diff --git a/.github/workflows/ci_pipeline.yaml b/.github/workflows/ci_pipeline.yaml index c040e07..6b8fc6e 100644 --- a/.github/workflows/ci_pipeline.yaml +++ b/.github/workflows/ci_pipeline.yaml @@ -37,7 +37,7 @@ jobs: - name: Get target directories id: get_dirs run: | - DIRS=$(find . -maxdepth 1 -type d ! -name "." ! -name ".github" | tr '\n' ' ') + DIRS=$(find . -maxdepth 1 -type d ! -name "." ! -name ".github" ! -name ".git" | tr '\n' ' ') echo "Target directories: $DIRS" diff --git a/audio/README.md b/audio/README.md index 56f628f..9e0aa0c 100644 --- a/audio/README.md +++ b/audio/README.md @@ -1,18 +1,18 @@ -# Audio Applications +## 音频应用 -This directory contains ready-to-use Audio application notebooks built with MindSpore. Each notebook demonstrates a complete or partial workflow—training, finetuning, or inference—along with a brief introduction to the model used. +本目录包含基于MindSpore 2.7.1版本+动态图+mint接口构建的即用型音频应用笔记。每个笔记演示了完整或部分工作流程（训练、微调或推理），并简要介绍了所用模型。 -## Application List +### 应用列表 -| No. | Model | Description | -| :-- | :---- | :-------------------------------- | -| 1 | [WaveNet](./wavenet/) | Includes notebooks for WaveNet training on tasks such as audio synthesis | +| 编号 | Model | 课程描述 | +|------|-----------------|-----------------------------------------------| +| 1 | WaveNet 神经网络 | 包含用于 WaveNet 训练的笔记本，涵盖音频合成等任务 | -## Contributing New Audio Applications +### 贡献新的音频应用 -To contribute a new Audio application: +贡献新的音频应用方法： -1. Place your notebook in the corresponding model directory. -2. If the model does not yet have its own directory, create a new one following the existing structure. -3. Follow the notebook writing and naming standards in the [Contributing Guidelines](https://github.com/mindspore-courses/applications/wiki/Contributing-Guidelines). -4. Update the application list in the README if required. +1. 请将您的笔记本放入对应的模型目录中。 +2. 若模型尚未拥有独立目录，请按现有结构创建新目录。 +3. 遵循贡献指南中的 notebook 编写与命名规范。 +4. 如需要，更新 README 中的应用程序列表。 diff --git a/audio/wavenet/train_wavenet_audio_generation.ipynb b/audio/wavenet/train_wavenet_audio_generation.ipynb index 38af62f..29f4abc 100644 --- a/audio/wavenet/train_wavenet_audio_generation.ipynb +++ b/audio/wavenet/train_wavenet_audio_generation.ipynb @@ -4,7 +4,7 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "# WaveNet音乐生成案例" + "# 基于MindSpore的WaveNet音乐生成" ] }, { @@ -77,100 +77,61 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "## 基于MindSpore的WaveNet音乐生成\n", - "\n", - "### 数据预处理\n", - "\n", - "一般情况下，原始音频存储为16位整数值序列，即音频信号共有65536种量化结果。WaveNet将样本点的预测视为量化区间的分类问题，这意味着最终的SoftMax层将要输出65536个类别的概率。为了降低网络参数量及预测难度，我们通常使用 $\\mu$ 率压扩变换将原始音频信号量化到256个值：\n", - "$$\n", - "f(x_t)=sign(x_t)\\frac{ln(1+\\mu|x_t|)}{ln(1+\\mu)}, -1\n", - " \"动画\"\n", - "\n", + "## 环境准备\n", "\n", - "如图所示，WaveNet**在推理时**，我们根据前n个时刻的样本预测当前时刻的样本值（**即网络的输入序列长度n为网络的感受野**），然后我们将当前时刻的预测值也作为n个输入中的一个输入网络中，预测下一时刻的样本点。\n", + "本案例的运行环境为：\n", "\n", - "而**在训练时**，我们只训练网络根据n个输入预测第$n+1$个值。为了提高效率，我们通常设定网络一次性预测长度为$o$的输出，根据一个预测样本对应网络感受野大小的样本的输入，网络的输入长度应为$n+o-1$。