msb-public
/
PaddleSpeech
mirror of https://github.com/PaddlePaddle/PaddleSpeech
1
0
Code Issues Projects Releases Wiki Activity

Merge pull request #2155 from Jackwaterveg/develop_dev

Browse Source 
[ASR] remove fluid（except gradclip）
pull/2168/head
Hui Zhang 3 years ago committed by GitHub
parent 3b1d71d63a 3d24316523
commit d098e027ca
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
8 changed files with 19 additions and 219 deletions
Show Stats Download Patch File Download Diff File

2
README.md
Unescape View File

@ -179,7 +179,7 @@ Via the easy-to-use, efficient, flexible and scalable implementation, our vision
## Installation ## Installation
We strongly recommend our users to install PaddleSpeech in **Linux** with *python>=3.7*. We strongly recommend our users to install PaddleSpeech in **Linux** with *python>=3.7* and *paddlepaddle>=2.3.1*.
Up to now, **Linux** supports CLI for the all our tasks, **Mac OSX** and **Windows** only supports PaddleSpeech CLI for Audio Classification, Speech-to-Text and Text-to-Speech. To install `PaddleSpeech`, please see [installation](./docs/source/install.md). Up to now, **Linux** supports CLI for the all our tasks, **Mac OSX** and **Windows** only supports PaddleSpeech CLI for Audio Classification, Speech-to-Text and Text-to-Speech. To install `PaddleSpeech`, please see [installation](./docs/source/install.md).

8
docs/source/install.md
Unescape View File

@ -117,9 +117,9 @@ conda install -y -c gcc_linux-64=8.4.0 gxx_linux-64=8.4.0
``` ```
(Hip: Do not use the last script if you want to install by **Hard** way): (Hip: Do not use the last script if you want to install by **Hard** way):
### Install PaddlePaddle ### Install PaddlePaddle
You can choose the `PaddlePaddle` version based on your system. For example, for CUDA 10.2, CuDNN7.5 install paddlepaddle-gpu 2.2.0: You can choose the `PaddlePaddle` version based on your system. For example, for CUDA 10.2, CuDNN7.5 install paddlepaddle-gpu 2.3.1:
```bash ```bash
python3 -m pip install paddlepaddle-gpu==2.2.0 -i https://mirror.baidu.com/pypi/simple python3 -m pip install paddlepaddle-gpu==2.3.1 -i https://mirror.baidu.com/pypi/simple
``` ```
### Install PaddleSpeech ### Install PaddleSpeech
You can install `paddlespeech` by the following command，then you can use the `ready-made` examples in `paddlespeech` : You can install `paddlespeech` by the following command，then you can use the `ready-made` examples in `paddlespeech` :
@ -180,9 +180,9 @@ Some users may fail to install `kaldiio` due to the default download source, you
```bash ```bash
pip install pytest-runner -i https://pypi.tuna.tsinghua.edu.cn/simple pip install pytest-runner -i https://pypi.tuna.tsinghua.edu.cn/simple
``` ```
Make sure you have GPU and the paddlepaddle version is right. For example, for CUDA 10.2, CuDNN7.5 install paddle 2.2.0: Make sure you have GPU and the paddlepaddle version is right. For example, for CUDA 10.2, CuDNN7.5 install paddle 2.3.1:
```bash ```bash
python3 -m pip install paddlepaddle-gpu==2.2.0 -i https://mirror.baidu.com/pypi/simple python3 -m pip install paddlepaddle-gpu==2.3.1 -i https://mirror.baidu.com/pypi/simple
``` ```
### Install PaddleSpeech in Developing Mode ### Install PaddleSpeech in Developing Mode
```bash ```bash

