wehub-resource-sync
62 lines
2.7 KiB
Python
62 lines
2.7 KiB
Python
#
|
|
# SPDX-FileCopyrightText: Copyright (c) 1993-2023 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
|
|
# SPDX-License-Identifier: Apache-2.0
|
|
#
|
|
# Licensed under the Apache License, Version 2.0 (the "License");
|
|
# you may not use this file except in compliance with the License.
|
|
# You may obtain a copy of the License at
|
|
#
|
|
# http://www.apache.org/licenses/LICENSE-2.0
|
|
#
|
|
# Unless required by applicable law or agreed to in writing, software
|
|
# distributed under the License is distributed on an "AS IS" BASIS,
|
|
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
# See the License for the specific language governing permissions and
|
|
# limitations under the License.
|
|
#
|
|
|
|
"""Some supportive functions"""
|
|
from absl import logging
|
|
|
|
import torch
|
|
from torch.autograd import Function
|
|
|
|
|
|
class ClipFunction(Function):
|
|
"""An universal tensor clip function
|
|
|
|
Pytorch's clamp() only supports scalar range and doesn't support broadcast. This implementation uses min/max which
|
|
is more genaral. The gradient is defined according to IBM's PACT paper https://arxiv.org/abs/1805.06085, which is
|
|
also the behavior of Tensorflow's clip_by_value()
|
|
"""
|
|
|
|
@staticmethod
|
|
def forward(ctx, input, clip_value_min, clip_value_max):
|
|
output = torch.min(input, clip_value_max)
|
|
output = torch.max(output, clip_value_min)
|
|
ctx.save_for_backward(input, clip_value_min, clip_value_max)
|
|
return output
|
|
|
|
@staticmethod
|
|
def backward(ctx, grad_output):
|
|
input, clip_value_min, clip_value_max = ctx.saved_tensors
|
|
min_mask = (input > clip_value_min).to(grad_output.dtype)
|
|
max_mask = (input < clip_value_max).to(grad_output.dtype)
|
|
grad_input = grad_output * min_mask * max_mask
|
|
|
|
if clip_value_min.requires_grad or clip_value_max.requires_grad:
|
|
logging.log_first_n(logging.WARNING, "Learning clip min/max is experimental, use at your own risk :).", 1)
|
|
if clip_value_min.numel() != 1 or clip_value_max.numel() != 1:
|
|
raise ValueError("Learnable min/max can only be scalar, got size %s and %s." % (clip_value_min.size(),
|
|
clip_value_max.size()))
|
|
|
|
# Ensure the dtypes of min/max grads matches the input dtype
|
|
# This might be necessary if running w/ AMP which will cast to fp32 before `sum()`
|
|
grad_clip_value_min = (grad_output * (1. - min_mask)).sum().to(clip_value_min.dtype) if clip_value_min.requires_grad else None
|
|
grad_clip_value_max = (grad_output * (1. - max_mask)).sum().to(clip_value_min.dtype) if clip_value_max.requires_grad else None
|
|
|
|
return grad_input, grad_clip_value_min, grad_clip_value_max
|
|
|
|
|
|
clip = ClipFunction.apply
|