/**
|
* @license
|
* Copyright 2018 Google Inc. 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 { Conv2DInfo, Conv3DInfo } from '../ops/conv_util';
|
import { FusedBatchMatMulConfig, FusedConv2DConfig } from '../ops/fused_util';
|
import { Backend, DataId, Scalar, Tensor, Tensor1D, Tensor2D, Tensor3D, Tensor4D, Tensor5D } from '../tensor';
|
import { BackendValues, DataType, Rank, ShapeMap } from '../types';
|
export declare const EPSILON_FLOAT32 = 1e-7;
|
export declare const EPSILON_FLOAT16 = 0.0001;
|
export interface BackendTimingInfo {
|
kernelMs: number | {
|
error: string;
|
};
|
getExtraProfileInfo?(): string;
|
}
|
export interface TensorStorage {
|
read(dataId: DataId): Promise<BackendValues>;
|
readSync(dataId: DataId): BackendValues;
|
disposeData(dataId: DataId): void;
|
write(values: BackendValues, shape: number[], dtype: DataType): DataId;
|
move(dataId: DataId, values: BackendValues, shape: number[], dtype: DataType): void;
|
memory(): {
|
unreliable: boolean;
|
};
|
/** Returns number of data ids currently in the storage. */
|
numDataIds(): number;
|
}
|
/** Convenient class for storing tensor-related data. */
|
export declare class DataStorage<T> {
|
private backend;
|
private dataMover;
|
private data;
|
private dataIdsCount;
|
constructor(backend: KernelBackend, dataMover: DataMover);
|
get(dataId: DataId): T;
|
set(dataId: DataId, value: T): void;
|
has(dataId: DataId): boolean;
|
delete(dataId: DataId): boolean;
|
numDataIds(): number;
|
}
|
export interface DataMover {
|
/**
|
* To be called by backends whenever they see a dataId that they don't own.
|
* Upon calling this method, the mover will fetch the tensor from another
|
* backend and register it with the current active backend.
|
*/
|
moveData(backend: KernelBackend, dataId: DataId): void;
|
}
|
export interface BackendTimer {
|
time(f: () => void): Promise<BackendTimingInfo>;
|
}
|
/**
|
* The interface that defines the kernels that should be implemented when
|
* adding a new backend. New backends don't need to implement every one of the
|
* methods, this can be done gradually (throw an error for unimplemented
|
* methods).
|
*/
|
export declare class KernelBackend implements TensorStorage, Backend, BackendTimer {
|
time(f: () => void): Promise<BackendTimingInfo>;
|
read(dataId: object): Promise<BackendValues>;
|
readSync(dataId: object): BackendValues;
|
numDataIds(): number;
|
disposeData(dataId: object): void;
|
write(values: BackendValues, shape: number[], dtype: DataType): DataId;
|
move(dataId: DataId, values: BackendValues, shape: number[], dtype: DataType): void;
|
memory(): {
|
unreliable: boolean;
|
reasons?: string[];
|
};
|
/** Returns the highest precision for floats in bits (e.g. 16 or 32) */
|
floatPrecision(): 16 | 32;
|
/** Returns the smallest representable number. */
|
epsilon(): number;
|
batchMatMul(a: Tensor3D, b: Tensor3D, transposeA: boolean, transposeB: boolean): Tensor3D;
|
fusedBatchMatMul({ a, b, transposeA, transposeB, bias, activation, preluActivationWeights }: FusedBatchMatMulConfig): Tensor3D;
|
slice<T extends Tensor>(x: T, begin: number[], size: number[]): T;
|
stridedSlice<T extends Tensor>(x: T, begin: number[], end: number[], strides: number[]): T;
|
unstack(x: Tensor, axis: number): Tensor[];
|
reverse<T extends Tensor>(a: T, axis: number[]): T;
|
concat(tensors: Tensor[], axis: number): Tensor;
|
neg<T extends Tensor>(a: T): T;
|
add(a: Tensor, b: Tensor): Tensor;
|
addN<T extends Tensor>(tensors: T[]): T;
|
subtract(a: Tensor, b: Tensor): Tensor;
|
multiply(a: Tensor, b: Tensor): Tensor;
|
realDivide(a: Tensor, b: Tensor): Tensor;
|