8
docs/source/install_cn.md
Unescape View File

@ -111,9 +111,9 @@ conda install -y -c gcc_linux-64=8.4.0 gxx_linux-64=8.4.0
``` ```
（提示： 如果你想使用**困难**方式完成安装，请不要使用最后一条命令） （提示： 如果你想使用**困难**方式完成安装，请不要使用最后一条命令）
### 安装 PaddlePaddle ### 安装 PaddlePaddle
你可以根据系统配置选择 PaddlePaddle 版本，例如系统使用 CUDA 10.2， CuDNN7.5 ，你可以安装 paddlepaddle-gpu 2.2.0： 你可以根据系统配置选择 PaddlePaddle 版本，例如系统使用 CUDA 10.2， CuDNN7.5 ，你可以安装 paddlepaddle-gpu 2.3.1：
```bash ```bash
python3 -m pip install paddlepaddle-gpu==2.2.0 -i https://mirror.baidu.com/pypi/simple python3 -m pip install paddlepaddle-gpu==2.3.1 -i https://mirror.baidu.com/pypi/simple
``` ```
### 安装 PaddleSpeech ### 安装 PaddleSpeech
最后安装 `paddlespeech`，这样你就可以使用 `paddlespeech` 中已有的 examples： 最后安装 `paddlespeech`，这样你就可以使用 `paddlespeech` 中已有的 examples：
@ -168,9 +168,9 @@ conda activate tools/venv
conda install -y -c conda-forge sox libsndfile swig bzip2 libflac bc conda install -y -c conda-forge sox libsndfile swig bzip2 libflac bc
``` ```
### 安装 PaddlePaddle ### 安装 PaddlePaddle
请确认你系统是否有 GPU，并且使用了正确版本的 paddlepaddle。例如系统使用 CUDA 10.2, CuDNN7.5 ，你可以安装 paddlepaddle-gpu 2.2.0： 请确认你系统是否有 GPU，并且使用了正确版本的 paddlepaddle。例如系统使用 CUDA 10.2, CuDNN7.5 ，你可以安装 paddlepaddle-gpu 2.3.1：
```bash ```bash
python3 -m pip install paddlepaddle-gpu==2.2.0 -i https://mirror.baidu.com/pypi/simple python3 -m pip install paddlepaddle-gpu==2.3.1 -i https://mirror.baidu.com/pypi/simple
``` ```
### 用开发者模式安装 PaddleSpeech ### 用开发者模式安装 PaddleSpeech
部分用户系统由于默认源的问题，安装中会出现 kaldiio 安转出错的问题，建议首先安装 pytest-runner: 部分用户系统由于默认源的问题，安装中会出现 kaldiio 安转出错的问题，建议首先安装 pytest-runner:

