wehub-resource-sync
26 lines
560 B
Python
26 lines
560 B
Python
from __future__ import annotations
|
|
|
|
from enum import Enum
|
|
|
|
|
|
class EvaluationFramework(str, Enum):
|
|
"""Frameworks to use when evaluating models."""
|
|
|
|
inspect_ai = "inspect_ai"
|
|
lm_eval = "lm_eval"
|
|
|
|
|
|
class PTQMethod(str, Enum):
|
|
"""Methods of post-training quantization."""
|
|
|
|
awq = "awq"
|
|
high_precision = "high_precision"
|
|
gptq = "gptq"
|
|
rtn = "rtn"
|
|
smoothquant = "smoothquant"
|
|
spinquant = "spinquant"
|
|
|
|
|
|
def get_model_size(model_name: str | None) -> float:
|
|
return float(model_name.split("-")[-1][:-1]) if model_name else 0.0
|