floorDiv(a: Tensor, b: Tensor): Tensor;
|
sum(x: Tensor, axes: number[]): Tensor;
|
prod(x: Tensor, axes: number[]): Tensor;
|
unsortedSegmentSum<T extends Tensor>(x: T, segmentIds: Tensor1D, numSegments: number): Tensor;
|
argMin(x: Tensor, axis: number): Tensor;
|
argMax(x: Tensor, axis: number): Tensor;
|
equal(a: Tensor, b: Tensor): Tensor;
|
notEqual(a: Tensor, b: Tensor): Tensor;
|
less(a: Tensor, b: Tensor): Tensor;
|
lessEqual(a: Tensor, b: Tensor): Tensor;
|
greater(a: Tensor, b: Tensor): Tensor;
|
greaterEqual(a: Tensor, b: Tensor): Tensor;
|
logicalNot<T extends Tensor>(a: T): T;
|
logicalAnd(a: Tensor, b: Tensor): Tensor;
|
logicalOr(a: Tensor, b: Tensor): Tensor;
|
where(condition: Tensor): Tensor2D;
|
select(condition: Tensor, a: Tensor, b: Tensor): Tensor;
|
topk<T extends Tensor>(x: T, k: number, sorted: boolean): [T, T];
|
min(x: Tensor, axes: number[]): Tensor;
|
minimum(a: Tensor, b: Tensor): Tensor;
|
mod(a: Tensor, b: Tensor): Tensor;
|
max(x: Tensor, axes: number[]): Tensor;
|
maximum(a: Tensor, b: Tensor): Tensor;
|
all(x: Tensor, axes: number[]): Tensor;
|
any(x: Tensor, axes: number[]): Tensor;
|
squaredDifference(a: Tensor, b: Tensor): Tensor;
|
ceil<T extends Tensor>(x: T): T;
|
floor<T extends Tensor>(x: T): T;
|
round<T extends Tensor>(x: T): T;
|
sign<T extends Tensor>(x: T): T;
|
isNaN<T extends Tensor>(x: T): T;
|
isInf<T extends Tensor>(x: T): T;
|
isFinite<T extends Tensor>(x: T): T;
|
pow<T extends Tensor>(a: T, b: Tensor): T;
|
exp<T extends Tensor>(x: T): T;
|
expm1<T extends Tensor>(x: T): T;
|
softmax<T extends Tensor>(x: T, dim: number): T;
|
log<T extends Tensor>(x: T): T;
|
log1p<T extends Tensor>(x: T): T;
|
sqrt<T extends Tensor>(x: T): T;
|
rsqrt<T extends Tensor>(x: T): T;
|
square<T extends Tensor>(x: T): T;
|
reciprocal<T extends Tensor>(x: T): T;
|
relu<T extends Tensor>(x: T): T;
|
relu6<T extends Tensor>(x: T): T;
|
prelu<T extends Tensor>(x: T, a: T): T;
|
elu<T extends Tensor>(x: T): T;
|
eluDer<T extends Tensor>(dy: T, y: T): T;
|
selu<T extends Tensor>(x: T): T;
|
int<T extends Tensor>(x: T): T;
|
clip<T extends Tensor>(x: T, min: number, max: number): T;
|
abs<T extends Tensor>(x: T): T;
|
complexAbs<T extends Tensor>(x: T): T;
|
sigmoid<T extends Tensor>(x: T): T;
|
softplus<T extends Tensor>(x: T): T;
|
sin<T extends Tensor>(x: T): T;
|
cos<T extends Tensor>(x: T): T;
|
tan<T extends Tensor>(x: T): T;
|
asin<T extends Tensor>(x: T): T;
|
acos<T extends Tensor>(x: T): T;
|
atan<T extends Tensor>(x: T): T;
|
atan2<T extends Tensor>(a: T, b: T): T;
|
sinh<T extends Tensor>(x: T): T;
|
cosh<T extends Tensor>(x: T): T;
|
tanh<T extends Tensor>(x: T): T;
|
asinh<T extends Tensor>(x: T): T;
|
acosh<T extends Tensor>(x: T): T;
|
atanh<T extends Tensor>(x: T): T;
|
erf<T extends Tensor>(x: T): T;
|
step<T extends Tensor>(x: T, alpha: number): T;
|
fusedConv2d({ input, filter, convInfo, bias, activation, preluActivationWeights }: FusedConv2DConfig): Tensor4D;
|
conv2d(x: Tensor4D, filter: Tensor4D, convInfo: Conv2DInfo): Tensor4D;
|
conv2dDerInput(dy: Tensor4D, filter: Tensor4D, convInfo: Conv2DInfo): Tensor4D;
|
conv2dDerFilter(x: Tensor4D, dY: Tensor4D, convInfo: Conv2DInfo): Tensor4D;
|
fusedDepthwiseConv2D({ input, filter, convInfo, bias, activation, preluActivationWeights }: FusedConv2DConfig): Tensor4D;
|
depthwiseConv2D(input: Tensor4D, filter: Tensor4D, convInfo: Conv2DInfo): Tensor4D;
|
depthwiseConv2DDerInput(dy: Tensor4D, filter: Tensor4D, convInfo: Conv2DInfo): Tensor4D;
|
depthwiseConv2DDerFilter(x: Tensor4D, dY: Tensor4D, convInfo: Conv2DInfo): Tensor4D;
|
conv3d(x: Tensor5D, filter: Tensor5D, convInfo: Conv3DInfo): Tensor5D;
|
