189 * @return value and derivatives at {@code t}
217 final double[] derivatives = t.getAllDerivatives(); in evaluate() local
227 derivatives[0] = dt0 - (i - 1) * stepSize; in evaluate()
228 … final DerivativeStructure deltaX = new DerivativeStructure(parameters, order, derivatives); in evaluate()
[all …]
|
| /aosp_15_r20/external/eigen/unsupported/Eigen/src/AutoDiff/ |
| H A D | AutoDiffScalar.h | 43 …* \param DerivativeType the vector type used to store/represent the derivatives. The base scalar t…
44 * as well as the number of derivatives to compute are determined from this type.
45 * Typical choices include, e.g., \c Vector4f for 4 derivatives, or \c VectorXf
46 * if the number of derivatives is not known at compile time, and/or, the number
47 * of derivatives is large.
52 …* This class represents a scalar value while tracking its respective derivatives using Eigen's exp…
62 * while derivatives are computed right away.
87 …and initializes the \a nbDer derivatives such that it corresponds to the \a derNumber -th variable…
95 * The derivatives are set to zero. */
103 /** Constructs an active scalar from its \a value and derivatives \a der */
[all …]
|
| /aosp_15_r20/external/eigen/unsupported/test/ |
| H A D | autodiff.cpp | 214 // TODO also check actual derivatives!
227 // TODO also check actual derivatives!
235 ap.x().derivatives() = Vector2f::UnitX(); in test_autodiff_vector()
236 ap.y().derivatives() = Vector2f::UnitY(); in test_autodiff_vector()
268 //set unit vectors for the derivative directions (partial derivatives of the input vector) in test_autodiff_hessian()
269 x(0).derivatives().resize(2); in test_autodiff_hessian()
270 x(0).derivatives().setZero(); in test_autodiff_hessian()
271 x(0).derivatives()(0)= 1; in test_autodiff_hessian()
272 x(1).derivatives().resize(2); in test_autodiff_hessian()
273 x(1).derivatives().setZero(); in test_autodiff_hessian()
[all …]
|
| H A D | autodiff_scalar.cpp | 32 VERIFY_IS_APPROX(res.derivatives(), x.derivatives()); in check_atan2()
36 VERIFY_IS_APPROX(res.derivatives(), x.derivatives()); in check_atan2()
52 VERIFY_IS_APPROX(res1.derivatives().x(), Scalar(1.0) / (cosh_px * cosh_px)); in check_hyperbolic_functions()
56 VERIFY_IS_APPROX(res2.derivatives().x(), cosh_px); in check_hyperbolic_functions()
60 VERIFY_IS_APPROX(res3.derivatives().x(), std::sinh(p.x())); in check_hyperbolic_functions()
66 VERIFY_IS_APPROX(res1.derivatives().x(), Scalar(0.896629559604914)); in check_hyperbolic_functions()
69 VERIFY_IS_APPROX(res2.derivatives().x(), Scalar(1.056071867829939)); in check_hyperbolic_functions()
72 VERIFY_IS_APPROX(res3.derivatives().x(), Scalar(0.339540557256150)); in check_hyperbolic_functions()
|
| /aosp_15_r20/external/pytorch/tools/autograd/ |
| H A D | load_derivatives.py | 1 # Parses derivatives.yaml into autograd functions
93 # prefer manually-defined derivatives if any
113 # so we can generate derivatives for them separately.
150 # Ensure that the old derivatives.yaml schema with no dispatch key can be loaded.
171 # that appear in derivatives.yaml. used_dispatch_keys is useful for generating
217 # Check that the referenced derivatives in the formula are in bounds
276 derivatives: list[Derivative],
321 f"Derivative definition of {defn_name} in derivatives.yaml defines the "
326 if not len(derivatives) == 1:
328 f"Derivative definition of {defn_name} in derivatives.yaml defines the "
[all …]
|
| /aosp_15_r20/external/pytorch/torchgen/api/ |
| H A D | autograd.py | 37 # Represents a backward formula that calculates derivatives for one
52 # Names of the arguments for which this formula calculates derivatives.
