MCPcopy Create free account
hub / github.com/zai-org/CodeGeeX / pad_seq

Method pad_seq

codegeex/data/processor.py:114–127  ·  view source on GitHub ↗
(self, prompt_tokens: List[int], label: int, extra: dict = None)

Source from the content-addressed store, hash-verified

112 self.start_time = 0
113
114 def pad_seq(self, prompt_tokens: List[int], label: int, extra: dict = None) -> Dict[str, List[int]]:
115 total_length = len(prompt_tokens)
116 assert total_length <= self._max_seq_len, f"padding sequence: {total_length} > {self._max_seq_len}"
117 pad_len = self._max_seq_len - total_length
118 input_ids = prompt_tokens + [self._pad_token] * pad_len
119 attention_mask = [1] * len(prompt_tokens) + [0] * pad_len
120 label = [label]
121
122 return {
123 "input_ids": input_ids,
124 "attention_mask": attention_mask,
125 "length": [len(prompt_tokens)],
126 "labels": label
127 }
128 def process_sample(self, sample: LabelSample) -> Iterable[Dict[str, List[int]]]:
129 """
130 Process a sample.

Callers 1

process_sampleMethod · 0.95

Calls

no outgoing calls

Tested by

no test coverage detected