dpr/data/biencoder_data.py (5 lines): - line 490: if len(row_lin) > 1: # TODO: change to checking cell value tokens - line 502: if len(row_lin) > 1: # TODO: change to checking cell value tokens - line 526: if len(row_lin) > 1: # TODO: change to checking cell value tokens - line 557: if len(row_lin) > 1: # TODO: change to checking cell value tokens - line 584: # TODO: support other types dpr/models/biencoder.py (4 lines): - line 41: # TODO: it is only used by _select_span_with_token. Move them to utils - line 141: # TODO delete once moved to the new method - line 314: # TODO: tmp workaround for EL, remove or revise - line 340: # TODO: make a long term HF compatibility fix dpr/utils/data_utils.py (3 lines): - line 125: # TODO: merge with iterate_ds_sampled_data - line 163: # TODO: reset the iteration status? - line 255: # TODO: clear iterators in some non-hacky way dense_retriever.py (2 lines): - line 62: # TODO: tmp workaround for EL, remove or revise - line 300: # TODO: long term HF state compatibility fix dpr/models/hf_models.py (2 lines): - line 109: # TODO: this is a hack-y logic that uses some private tokenizer structure which can be changed in HF code - line 240: # TODO: move max len to methods params? train_dense_encoder.py (2 lines): - line 174: # TODO: ideally we'd want to just call - line 586: # TODO: offset doesn't work for multiset currently conf/biencoder_train_cfg.yaml (1 line): - line 25: # TODO: move to a conf group dpr/data/reader_data.py (1 line): - line 173: # TODO: check if pytorch process group is initialized dpr/models/reader.py (1 line): - line 52: # TODO: provide segment values conf/extractive_reader_train_cfg.yaml (1 line): - line 58: # TODO: move to a conf group conf/dense_retriever.yaml (1 line): - line 59: # TODO: move to a conf group generate_dense_embeddings.py (1 line): - line 73: # TODO: refactor to avoid 'if' dpr/data/download_data.py (1 line): - line 23: # TODO: move to hydra config group conf/gen_embs.yaml (1 line): - line 40: # TODO: move to a conf group dpr/options.py (1 line): - line 24: # TODO: to be merged with conf_utils.py train_extractive_reader.py (1 line): - line 487: # TODO: remove?