MCPcopy Create free account
hub / github.com/LUMIA-Group/MemoryDecoder / parse_args

Function parse_args

train_memdec.py:77–315  ·  view source on GitHub ↗
()

Source from the content-addressed store, hash-verified

75MODEL_TYPES = tuple(conf.model_type for conf in MODEL_CONFIG_CLASSES)
76
77def parse_args():
78 parser = argparse.ArgumentParser(description="Finetune a transformers model on a causal language modeling task")
79 parser.add_argument(
80 "--dataset_name",
81 type=str,
82 default=None,
83 help="The name of the dataset to use (via the datasets library).",
84 )
85 parser.add_argument(
86 "--dataset_config_name",
87 type=str,
88 default=None,
89 help="The configuration name of the dataset to use (via the datasets library).",
90 )
91 parser.add_argument(
92 "--dataset_split_name",
93 type=str,
94 default="test",
95 )
96 parser.add_argument(
97 "--train_file", type=str, default=None, help="A csv, txt or a json file containing the training data."
98 )
99 parser.add_argument(
100 "--validation_file", type=str, default=None, help="A csv, txt or a json file containing the validation data."
101 )
102 parser.add_argument(
103 "--validation_split_percentage",
104 default=5,
105 help="The percentage of the train set used as validation set in case there's no validation split",
106 )
107 parser.add_argument(
108 "--model_name_or_path",
109 type=str,
110 help="Path to pretrained model or model identifier from huggingface.co/models.",
111 required=False,
112 )
113 parser.add_argument(
114 "--config_name",
115 type=str,
116 default=None,
117 help="Pretrained config name or path if not the same as model_name",
118 )
119 parser.add_argument(
120 "--tokenizer_name",
121 type=str,
122 default=None,
123 help="Pretrained tokenizer name or path if not the same as model_name",
124 )
125 parser.add_argument(
126 "--use_slow_tokenizer",
127 action="store_true",
128 help="If passed, will use a slow tokenizer (not backed by the 🤗 Tokenizers library).",
129 )
130 parser.add_argument(
131 "--per_device_train_batch_size",
132 type=int,
133 default=8,
134 help="Batch size (per device) for the training dataloader.",

Callers 1

mainFunction · 0.70

Calls

no outgoing calls

Tested by

no test coverage detected