65 # Represents a forward formula that calculates forward derivatives
75 # derivatives
79 # derivatives
82 # Inputs for which the forward derivatives are required for this formula
92 # If this formula is specified in derivatives.yaml or if we are re-using the
100 # The base name read from derivatives.yaml.
109 # We first use the schema string (under the 'name' key) in derivatives.yaml
113 # derivatives.yaml entry. If there is no exact match, then we choose the
[all …]
|
| /aosp_15_r20/external/eigen/unsupported/Eigen/src/Splines/ |
| H A D | SplineFitting.h | 243 * derivatives.
246 * \param derivatives The desired derivatives of the interpolating spline at interpolation
249 * must be the same size as @a derivatives.
252 * \returns A spline interpolating @a points with @a derivatives at those points.
260 const PointArrayType& derivatives,
265 * \brief Fits an interpolating spline to the given data points and derivatives.
268 … * \param derivatives The desired derivatives of the interpolating spline at interpolation points.
270 * must be the same size as @a derivatives.
274 * \returns A spline interpolating @a points with @a derivatives at those points.
282 const PointArrayType& derivatives,
[all …]
|
| H A D | Spline.h | 54 /** \brief The data type used to store the values of the basis function derivatives. */
115 * \brief Evaluation of spline derivatives of up-to given order.
124 * \param order The order up to which the derivatives are computed.
127 derivatives(Scalar u, DenseIndex order) const;
130 * \copydoc Spline::derivatives
136 derivatives(Scalar u, DenseIndex order = DerivativeOrder) const;
158 * \brief Computes the non-zero spline basis function derivatives up to given order.
167 * derivatives are computed.
168 * \param order The order up to which the basis function derivatives are computes.
328 // Retrieve the basis function derivatives up to the desired order... in derivativesImpl()
[all …]
|
| /aosp_15_r20/external/tensorflow/tensorflow/python/ops/ |
| H A D | gradients_impl.py | 49 """Constructs symbolic derivatives of sum of `ys` w.r.t. x in `xs`.
55 `gradients()` adds ops to the graph to output the derivatives of `ys` with
63 derivatives using a different initial gradient for each y (e.g., if
70 other things, this allows computation of partial derivatives as opposed to
71 total derivatives. For example:
79 Here the partial derivatives `g` evaluate to `[1.0, 1.0]`, compared to the
80 total derivatives `tf.gradients(a + b, [a, b])`, which take into account the
116 phase. This function is used to evaluate the derivatives of the cost function
181 """Constructs symbolic derivatives of sum of `ys` w.r.t. x in `xs`.
191 `gradients()` adds ops to the graph to output the derivatives of `ys` with
[all …]
|
| /aosp_15_r20/external/apache-commons-math/src/main/java/org/apache/commons/math3/analysis/interpolation/ |
| H A D | FieldHermiteInterpolator.java | 32 /** Polynomial interpolator using both sample values and sample derivatives.
35 * and provided derivatives. There is one polynomial for each component of
37 * polynomials depends on the number of points and number of derivatives at each
39 * any derivatives all have degree n-1. The interpolation polynomials for n
73 * derivatives.
79 * @param value value and derivatives of the sample point
85 * @exception MathArithmeticException if the number of derivatives is larger
162 /** Interpolate value and first derivatives at a specified abscissa.
165 * @return interpolated value and derivatives (value in row 0,
170 public T[][] derivatives(T x, int order) throws NoDataException, NullArgumentException { in derivatives() method in FieldHermiteInterpolator
[all …]
|
| H A D | BicubicSplineInterpolatingFunction.java | 44 * and function derivatives values
72 * Partial derivatives.