据此我们构建数据集：" + "| Python | MindSpore |\n", + "| :----- | :-------- |\n", + "| 3.11 | 2.7.1 |" ] }, { "cell_type": "code", - "execution_count": 1, + "execution_count": null, "metadata": {}, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "Looking in indexes: http://192.168.0.122:8888/repository/pypi/simple\n", - "Requirement already satisfied: librosa in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (0.9.2)\n", - "Requirement already satisfied: soundfile>=0.10.2 in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from librosa) (0.11.0)\n", - "Requirement already satisfied: numba>=0.45.1 in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from librosa) (0.56.4)\n", - "Requirement already satisfied: pooch>=1.0 in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from librosa) (1.6.0)\n", - "Requirement already satisfied: scipy>=1.2.0 in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from librosa) (1.5.4)\n", - "Requirement already satisfied: numpy>=1.17.0 in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from librosa) (1.21.6)\n", - "Requirement already satisfied: decorator>=4.0.10 in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from librosa) (4.4.1)\n", - "Requirement already satisfied: packaging>=20.0 in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from librosa) (21.3)\n", - "Requirement already satisfied: scikit-learn>=0.19.1 in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from librosa) (0.24.0)\n", - "Requirement already satisfied: resampy>=0.2.2 in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from librosa) (0.4.2)\n", - "Requirement already satisfied: audioread>=2.1.9 in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from librosa) (3.0.0)\n", - "Requirement already satisfied: joblib>=0.14 in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from librosa) (1.1.0)\n", - "Requirement already satisfied: importlib-metadata in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from numba>=0.45.1->librosa) (4.12.0)\n", - "Requirement already satisfied: llvmlite<0.40,>=0.39.0dev0 in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from numba>=0.45.1->librosa) (0.39.1)\n", - "Requirement already satisfied: setuptools in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from numba>=0.45.1->librosa) (58.0.4)\n", - "Requirement already satisfied: pyparsing!=3.0.5,>=2.0.2 in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from packaging>=20.0->librosa) (3.0.6)\n", - "Requirement already satisfied: appdirs>=1.3.0 in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from pooch>=1.0->librosa) (1.4.4)\n", - "Requirement already satisfied: requests>=2.19.0 in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from pooch>=1.0->librosa) (2.23.0)\n", - "Requirement already satisfied: urllib3!=1.25.0,!=1.25.1,<1.26,>=1.21.1 in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from requests>=2.19.0->pooch>=1.0->librosa) (1.25.11)\n", - "Requirement already satisfied: certifi>=2017.4.17 in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from requests>=2.19.0->pooch>=1.0->librosa) (2021.10.8)\n", - "Requirement already satisfied: idna<3,>=2.5 in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from requests>=2.19.0->pooch>=1.0->librosa) (2.10)\n", - "Requirement already satisfied: chardet<4,>=3.0.2 in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from requests>=2.19.0->pooch>=1.0->librosa) (3.0.4)\n", - "Requirement already satisfied: threadpoolctl>=2.0.0 in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from scikit-learn>=0.19.1->librosa) (3.0.0)\n", - "Requirement already satisfied: cffi>=1.0 in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from