help.txt

usage: extract.py [-h] [--n N] [--k K] [--alpha ALPHA]
                  [--batch_size BATCH_SIZE] [--temperature TEMPERATURE]
                  [--repetition_penalty REPETITION_PENALTY]
                  [--min_length MIN_LENGTH] [--max_length MAX_LENGTH]
                  [--num_return_sequences NUM_RETURN_SEQUENCES]
                  [--top_p TOP_P] [--top_k TOP_K] [--device DEVICE]
                  [--pretrained_model_name PRETRAINED_MODEL_NAME]
                  [--revision REVISION] [--assets ASSETS] [-d]

optional arguments:
  -h, --help            show this help message and exit
  --n N                 The number of texts you want to sample. Default=10000
  --k K                 The number of texts you want to screen out of the
                        sampled texts. Default=100
  --alpha ALPHA         Threshold value for similarity. Default=0.5
  --batch_size BATCH_SIZE
                        The number of sentences to generate at once. It
                        depends on the maximum VRAM size of the GPU.
                        Default=16
  --temperature TEMPERATURE
                        The value used to module the next token probabilities.
                        Default=1.0
  --repetition_penalty REPETITION_PENALTY
                        The parameter for repetition penalty. 1.0 means no
                        penalty. Default=1.0
  --min_length MIN_LENGTH
                        The minimum length of the sequence to be generated.
                        Default=256
  --max_length MAX_LENGTH
                        The maximum length of the sequence to be generated.
                        Default=256
  --num_return_sequences NUM_RETURN_SEQUENCES
                        Number of generating output statements from LM using a
                        single input.Default=1
  --top_p TOP_P         If set to float < 1, only the most probable tokens
                        with probabilities that add up to top_p or higher are
                        kept for generation. Default=1.0
  --top_k TOP_K         The number of highest probability vocabulary tokens to
                        keep for top-k-filtering. Default=40
  --device DEVICE       The device on which the model will be loaded.
                        Default=cuda:0
  --pretrained_model_name PRETRAINED_MODEL_NAME
                        The pretrained model to use. Default=kakaobrain/kogpt
  --revision REVISION   The specific model version to use. Default=main
  --assets ASSETS       The folder where the output result file (*.csv) will
                        be saved. The file name is saved as
                        '{revision}-{nowtime}-{n}.csv' by default.
                        Default=assets
  -d, --debug           Specifies the debugging mode. Default=False