73 * The value of the first index determines the kind of derivatives:
74 * 0 = first partial derivatives wrt x
75 * 1 = first partial derivatives wrt y
76 * 2 = second partial derivatives wrt x
77 * 3 = second partial derivatives wrt y
78 * 4 = cross partial derivatives
121 * needed for calling any of the methods that compute the partial derivatives
203 // Compute all partial derivatives. in BicubicSplineInterpolatingFunction()
[all …]
|
| /aosp_15_r20/external/apache-commons-math/src/main/java/org/apache/commons/math/ode/nonstiff/ |
| H A D | AdamsNordsieckTransformer.java | 36 * classical representation with several previous first derivatives and Nordsieck
37 * representation with higher order scaled derivatives.</p>
39 * <p>We define scaled derivatives s<sub>i</sub>(n) at step n as:
49 * uses first derivatives only, i.e. it handles y<sub>n</sub>, s<sub>1</sub>(n) and
57 * higher degrees scaled derivatives all taken at the same step, i.e it handles y<sub>n</sub>,
139 /** Initialization matrix for the higher order derivatives wrt y'', y''' ... */
142 … /** Update matrix for the higher order derivatives h<sup>2</sup>/2y'', h<sup>3</sup>/6 y''' ... */
145 /** Update coefficients of the higher order derivatives wrt y'. */
254 /** Initialize the high order scaled derivatives at step start.
256 * @param multistep scaled derivatives after step start (hy'1, ..., hy'k-1)
[all …]
|
| H A D | AdamsIntegrator.java | 97 /** Update the high order scaled derivatives for Adams integrators (phase 1).
98 * <p>The complete update of high order derivatives has a form similar to:
103 * @param highOrder high order scaled derivatives
105 * @return updated high order derivatives
112 /** Update the high order scaled derivatives Adams integrators (phase 2).
113 * <p>The complete update of high order derivatives has a form similar to:
119 * @param start first order scaled derivatives at step start
120 * @param end first order scaled derivatives at step end
121 * @param highOrder high order scaled derivatives, will be modified
|
| /aosp_15_r20/external/apache-commons-math/src/main/java/org/apache/commons/math3/ode/nonstiff/ |
| H A D | AdamsNordsieckTransformer.java | 38 * classical representation with several previous first derivatives and Nordsieck
39 * representation with higher order scaled derivatives.</p>
41 * <p>We define scaled derivatives s<sub>i</sub>(n) at step n as:
51 * uses first derivatives only, i.e. it handles y<sub>n</sub>, s<sub>1</sub>(n) and
59 * higher degrees scaled derivatives all taken at the same step, i.e it handles y<sub>n</sub>,
141 …/** Update matrix for the higher order derivatives h<sup>2</sup>/2 y'', h<sup>3</sup>/6 y''' ... */
144 /** Update coefficients of the higher order derivatives wrt y'. */
246 /** Initialize the high order scaled derivatives at step start.
250 * @param yDot first steps derivatives
320 /** Update the high order scaled derivatives for Adams integrators (phase 1).
[all …]
|
| H A D | AdamsNordsieckFieldTransformer.java | 36 * classical representation with several previous first derivatives and Nordsieck
37 * representation with higher order scaled derivatives.</p>
39 * <p>We define scaled derivatives s<sub>i</sub>(n) at step n as:
49 * uses first derivatives only, i.e. it handles y<sub>n</sub>, s<sub>1</sub>(n) and
57 * higher degrees scaled derivatives all taken at the same step, i.e it handles y<sub>n</sub>,
146 …/** Update matrix for the higher order derivatives h<sup>2</sup>/2 y'', h<sup>3</sup>/6 y''' ... */
149 /** Update coefficients of the higher order derivatives wrt y'. */
247 /** Initialize the high order scaled derivatives at step start.
251 * @param yDot first steps derivatives
322 /** Update the high order scaled derivatives for Adams integrators (phase 1).
[all …]
|
| H A D | AdamsIntegrator.java | 102 /** Update the high order scaled derivatives for Adams integrators (phase 1).
103 * <p>The complete update of high order derivatives has a form similar to:
108 * @param highOrder high order scaled derivatives
110 * @return updated high order derivatives
117 /** Update the high order scaled derivatives Adams integrators (phase 2).
118 * <p>The complete update of high order derivatives has a form similar to:
124 * @param start first order scaled derivatives at step start
125 * @param end first order scaled derivatives at step end
126 * @param highOrder high order scaled derivatives, will be modified
|
| H A D | AdamsFieldIntegrator.java | 112 /** Update the high order scaled derivatives for Adams integrators (phase 1).
113 * <p>The complete update of high order derivatives has a form similar to:
118 * @param highOrder high order scaled derivatives
120 * @return updated high order derivatives
127 /** Update the high order scaled derivatives Adams integrators (phase 2).
