resolve dependancy of variables @param triplets: all sparql triplet lines @param filter_lines: filter lines that start with `Filter (s

resolve dependancy of variables @param triplets: all sparql triplet lines @param filter_lines: filter lines that start with `Filter (s

r""" EMA implementation according to TensorBoard.

Convert a triplet to S_expression clause @param tgt_var: target variable @param triplet: triplet in sparql @param parsed_dict:

Convert a triplet to S_expression clause @param tgt_var: target variable @param triplet: triplet in sparql @param parsed_dict:

Create mapping from MID to offset/ID. :return:

Build the surface index. Reads from the surface map on disk and creates a map from surface_form -> offset, score .... :retur

Some entities are identified twice, once with a prefix/suffix and once without. :param identified_entities: :return

Return vocabulary by building a new or reading an existing one. :param index_prefix: :return:

Return surface index by building new or reading existing one. :param index_prefix: :return:

Instantiate entity from string representation. :param line: :return:

Check if the entity name is a perfect match on the text. :param entity: :param text: :return:

return a relation string with '_' and '.' replaced

Run top k predictions, using linear origin map

Run top k predictions, using linear origin map

Combine the linking results from FACC1 and ELQ

convert a cwq sparql to a s_expr

convert a webqsp sparql to a s_expr

Convert dependancy graph to s_expression @param var_dep_list: varialbe dependancy list @param ret_var: return var @param spec_

Convert dependancy graph to s_expression @param var_dep_list: varialbe dependancy list @param ret_var: return var @param spec_

r""" Dispatches a pre-trained model to GPUs with balanced memory. Borrowed from: https://github.com/huggingface/transformers/blob/v4.31.0/src/

extract relation from sparql

r""" Generates model's responses given queries.

r""" Computes scores using given reward model.

Return the correct value representation for a year.

r""" Initializes the adapters. Support full-parameter, freeze and LoRA training. Note that the trainable parameters must be cast to floa

Return true if the tokens marked by start and end indices are a valid entity occurrence. :param tokens: :param start:

return normalized body lines of sparql, specially return filter lines starting with `FILTER (str(` @param lines: sparql lines list

return normalized body lines of sparql, specially return filter lines starting with `FILTER (str(` @param lines: sparql lines list

r""" Event called after logging the last logs.

r""" Event called after a checkpoint save.

r""" Event called at the end of a training step.

r""" Event called at the beginning of training.

Parse CWQ sparqls into s-expressions

Parse body lines @param body_lines: list of sparql body lines @param filter_lines: lines that start with `FILTER (str(?` @para

Parse body lines @param body_lines: list of sparql body lines @param ret_var: return var, default `?x` @param filter_lines: li

parse a sparql query into a s-expression @param query: sparql query @param mid_list: all mids appeared in the sparql query