MCPcopy Create free account
hub / github.com/THUDM/LongWriter / DataArguments

Class DataArguments

train/main.py:19–34  ·  view source on GitHub ↗

Source from the content-addressed store, hash-verified

17
18@dataclass
19class DataArguments:
20 train_file: str = field(default=None, metadata={"help": "Path to the training data."})
21 validation_file: str = field(default=None, metadata={"help": "Path to the training data."})
22 preprocessing_num_workers: Optional[int] = field(
23 default=1,
24 metadata={"help": "The number of processes to use for the preprocessing."},
25 )
26 prompt_column: Optional[str] = field(
27 default=None,
28 metadata={"help": "The name of the column in the datasets containing the full texts (for summarization)."},
29 )
30 response_column: Optional[str] = field(
31 default=None,
32 metadata={"help": "The name of the column in the datasets containing the summaries (for summarization)."},
33 )
34 batch_method: str = field(default="naive")
35
36@dataclass
37class TrainingArguments(transformers.Seq2SeqTrainingArguments):

Callers

nothing calls this directly

Calls

no outgoing calls

Tested by

no test coverage detected