128 * <p>The complete update of high order derivatives has a form similar to:
134 * @param start first order scaled derivatives at step start
135 * @param end first order scaled derivatives at step end
136 * @param highOrder high order scaled derivatives, will be modified
|
| /aosp_15_r20/external/apache-commons-math/src/main/java/org/apache/commons/math/ode/jacobians/ |
| H A D | StepInterpolatorWithJacobians.java | 99 * Get the partial derivatives of the state vector with respect to
104 * @return partial derivatives of the state vector with respect to
113 * Get the partial derivatives of the state vector with respect to
118 * @return partial derivatives of the state vector with respect to
127 * Get the time derivatives of the state vector of the interpolated point.
131 * @return derivatives of the state vector at time {@link #getInterpolatedTime}
139 * Get the time derivatives of the jacobian of the state vector
144 * @return time derivatives of the jacobian of the state vector
153 * Get the time derivatives of the jacobian of the state vector
158 * @return time derivatives of the jacobian of the state vector
|
| /aosp_15_r20/external/mesa3d/src/panfrost/midgard/ |
| H A D | midgard_derivatives.c | 30 /* Derivatives in Midgard are implemented on the texture pipe, rather than the
32 * instructions require (implicit) derivatives to be calculated anyway, so it
34 * texturing ops that calculate derivatives, there are two explicit texture ops
38 * One major caveat is that derivatives can only be calculated on up to a vec2
40 * derivatives will be vec2 (autocalculating mip levels of 2D texture
45 * generation), we generate texture ops 1:1 to the incoming NIR derivatives.
47 * scan for vec3/vec4 derivatives and lower (split) to multiple instructions.
72 /* Returns true if a texturing op computes derivatives either explicitly or
78 /* Only fragment shaders may compute derivatives, but the sense of in mir_op_computes_derivatives()
|
| /aosp_15_r20/external/tensorflow/tensorflow/java/src/main/java/org/tensorflow/op/core/ |
| H A D | Gradients.java | 31 * Adds operations to compute the partial derivatives of sum of {@code y}s w.r.t {@code x}s,
34 …* If {@code Options.dx()} values are set, they are as the initial symbolic partial derivatives of …
40 * The partial derivatives are returned in output {@code dy}, with the size of {@code x}.
60 * @param dx partial derivatives of some loss function {@code L} w.r.t. {@code y}
79 * @param x inputs of the function for which partial derivatives are computed
116 * @param x inputs of the function for which partial derivatives are computed
128 * @param dx partial derivatives of some loss function {@code L} w.r.t. {@code y}
142 * Partial derivatives of {@code y}s w.r.t. {@code x}s, with the size of {@code x}
|
| /aosp_15_r20/external/apache-commons-math/src/main/java/org/apache/commons/math/ode/ |
| H A D | MultistepIntegrator.java | 34 * <p>We define scaled derivatives s<sub>i</sub>(n) at step n as:
43 * the Nordsieck vector with higher degrees scaled derivatives all taken at the same
67 /** Nordsieck matrix of the higher scaled derivatives.
240 /** Initialize the high order scaled derivatives at step start.
242 * @param multistep scaled derivatives after step start (hy'1, ..., hy'k-1)
244 * @return high order scaled derivatives at step start
301 /** Initialize the high order scaled derivatives at step start.
303 * @param multistep scaled derivatives after step start (hy'1, ..., hy'k-1)
305 * @return high order derivatives at step start
339 // compute the high order scaled derivatives in handleStep()
|
| /aosp_15_r20/external/apache-commons-math/src/main/java/org/apache/commons/math/analysis/interpolation/ |
| H A D | BicubicSplineInterpolatingFunction.java | 38 * and function derivatives values
66 * Partial derivatives
67 * The value of the first index determines the kind of derivatives:
68 * 0 = first partial derivatives wrt x
69 * 1 = first partial derivatives wrt y
70 * 2 = second partial derivatives wrt x
71 * 3 = second partial derivatives wrt y
72 * 4 = cross partial derivatives
259 * Compute all partial derivatives.
302 * function partial derivatives values at the four corners of a grid
[all …]
|