63
paddlespeech/s2t/__init__.py
Unescape View File

@ -18,7 +18,6 @@ from typing import Union
import paddle import paddle
from paddle import nn from paddle import nn
from paddle.fluid import core
from paddle.nn import functional as F from paddle.nn import functional as F
from paddlespeech.s2t.utils.log import Log from paddlespeech.s2t.utils.log import Log
@ -39,46 +38,6 @@ paddle.long = 'int64'
paddle.uint16 = 'uint16' paddle.uint16 = 'uint16'
paddle.cdouble = 'complex128' paddle.cdouble = 'complex128'
def convert_dtype_to_string(tensor_dtype):
"""
Convert the data type in numpy to the data type in Paddle
Args:
tensor_dtype(core.VarDesc.VarType): the data type in numpy.
Returns:
core.VarDesc.VarType: the data type in Paddle.
"""
dtype = tensor_dtype
if dtype == core.VarDesc.VarType.FP32:
return paddle.float32
elif dtype == core.VarDesc.VarType.FP64:
return paddle.float64
elif dtype == core.VarDesc.VarType.FP16:
return paddle.float16
elif dtype == core.VarDesc.VarType.INT32:
return paddle.int32
elif dtype == core.VarDesc.VarType.INT16:
return paddle.int16
elif dtype == core.VarDesc.VarType.INT64:
return paddle.int64
elif dtype == core.VarDesc.VarType.BOOL:
return paddle.bool
elif dtype == core.VarDesc.VarType.BF16:
# since there is still no support for bfloat16 in NumPy,
# uint16 is used for casting bfloat16
return paddle.uint16
elif dtype == core.VarDesc.VarType.UINT8:
return paddle.uint8
elif dtype == core.VarDesc.VarType.INT8:
return paddle.int8
elif dtype == core.VarDesc.VarType.COMPLEX64:
return paddle.complex64
elif dtype == core.VarDesc.VarType.COMPLEX128:
return paddle.complex128
else:
raise ValueError("Not supported tensor dtype %s" % dtype)
if not hasattr(paddle, 'softmax'): if not hasattr(paddle, 'softmax'):
logger.debug("register user softmax to paddle, remove this when fixed!") logger.debug("register user softmax to paddle, remove this when fixed!")
setattr(paddle, 'softmax', paddle.nn.functional.softmax) setattr(paddle, 'softmax', paddle.nn.functional.softmax)
@ -155,28 +114,6 @@ if not hasattr(paddle.Tensor, 'new_full'):
paddle.Tensor.new_full = new_full paddle.Tensor.new_full = new_full
paddle.static.Variable.new_full = new_full paddle.static.Variable.new_full = new_full
def eq(xs: paddle.Tensor, ys: Union[paddle.Tensor, float]) -> paddle.Tensor:
if convert_dtype_to_string(xs.dtype) == paddle.bool:
xs = xs.astype(paddle.int)
return xs.equal(
paddle.to_tensor(
ys, dtype=convert_dtype_to_string(xs.dtype), place=xs.place))
if not hasattr(paddle.Tensor, 'eq'):
logger.debug(
"override eq of paddle.Tensor if exists or register, remove this when fixed!"
)
paddle.Tensor.eq = eq
paddle.static.Variable.eq = eq
if not hasattr(paddle, 'eq'):
logger.debug(
"override eq of paddle if exists or register, remove this when fixed!")
paddle.eq = eq
def contiguous(xs: paddle.Tensor) -> paddle.Tensor: def contiguous(xs: paddle.Tensor) -> paddle.Tensor:
return xs return xs

2
paddlespeech/s2t/models/u2/u2.py
Unescape View File

@ -318,7 +318,7 @@ class U2BaseModel(ASRInterface, nn.Layer):
dim=1) # (B*N, i+1) dim=1) # (B*N, i+1)
# 2.6 Update end flag # 2.6 Update end flag
end_flag = paddle.eq(hyps[:, -1], self.eos).view(-1, 1) end_flag = paddle.equal(hyps[:, -1], self.eos).view(-1, 1)
# 3. Select best of best # 3. Select best of best
scores = scores.view(batch_size, beam_size) scores = scores.view(batch_size, beam_size)

