text-generation-inference/server/text_generation_server/utils/__init__.py
OlivierDehaene 62f91f78ac
feat(server): support vectorized warpers in flash causal lm (#317)
Co-authored-by: Joel Lamy-Poirier <joel.lamy-poirier@servicenow.com>
2023-05-26 12:30:27 +02:00

39 lines
953 B
Python

from text_generation_server.utils.convert import convert_file, convert_files
from text_generation_server.utils.dist import initialize_torch_distributed
from text_generation_server.utils.hub import (
weight_files,
weight_hub_files,
download_weights,
EntryNotFoundError,
LocalEntryNotFoundError,
RevisionNotFoundError,
)
from text_generation_server.utils.tokens import (
NextTokenChooser,
HeterogeneousNextTokenChooser,
StoppingCriteria,
StopSequenceCriteria,
FinishReason,
Sampling,
Greedy,
)
__all__ = [
"convert_file",
"convert_files",
"initialize_torch_distributed",
"weight_files",
"weight_hub_files",
"download_weights",
"EntryNotFoundError",
"HeterogeneousNextTokenChooser",
"LocalEntryNotFoundError",
"RevisionNotFoundError",
"Greedy",
"NextTokenChooser",
"Sampling",
"StoppingCriteria",
"StopSequenceCriteria",
"FinishReason",
]