fffiloni's picture
Upload 164 files
2ada650 verified
raw
history blame
1.06 kB
# Copyright (c) Meta Platforms, Inc. and affiliates.
# This software may be used and distributed according to the terms of the Llama 2 Community License Agreement.
import torch
from torch.distributed.fsdp import (
# FullyShardedDataParallel as FSDP,
# CPUOffload,
MixedPrecision,
# BackwardPrefetch,
# ShardingStrategy,
)
# requires grad scaler in main loop
fpSixteen = MixedPrecision(
param_dtype=torch.float16,
# Gradient communication precision.
reduce_dtype=torch.float16,
# Buffer precision.
buffer_dtype=torch.float16,
)
bfSixteen = MixedPrecision(
param_dtype=torch.bfloat16,
# Gradient communication precision.
reduce_dtype=torch.bfloat16,
# Buffer precision.
buffer_dtype=torch.bfloat16,
cast_forward_inputs=True,
)
bfSixteen_mixed = MixedPrecision(
param_dtype=torch.float32,
reduce_dtype=torch.bfloat16,
buffer_dtype=torch.bfloat16,
)
fp32_policy = MixedPrecision(
param_dtype=torch.float32,
reduce_dtype=torch.float32,
buffer_dtype=torch.float32,
)