conv3dDerInput(dy: Tensor5D, filter: Tensor5D, convInfo: Conv3DInfo): Tensor5D;
|
conv3dDerFilter(x: Tensor5D, dY: Tensor5D, convInfo: Conv3DInfo): Tensor5D;
|
maxPool(x: Tensor4D, convInfo: Conv2DInfo): Tensor4D;
|
maxPoolBackprop(dy: Tensor4D, x: Tensor4D, y: Tensor4D, convInfo: Conv2DInfo): Tensor4D;
|
avgPool(x: Tensor4D, convInfo: Conv2DInfo): Tensor4D;
|
avgPoolBackprop(dy: Tensor4D, x: Tensor4D, convInfo: Conv2DInfo): Tensor4D;
|
avgPool3d(x: Tensor5D, convInfo: Conv3DInfo): Tensor5D;
|
avgPool3dBackprop(dy: Tensor5D, x: Tensor5D, convInfo: Conv3DInfo): Tensor5D;
|
maxPool3d(x: Tensor5D, convInfo: Conv3DInfo): Tensor5D;
|
maxPool3dBackprop(dy: Tensor5D, x: Tensor5D, y: Tensor5D, convInfo: Conv3DInfo): Tensor5D;
|
reshape<T extends Tensor, R extends Rank>(x: T, shape: ShapeMap[R]): Tensor<R>;
|
cast<T extends Tensor>(x: T, dtype: DataType): T;
|
tile<T extends Tensor>(x: T, reps: number[]): T;
|
pad<T extends Tensor>(x: T, paddings: Array<[number, number]>, constantValue: number): T;
|
transpose<T extends Tensor>(x: T, perm: number[]): T;
|
gather<T extends Tensor>(x: T, indices: Tensor1D, axis: number): T;
|
gatherND(x: Tensor, indices: Tensor): Tensor;
|
scatterND<R extends Rank>(indices: Tensor, updates: Tensor, shape: ShapeMap[R]): Tensor<R>;
|
batchToSpaceND<T extends Tensor>(x: T, blockShape: number[], crops: number[][]): T;
|
spaceToBatchND<T extends Tensor>(x: T, blockShape: number[], paddings: number[][]): T;
|
resizeBilinear(x: Tensor4D, newHeight: number, newWidth: number, alignCorners: boolean): Tensor4D;
|
resizeBilinearBackprop(dy: Tensor4D, x: Tensor4D, alignCorners: boolean): Tensor4D;
|
resizeNearestNeighbor(x: Tensor4D, newHEight: number, newWidth: number, alignCorners: boolean): Tensor4D;
|
resizeNearestNeighborBackprop(dy: Tensor4D, x: Tensor4D, alignCorners: boolean): Tensor4D;
|
batchNormalization(x: Tensor4D, mean: Tensor4D | Tensor1D, variance: Tensor4D | Tensor1D, varianceEpsilon: number, scale?: Tensor4D | Tensor1D, offset?: Tensor4D | Tensor1D): Tensor4D;
|
localResponseNormalization4D(x: Tensor4D, radius: number, bias: number, alpha: number, beta: number): Tensor4D;
|
LRNGrad(dy: Tensor4D, inputImage: Tensor4D, outputImage: Tensor4D, radius: number, bias: number, alpha: number, beta: number): Tensor4D;
|
multinomial(logits: Tensor2D, normalized: boolean, numSamples: number, seed: number): Tensor2D;
|
oneHot(indices: Tensor1D, depth: number, onValue: number, offValue: number): Tensor2D;
|
cumsum(x: Tensor, axis: number, exclusive: boolean, reverse: boolean): Tensor;
|
nonMaxSuppression(boxes: Tensor2D, scores: Tensor1D, maxOutputSize: number, iouThreshold: number, scoreThreshold?: number): Tensor1D;
|
fft(x: Tensor2D): Tensor2D;
|
ifft(x: Tensor2D): Tensor2D;
|
complex<T extends Tensor>(real: T, imag: T): T;
|
real<T extends Tensor>(input: T): T;
|
imag<T extends Tensor>(input: T): T;
|
cropAndResize(image: Tensor4D, boxes: Tensor2D, boxIndex: Tensor1D, cropSize: [number, number], method: 'bilinear' | 'nearest', extrapolationValue: number): Tensor4D;
|
depthToSpace(x: Tensor4D, blockSize: number, dataFormat: string): Tensor4D;
|
split<T extends Tensor>(value: T, sizeSplits: number[], axis: number): T[];
|
sparseToDense<R extends Rank>(sparseIndices: Tensor, sparseValues: Tensor, outputShape: ShapeMap[R], defaultValue: Scalar): Tensor<R>;
|
diag(x: Tensor): Tensor;
|
fill<R extends Rank>(shape: ShapeMap[R], value: number | string, dtype?: DataType): Tensor<R>;
|
onesLike<R extends Rank>(x: Tensor<R>): Tensor<R>;
|
zerosLike<R extends Rank>(x: Tensor<R>): Tensor<R>;
|
linspace(start: number, stop: number, num: number): Tensor1D;
|
dispose(): void;
|
}
|
