120 lines
5.3 KiB
Python
120 lines
5.3 KiB
Python
# Copyright (c) 2025 PaddlePaddle Authors. All Rights Reserved.
|
|
#
|
|
# 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.
|
|
|
|
from __future__ import annotations
|
|
|
|
from typing import TYPE_CHECKING
|
|
|
|
from paddle import _C_ops
|
|
from paddle.base.framework import in_dynamic_or_pir_mode
|
|
|
|
if TYPE_CHECKING:
|
|
from paddle import Tensor
|
|
|
|
|
|
def moe_unpermute(
|
|
hidden_states_unzipped: Tensor,
|
|
zipped_expertwise_rowmap: Tensor,
|
|
expert_routemap_topk: Tensor,
|
|
token_prob_unzipped: Tensor,
|
|
total_zipped_tokens: int,
|
|
num_experts: int,
|
|
using_mix_precision: bool = True,
|
|
using_weighted_combine: bool = False,
|
|
name: str | None = None,
|
|
) -> tuple[Tensor, Tensor]:
|
|
r"""
|
|
Args:
|
|
hidden_states_unzipped (Tensor): The input Tensor containing broadcasted and permuted hidden states.
|
|
Shape: (seqlen_broadcasted, token_len). Dtype: bfloat16.
|
|
zipped_expertwise_rowmap (Tensor): The input Tensor recording the mapping relationship for unpermute operation.
|
|
Shape: (seqlen, num_experts). Dtype: int32.
|
|
expert_routemap_topk (Tensor): The input Tensor indicating which expert each token is assigned to.
|
|
Shape: (seqlen, 8). Value range: [-1, num_experts]. Dtype: int32.
|
|
token_prob_unzipped (Tensor): The input Tensor containing flattened expert probabilities corresponding to hidden_states_unzipped.
|
|
Shape: (seqlen_broadcasted, 1). Dtype: float32.
|
|
total_zipped_tokens_num (int): The total number of tokens before permutation for output buffer allocation. Dtype: int32.
|
|
num_experts (int): The number of experts. Dtype: int32.
|
|
using_mix_precision (bool, optional): Whether to use mixed precision during accumulation.
|
|
This option significantly improves precision when number of experts > 4. Default: True.
|
|
using_weighted_combine (bool, optional): Whether to use weighted token accumulation during unpermute.
|
|
Which utilize probs as weights to accumulate tokens. Default: False.
|
|
name (str|None, optional): Name for the operation. Default: None.
|
|
|
|
Returns:
|
|
tuple[Tensor, Tensor]: A tuple containing:
|
|
- hidden_states (Tensor): The output Tensor with unpermuted tokens.
|
|
Shape: (seqlen, token_len). Dtype: bfloat16.
|
|
- expert_prob_topk (Tensor): The output Tensor with unpermuted probabilities.
|
|
Shape: (seqlen, topk). Dtype: float32.
|
|
|
|
Examples:
|
|
.. code-block:: pycon
|
|
|
|
>>> # doctest: +REQUIRES(env:GPU)
|
|
>>> # doctest: +SKIP('This is only support in cuda 12.0+')
|
|
>>> import paddle
|
|
>>> import numpy as np
|
|
>>> import paddle.nn.functional as F
|
|
>>> hidden_states = paddle.randn([3, 128], dtype='bfloat16')
|
|
>>> expert_routemap_topk = paddle.to_tensor(
|
|
... [
|
|
... [-1, 0, -1, -1, 2, -1, -1, -1],
|
|
... [1, -1, -1, -1, -1, -1, -1, -1],
|
|
... [-1, -1, -1, -1, -1, -1, 1, -1],
|
|
... ],
|
|
... dtype='int32',
|
|
... )
|
|
>>> expert_prob_topk = paddle.to_tensor(
|
|
... [
|
|
... [0.0, 0.6, 0.0, 0.0, 0.4, 0.0, 0.0, 0.0],
|
|
... [1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0],
|
|
... [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0, 0.0],
|
|
... ],
|
|
... dtype='float32',
|
|
... )
|
|
>>> num_experts = 3
|
|
>>> tokens_per_expert = [1, 2, 1]
|
|
>>> padding_alignment = 2
|
|
>>> hidden_states_unzipped, zipped_expertwise_rowmap, token_prob_unzipped, scale_unzipped = F.moe_permute(
|
|
... hidden_states,
|
|
... None,
|
|
... expert_routemap_topk,
|
|
... expert_prob_topk,
|
|
... num_experts,
|
|
... tokens_per_expert,
|
|
... padding_alignment,
|
|
... )
|
|
>>> # weighted by probs.
|
|
>>> hidden_states_unzipped = (
|
|
... hidden_states_unzipped.astype("float32") * token_prob_unzipped.astype("float32").unsqueeze(-1)
|
|
... ).astype("bfloat16")
|
|
>>> zipped_tokens, zipped_probs = F.moe_unpermute(
|
|
... hidden_states_unzipped, zipped_expertwise_rowmap, expert_routemap_topk, token_prob_unzipped, 3, 3
|
|
... )
|
|
>>> np.testing.assert_allclose(zipped_tokens.numpy(), hidden_states.numpy(), rtol=1e-05, atol=1e-06)
|
|
"""
|
|
if in_dynamic_or_pir_mode():
|
|
zipped_tokens, zipped_probs_topk = _C_ops.moe_unpermute(
|
|
hidden_states_unzipped,
|
|
zipped_expertwise_rowmap,
|
|
expert_routemap_topk,
|
|
token_prob_unzipped,
|
|
total_zipped_tokens,
|
|
num_experts,
|
|
using_mix_precision,
|
|
using_weighted_combine,
|
|
)
|
|
return (zipped_tokens, zipped_probs_topk)
|