atomgen.data package

Data module for the AtomGen library.

This module contains the data classes and functions for pre-processing and collating data for training/inference.

Submodules

• atomgen.data.data_collator module
  • DataCollatorForAtomModeling
    • DataCollatorForAtomModeling.causal
    • DataCollatorForAtomModeling.coords_perturb
    • DataCollatorForAtomModeling.flatten_batch()
    • DataCollatorForAtomModeling.k
    • DataCollatorForAtomModeling.mam
    • DataCollatorForAtomModeling.max_neighbors
    • DataCollatorForAtomModeling.max_radius
    • DataCollatorForAtomModeling.pad
    • DataCollatorForAtomModeling.pad_to_multiple_of
    • DataCollatorForAtomModeling.return_edge_indices
    • DataCollatorForAtomModeling.return_lap_pe
    • DataCollatorForAtomModeling.return_tensors
    • DataCollatorForAtomModeling.tokenizer
    • DataCollatorForAtomModeling.torch_call()
    • DataCollatorForAtomModeling.torch_compute_edges()
    • DataCollatorForAtomModeling.torch_compute_lap_pe()
    • DataCollatorForAtomModeling.torch_mask_tokens()
    • DataCollatorForAtomModeling.torch_perturb_coords()
• atomgen.data.tokenizer module
  • AtomTokenizer
    • AtomTokenizer.build_inputs_with_special_tokens()
    • AtomTokenizer.convert_tokens_to_string()
    • AtomTokenizer.from_pretrained()
    • AtomTokenizer.get_vocab()
    • AtomTokenizer.get_vocab_size()
    • AtomTokenizer.load_vocab()
    • AtomTokenizer.pad()
    • AtomTokenizer.pad_coords()
    • AtomTokenizer.pad_fixed()
    • AtomTokenizer.pad_forces()
    • AtomTokenizer.save_vocabulary()