MCPcopy Create free account
hub / github.com/THUDM/GLM / __getitem__

Method __getitem__

data_utils/corpora.py:80–86  ·  view source on GitHub ↗
(self, index)

Source from the content-addressed store, hash-verified

78 return self.prompt_lens[idx] + self.text_lens[idx]
79
80 def __getitem__(self, index):
81 prompt = self.prompts[index]
82 text = self.texts[index]
83 if self.to_tokenize:
84 prompt = self.tokenizer.EncodeAsIds(prompt).tokenization
85 text = self.tokenizer.EncodeAsIds(text).tokenization
86 return {"tokens": prompt + text, "loss_masks": [0] * len(prompt) + [1] * len(text)}
87
88 def __len__(self):
89 return len(self.prompts)

Callers

nothing calls this directly

Calls 1

EncodeAsIdsMethod · 0.45

Tested by

no test coverage detected