15
paddlespeech/s2t/modules/align.py
Unescape View File

@ -13,8 +13,7 @@
# limitations under the License. # limitations under the License.
import paddle import paddle
from paddle import nn from paddle import nn
import math
from paddlespeech.s2t.modules.initializer import KaimingUniform
""" """
To align the initializer between paddle and torch, To align the initializer between paddle and torch,
the API below are set defalut initializer with priority higger than global initializer. the API below are set defalut initializer with priority higger than global initializer.
@ -82,10 +81,10 @@ class Linear(nn.Linear):
name=None): name=None):
if weight_attr is None: if weight_attr is None:
if global_init_type == "kaiming_uniform": if global_init_type == "kaiming_uniform":
weight_attr = paddle.ParamAttr(initializer=KaimingUniform()) weight_attr = paddle.ParamAttr(initializer=nn.initializer.KaimingUniform(fan_in=None, negative_slope=math.sqrt(5), nonlinearity='leaky_relu'))
if bias_attr is None: if bias_attr is None:
if global_init_type == "kaiming_uniform": if global_init_type == "kaiming_uniform":
bias_attr = paddle.ParamAttr(initializer=KaimingUniform()) bias_attr = paddle.ParamAttr(initializer=nn.initializer.KaimingUniform(fan_in=None, negative_slope=math.sqrt(5), nonlinearity='leaky_relu'))
super(Linear, self).__init__(in_features, out_features, weight_attr, super(Linear, self).__init__(in_features, out_features, weight_attr,
bias_attr, name) bias_attr, name)
@ -105,10 +104,10 @@ class Conv1D(nn.Conv1D):
data_format='NCL'): data_format='NCL'):
if weight_attr is None: if weight_attr is None:
if global_init_type == "kaiming_uniform": if global_init_type == "kaiming_uniform":
weight_attr = paddle.ParamAttr(initializer=KaimingUniform()) weight_attr = paddle.ParamAttr(initializer=nn.initializer.KaimingUniform(fan_in=None, negative_slope=math.sqrt(5), nonlinearity='leaky_relu'))
if bias_attr is None: if bias_attr is None:
if global_init_type == "kaiming_uniform": if global_init_type == "kaiming_uniform":
bias_attr = paddle.ParamAttr(initializer=KaimingUniform()) bias_attr = paddle.ParamAttr(initializer=nn.initializer.KaimingUniform(fan_in=None, negative_slope=math.sqrt(5), nonlinearity='leaky_relu'))
super(Conv1D, self).__init__( super(Conv1D, self).__init__(
in_channels, out_channels, kernel_size, stride, padding, dilation, in_channels, out_channels, kernel_size, stride, padding, dilation,
groups, padding_mode, weight_attr, bias_attr, data_format) groups, padding_mode, weight_attr, bias_attr, data_format)
@ -129,10 +128,10 @@ class Conv2D(nn.Conv2D):
data_format='NCHW'): data_format='NCHW'):
if weight_attr is None: if weight_attr is None:
if global_init_type == "kaiming_uniform": if global_init_type == "kaiming_uniform":
weight_attr = paddle.ParamAttr(initializer=KaimingUniform()) weight_attr = paddle.ParamAttr(initializer=nn.initializer.KaimingUniform(fan_in=None, negative_slope=math.sqrt(5), nonlinearity='leaky_relu'))
if bias_attr is None: if bias_attr is None:
if global_init_type == "kaiming_uniform": if global_init_type == "kaiming_uniform":
bias_attr = paddle.ParamAttr(initializer=KaimingUniform()) bias_attr = paddle.ParamAttr(initializer=nn.initializer.KaimingUniform(fan_in=None, negative_slope=math.sqrt(5), nonlinearity='leaky_relu'))
super(Conv2D, self).__init__( super(Conv2D, self).__init__(
in_channels, out_channels, kernel_size, stride, padding, dilation, in_channels, out_channels, kernel_size, stride, padding, dilation,
groups, padding_mode, weight_attr, bias_attr, data_format) groups, padding_mode, weight_attr, bias_attr, data_format)

2
paddlespeech/s2t/modules/attention.py
Unescape View File

@ -109,7 +109,7 @@ class MultiHeadedAttention(nn.Layer):
# 1. onnx(16/-1, -1/-1, 16/0) # 1. onnx(16/-1, -1/-1, 16/0)
# 2. jit (16/-1, -1/-1, 16/0, 16/4) # 2. jit (16/-1, -1/-1, 16/0, 16/4)
if paddle.shape(mask)[2] > 0: # time2 > 0 if paddle.shape(mask)[2] > 0: # time2 > 0
mask = mask.unsqueeze(1).eq(0) # (batch, 1, *, time2) mask = mask.unsqueeze(1).equal(0) # (batch, 1, *, time2)
# for last chunk, time2 might be larger than scores.size(-1) # for last chunk, time2 might be larger than scores.size(-1)
mask = mask[:, :, :, :paddle.shape(scores)[-1]] mask = mask[:, :, :, :paddle.shape(scores)[-1]]
scores = scores.masked_fill(mask, -float('inf')) scores = scores.masked_fill(mask, -float('inf'))

