Read an INI file into easy-to-access name/value pairs. (Note that I've gone for simplicity here rather than speed, but it should be pretty decent.)

Computes and stores the average and current value

Fake Tensor quantizer module This module quantize a tensor and wraps variable. It also can collect relevant variables for calibration. Args:

Supportive descriptor of quantization. Describe how a tensor should be quantized. Args: collection_name_prefix: A string. Determine which colle

r"""Fused Linear and Activation Module.

r"""Fused Linear and activation Module.

BERT model ("Bidirectional Embedding Representations from a Transformer"). Params: config: a BertConfig class instance with the configura

r"""Fused Linear and Activation Module.

Computes and stores the average and current value

A single training/test example for simple sequence classification.

Runs end-to-end tokenization: punctuation splitting + wordpiece

Writes InputFeature to TF example file.

A single training/test example for simple sequence classification.

A single training/test example for simple sequence classification.

Implements BERT version of Adam algorithm with weight decay fix. Params: lr: learning rate warmup: portion of t_total for the warm

Configuration class to store the configuration of a `BertModel`.

Construct the embeddings from patch, position embeddings.

Runs basic tokenization (punctuation splitting, lower casing, etc.).

Runs basic tokenization (punctuation splitting, lower casing, etc.).

Construct the embeddings from word, position and token_type embeddings.

BERT model ("Bidirectional Encoder Representations from Transformers"). Example usage: ```python # Already been converted into WordPiece token

A single set of features of data.

A single set of features of data.

SentencePiece BPE tokenizer with Jieba integration

Applies a warm up period to the learning rate.

Encodes positions with sine waves as described in https://arxiv.org/abs/1706.03762.

Encodes positions with sine waves as described in https://arxiv.org/abs/1706.03762.

Runs WordPiece tokenization.

A basic Adam optimizer that includes "correct" L2 weight decay.

Parameters structure

A basic Swin Transformer layer for one stage. Args: dim (int): Number of input channels. input_resolution (tuple[int]): Input re

A basic Swin Transformer layer for one stage. Args: dim (int): Number of input channels. input_resolution (tuple[int]): Input re

Configuration for `BertModel`.

Configuration for `BertModel`.

BERT model for multiple choice tasks. This module is composed of the BERT model with a linear layer on top of the pooled output. Params:

BERT model for Question Answering (span extraction). This module is composed of the BERT model with a linear layer on top of the sequence outp