Check the content type of the input. Four types are allowed: str, dict, liststr, listdict. Args: msgs: Raw input messages.

Create prompt templates for various tasks

Computes TEDS score between the prediction and the ground truth of a given sample

Try to parse into a date. Return: tuple (year, month, date) if successful; otherwise None.

Try to cast an object as a dict.

The main function for multi-turn chatting. Will call `chat_inner` with the preprocessed input messages.

Check the content type of the input. Four types are allowed: str, dict, liststr, listdict.

Merge overlapping intervals in a list. Assumes each interval is a list [start, end].

Select a crop size from `crop_size_list` favoring aspect-ratio match. It ranks candidate crop sizes by how well they match the original image's

Add newline characters to sequence

Converts a markdown table to the corresponding html string for TEDS computation.

Enumerate candidate crop sizes (multiples of `patch_size`). The generated sizes are constrained by a maximum aspect ratio. This list is later

Exact match of math if and only if: 1. numerical equal: both can convert to float and are equal 2. symbolic equal: both can convert to sy

Compares attributes of trees

Args: dim (int): Number of input channels. num_heads (int): Number of attention heads in each ViT block.

Compute the accuracy of two bounding boxes based on a specified threshold. Parameters: - box1 (list of float): Bounding box [x_min, y_mi

Compute the accuracy of two bounding boxes based on a specified threshold. Parameters: - box1 (list of float): Bounding box [x_min, y_mi

Parses the 'predict' string and returns a dictionary. Missing or unparseable content is handled gracefully. Parameters: - predict_st

Sort by value, while iterate over element if data is list

Prepare messages for score evaluation Args: item: DataFrame row containing question, answer, and prediction Returns: li

Calculate the IoU between predicted and target bounding boxes.

Calculate VQ related parameters Args: image_height: Image height image_width: Image width vae_scale: VAE s

Compute the F1-score for KIE task between predicted and ground truth dictionaries. Args: preds (dict): The predicted key-value pairs.