136
paddlespeech/s2t/modules/initializer.py
Unescape View File

@ -12,142 +12,6 @@
# See the License for the specific language governing permissions and # See the License for the specific language governing permissions and
# limitations under the License. # limitations under the License.
import numpy as np import numpy as np
from paddle.fluid import framework
from paddle.fluid import unique_name
from paddle.fluid.core import VarDesc
from paddle.fluid.initializer import MSRAInitializer
__all__ = ['KaimingUniform']
class KaimingUniform(MSRAInitializer):
r"""Implements the Kaiming Uniform initializer
This class implements the weight initialization from the paper
`Delving Deep into Rectifiers: Surpassing Human-Level Performance on
ImageNet Classification <https://arxiv.org/abs/1502.01852>`_
by Kaiming He, Xiangyu Zhang, Shaoqing Ren and Jian Sun. This is a
robust initialization method that particularly considers the rectifier
nonlinearities.
In case of Uniform distribution, the range is [-x, x], where
.. math::
x = \sqrt{\frac{1.0}{fan\_in}}
In case of Normal distribution, the mean is 0 and the standard deviation
is
.. math::
\sqrt{\\frac{2.0}{fan\_in}}
Args:
fan_in (float32|None): fan_in for Kaiming uniform Initializer. If None, it is\
inferred from the variable. default is None.
Note:
It is recommended to set fan_in to None for most cases.
Examples:
.. code-block:: python
import paddle
import paddle.nn as nn
linear = nn.Linear(2,
4,
weight_attr=nn.initializer.KaimingUniform())
data = paddle.rand([30, 10, 2], dtype='float32')
res = linear(data)
"""
def __init__(self, fan_in=None):
super(KaimingUniform, self).__init__(
uniform=True, fan_in=fan_in, seed=0)
def __call__(self, var, block=None):
"""Initialize the input tensor with MSRA initialization.
Args:
var(Tensor): Tensor that needs to be initialized.
block(Block, optional): The block in which initialization ops
should be added. Used in static graph only, default None.
Returns:
The initialization op
"""
block = self._check_block(block)
assert isinstance(var, framework.Variable)
assert isinstance(block, framework.Block)
f_in, f_out = self._compute_fans(var)
# If fan_in is passed, use it
fan_in = f_in if self._fan_in is None else self._fan_in
if self._seed == 0:
self._seed = block.program.random_seed
# to be compatible of fp16 initalizers
if var.dtype == VarDesc.VarType.FP16 or (
var.dtype == VarDesc.VarType.BF16 and not self._uniform):
out_dtype = VarDesc.VarType.FP32
out_var = block.create_var(
name=unique_name.generate(
".".join(['masra_init', var.name, 'tmp'])),
shape=var.shape,
dtype=out_dtype,
type=VarDesc.VarType.LOD_TENSOR,
persistable=False)
else:
out_dtype = var.dtype
out_var = var
if self._uniform:
limit = np.sqrt(1.0 / float(fan_in))
op = block.append_op(
type="uniform_random",
inputs={},
outputs={"Out": out_var},
attrs={
"shape": out_var.shape,
"dtype": int(out_dtype),
"min": -limit,
"max": limit,
"seed": self._seed
},
stop_gradient=True)
else:
std = np.sqrt(2.0 / float(fan_in))
op = block.append_op(
type="gaussian_random",
outputs={"Out": out_var},
attrs={
"shape": out_var.shape,
"dtype": int(out_dtype),
"mean": 0.0,
"std": std,
"seed": self._seed
},
stop_gradient=True)
if var.dtype == VarDesc.VarType.FP16 or (
var.dtype == VarDesc.VarType.BF16 and not self._uniform):
block.append_op(
type="cast",
inputs={"X": out_var},
outputs={"Out": var},
attrs={"in_dtype": out_var.dtype,
"out_dtype": var.dtype})
if not framework.in_dygraph_mode():
var.op = op
return op
class DefaultInitializerContext(object): class DefaultInitializerContext(object):
""" """

Write Preview
Loading…
Cancel
Save