Python Tensor, which wraps a swig converted Tensor from CPP Tensor. Args: shape (tuple<int>): a tuple of integers for the tensor shape. I

Python Tensor, which wraps a swig converted Tensor from CPP Tensor. Args: shape (tuple<int>): a tuple of integers for the tensor shape. I

Class for 1.31 unsigned floating-point computation

Generate a SoftMax operator

Generate a Reshape Operator

Generate a Add operator

Generate a Linear3D operator

Generate a Linear3D operator

Element-wise sum of each of the input tensors

`y = exp(x)`, is applied to the tensor elementwise.

Dataset loader for Libsvm data format

Eva worker get model from search strategy The serialize/deserialize is for good scalability. The project can be decouple into multiple servic

`f(x) = a^b`, is applied to the tensor elementwise.

`y = x^0.5`, is applied to the tensor elementwise.

Calculates the hyperbolic tangent of the given input tensor element-wise.

`y = abs(x)`, is applied to the tensor elementwise.

Performs element-wise binary division (with Numpy-style broadcasting support).

`y = log(x)`, is applied to the tensor elementwise.

Calculate the sign of the given input tensor element-wise. If input > 0, output 1. if input < 0, output -1. if input == 0, output 0.

`y = ln(exp(x) + 1)` is applied to the tensor elementwise.

Calculates the softsign `(x/(1+|x|))` of the given input tensor element-wise.

Transpose the input tensor similar to numpy.transpose.

Fetch data from file and generate batches. Load data from folder as PIL.Images and convert them into batch array. Args: img_folder (

Compose several transforms together. Args: transforms: list of transforms to compose. Example: >>> transforms.Compose([

Generate a Conv 2d operator

Init a General Matrix multiplication(Gemm) operator. Compute `Y = alpha * A' * B' + beta * C`, where input tensor A has shape (M, K) or (K, M

Dataset loader for Libsvm data format

LinearLoraConfig: linear lora config class

LogMessageFatal ensures the process will exit in failure after logging this message.

Eva worker send score to search strategy

Normalize a ``numpy.array`` image with mean and standard deviation. This transform does not support PIL Image. Given mean: ``(mean[1],...,mea

Reimplementation of NodeProto from ONNX, but in a form more convenient to work with from Python.

Generate a ReLU operator

Takes a tensor as input and outputs a tensor containing the shape of the input tensor.

Generate a ReLU operator

This will preoritically query data from cache-service

The class of vocab, include 2 dicts of token to index and index to token

Init a conv 2d operator

Calculates the arccosine (inverse of cosine) of the given input tensor, element-wise.

Calculates the hyperbolic arccosine of the given input tensor element-wise.

Performs element-wise binary addition.

Add Bias to each row / column of the Tensor, depending on the axis arg.