soundfile>=0.10.2->librosa) (1.14.0)\n", - "Requirement already satisfied: pycparser in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from cffi>=1.0->soundfile>=0.10.2->librosa) (2.21)\n", - "Requirement already satisfied: typing-extensions>=3.6.4 in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from importlib-metadata->numba>=0.45.1->librosa) (3.10.0.2)\n", - "Requirement already satisfied: zipp>=0.5 in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from importlib-metadata->numba>=0.45.1->librosa) (3.8.1)\n", - "Looking in indexes: http://192.168.0.122:8888/repository/pypi/simple\n", - "Requirement already satisfied: soundfile in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (0.11.0)\n", - "Requirement already satisfied: cffi>=1.0 in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from soundfile) (1.14.0)\n", - "Requirement already satisfied: pycparser in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from cffi>=1.0->soundfile) (2.21)\n", - "Looking in indexes: http://192.168.0.122:8888/repository/pypi/simple\n", - "Requirement already satisfied: nnmnkwii in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (0.1.1)\n", - "Requirement already satisfied: fastdtw in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from nnmnkwii) (0.3.4)\n", - "Requirement already satisfied: pysptk>=0.1.17 in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from nnmnkwii) (0.2.0)\n", - "Requirement already satisfied: cython>=0.28.0 in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from nnmnkwii) (0.29.14)\n", - "Requirement already satisfied: scikit-learn in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from nnmnkwii) (0.24.0)\n", - "Requirement already satisfied: tqdm in /home/ma-user/modelarts-dev/modelarts-sdk (from nnmnkwii) (4.62.3)\n", - "Requirement already satisfied: scipy in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from nnmnkwii) (1.5.4)\n", - "Requirement already satisfied: decorator in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from pysptk>=0.1.17->nnmnkwii) (4.4.1)\n", - "Requirement already satisfied: numpy in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from fastdtw->nnmnkwii) (1.21.6)\n", - "Requirement already satisfied: joblib>=0.11 in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from scikit-learn->nnmnkwii) (1.1.0)\n", - "Requirement already satisfied: threadpoolctl>=2.0.0 in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from scikit-learn->nnmnkwii) (3.0.0)\n" - ] - } - ], + "outputs": [], "source": [ - "# 安装音频处理所需要的依赖包，通过librosa对音频进行导入, 通过soundfile对音频导出，利用nnmnkwii提供的算法接口对音频进行μ率扩缩\n", - "!pip install librosa\n", - "!pip install soundfile\n", - "!pip install nnmnkwii" + "# 检查mindspore版本是否为2.7.1，如果不是则取消下个单元格注释进行安装\n", + "!pip show mindspore" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ - "我们首先调用```generate_dataset```方法将原始音频文件进行μ率压缩及量化得到用于网络训练和推理的数据集。" + "如果你在如[昇思大模型平台](https://xihe.mindspore.cn/training-projects)、[华为云ModelArts](https://www.huaweicloud.com/product/modelarts.html)、[启智社区](https://openi.pcl.ac.cn/)等算力平台的Jupyter在线编程环境中运行本案例，可取消如下代码的注释，进行依赖库安装：" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "# 安装mindspore==2.7.1版本，如需更换mindspore版本，可更改下面 MINDSPORE_VERSION 变量\n", + "# !pip uninstall mindspore -y\n", + "# %env MINDSPORE_VERSION=2.7.1\n", + "# !pip install mindspore==2.7.1 -i https://repo.mindspore.cn/pypi/simple --trusted-host repo.mindspore.cn --extra-index-url https://repo.huaweicloud.com/repository/pypi/simple" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "# 安装音频处理所需要的依赖包，通过librosa对音频进行导入, 通过soundfile对音频导出，利用nnmnkwii提供的算法接口对音频进行μ率扩缩\n", + "!pip install librosa\n", + "!pip install soundfile\n", + "!pip install nnmnkwii" ] }, { @@ -187,9 +148,57 @@ "\n", "import librosa\n", "from nnmnkwii import preprocessing as pre\n", - "import soundfile as sf\n", + "import soundfile as sf" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "其他场景可参考[MindSpore安装指南](https://www.mindspore.cn/install)进行环境搭建。" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "### 数据预处理\n", + "\n", + "一般情况下，原始音频存储为16位整数值序列，即音频信号共有65536种量化结果。WaveNet将样本点的预测视为量化区间的分类问题，这意味着最终的SoftMax层将要输出65536个类别的概率。为了降低网络参数量及预测难度，我们通常使用 $\\mu$ 率压扩变换将原始音频信号量化到256个值：\n", + "$$\n", + "f(x_t)=sign(x_t)\\frac{ln(1+\\mu|x_t|)}{ln(1+\\mu)}, -1\n", + " \"动画\"\n", + "\n", "\n", + "如图所示，WaveNet**在推理时**，我们根据前n个时刻的样本预测当前时刻的样本值（**即网络的输入序列长度n为网络的感受野**），然后我们将当前时刻的预测值也作为n个输入中的一个输入网络中，预测下一时刻的样本点。\n", + "\n", + "而**在训练时**，我们只训练网络根据n个输入预测第$n+1$个值。为了提高效率，我们通常设定网络一次性预测长度为$o$的输出，根据一个预测样本对应网络感受野大小的样本的输入，网络的输入长度应为$n+o-1$。据此我们构建数据集：" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "我们首先调用```generate_dataset```方法将原始音频文件进行μ率压缩及量化得到用于网络训练和推理的数据集。" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ "def generate_dataset(file_location, out_file, sampling_rate=16000, mono=True):\n", " audio_files = Path(file_location).glob(\"*\")\n", " processed_files = []\n", @@ -295,7 +304,8 @@ "metadata": {}, "outputs": [], "source": [ - "from mindspore import nn\n", + "from mindspore import nn,mint\n", + "import math\n", "\n", "\n", "class ResidualConv1dGLU(nn.Cell):\n", @@ -319,21 +329,17 @@ " dropout=1 - 0.95, dilation=1, cin_channels=-1, gin_channels=-1, padding=None, causal=True):\n", " super(ResidualConv1dGLU, self).__init__()\n", " self.dropout = dropout\n", - " self.dropout_op = nn.Dropout(keep_prob=1. - self.dropout)\n", - " # self.eval_split_op = P.Split(axis=-1, output_num=2)\n", - " self.train_split_op = P.Split(axis=1, output_num=2)\n", - " self.tanh = P.Tanh()\n", - " self.sigmoid = P.Sigmoid()\n", - " self.mul = P.Mul()\n", - " self.add = P.Add()\n", + " self.dropout_op = mint.nn.Dropout(p=self.dropout)\n", "\n", " padding = (kernel_size - 1) * dilation\n", " # print(padding)\n", - " self.conv = nn.Conv1d(residual_channels, gate_channels, kernel_size, pad_mode='pad',\n", - " padding=padding, dilation=dilation, has_bias=bias)\n", + " self.conv = mint.nn.Conv1d(residual_channels, gate_channels, kernel_size, \n", + " padding=padding, dilation=dilation, bias=bias)\n", " gate_out_channels = gate_channels // 2\n", - " self.conv1x1_out = nn.Conv1d(gate_out_channels, residual_channels, kernel_size=1, pad_mode='pad', padding=0, dilation=1, has_bias=True)\n", - " self.conv1x1_skip = nn.Conv1d(gate_out_channels, skip_out_channels, kernel_size=1, pad_mode='pad', padding=0, dilation=1, has_bias=True)\n", + " self.conv1x1_out = mint.nn.Conv1d(gate_out_channels, residual_channels, kernel_size=1, \n", + " padding=0, dilation=1, bias=True)\n", + " self.conv1x1_skip = mint.nn.Conv1d(gate_out_channels, skip_out_channels, kernel_size=1, \n", + " padding=0, dilation=1, bias=True)\n", " self.factor = math.sqrt(0.5)\n", "\n", " def construct(self, x):\n", @@ -351,14 +357,15 @@ " x = self.conv(x)\n", " # remove future time steps\n", " x = x[:, :, :residual.shape[-1]]\n", - " split_op = self.train_split_op\n", - " a, b = split_op(x)\n", - " x = self.mul(self.tanh(a), self.sigmoid(b))\n", + "\n", + " a, b = mint.chunk(x, chunks=2, dim=1)\n", + " \n", + " x = mint.mul(mint.tanh(a), mint.sigmoid(b))\n", "\n", " s = self.conv1x1_skip(x)\n", " x = self.conv1x1_out(x)\n", "\n", - " x = self.add(x, residual) * self.factor\n", + " x = mint.mul(mint.add(x, residual), self.factor)\n", " return x, s\n" ] }, @@ -387,22 +394,14 @@ " skip_out_channels=512,\n", " kernel_size=3, dropout=1 - 0.95):\n", " super().__init__()\n", - " self.transpose_op = P.Transpose()\n", - " self.softmax = P.Softmax(axis=1)\n", - " self.reshape_op = P.Reshape()\n", - " self.zeros_op = P.Zeros()\n", - " self.ones_op = P.Ones()\n", - " self.squeeze_op = P.Squeeze()\n", - " self.expandim_op = P.ExpandDims()\n", - " self.transpose_op = P.Transpose()\n", - " self.tile_op = P.Tile()\n", + " \n", " self.out_channels = out_channels\n", - " self.fack_data = P.Zeros()\n", + " \n", " print(f\"network info: \\n\\tlayers: {layers}\\n\\tblocks:{blocks}\")\n", " assert layers % blocks == 0\n", "\n", " self.layers_per_block = layers // blocks # 24 / 4 = 6\n", - " self.first_conv = nn.Conv1d(out_channels, residual_channels, kernel_size=1)\n", + " self.first_conv = mint.nn.Conv1d(out_channels, residual_channels, kernel_size=1)\n", " conv_layers = []\n", " for layer in range(layers):\n", " dilation = 2 ** (layer % self.layers_per_block) # 1, 2, 4, 8, 16, 32\n", @@ -416,10 +415,10 @@ " conv_layers.append(conv)\n", " self.conv_layers = nn.CellList(conv_layers)\n", " self.last_conv_layers = nn.CellList([\n", - " nn.ReLU(),\n", - " nn.Conv1d(skip_out_channels, skip_out_channels, kernel_size=1),\n", - " nn.ReLU(),\n", - " nn.Conv1d(skip_out_channels, out_channels, kernel_size=1)])\n", + " mint.nn.ReLU(),\n", + " mint.nn.Conv1d(skip_out_channels, skip_out_channels, kernel_size=1),\n", + " mint.nn.ReLU(),\n", + " mint.nn.Conv1d(skip_out_channels, out_channels, kernel_size=1)])\n", " self.factor = math.sqrt(1.0 / len(self.conv_layers)) # sqrt( 1 / 24)\n", "\n", " self.receptive_field = 1\n", @@ -434,15 +433,20 @@ "\n", " B, _, T = x.shape\n", " x = self.first_conv(x)\n", - " skips = 0\n", + " skips = None\n", " for f in self.conv_layers:\n", " x, hidden = f(x) # x=[B, 128, 10240], hidden=[B, 128, 10240]\n", - " skips += hidden\n", - " skips *= self.factor\n", + " if skips is None:\n", + " skips = hidden\n", + " else:\n", + " skips = mint.add(skips, hidden)\n", + " skips = mint.mul(skips, self.factor)\n", " x = skips # x=[B, 128, 10240]\n", " for f in self.last_conv_layers:\n", " x = f(x) # x=[B, 2, 10240]\n", - " x = self.softmax(x) if softmax else x\n", + " if softmax:\n", + " x = mint.softmax(x, dim=1)\n", + " \n", " return x" ] }, @@ -472,7 +476,6 @@ "source": [ "from mindspore import ops\n", "from mindspore.amp import all_finite\n", - "from pathlib import Path\n", "\n", "\n", "def train_loop(model, dataset, loss_fn, optimizer, logger):\n", @@ -489,7 +492,7 @@ " def train_step(data, label):\n", " (loss, logits), grads = grad_fn(data, label)\n", " if all_finite(grads):\n", - " loss = ops.depend(loss, optimizer(grads))\n", + " optimizer(grads)\n", " return loss\n", "\n", " size = dataset.get_dataset_size()\n", @@ -518,170 +521,13 @@ }, { "cell_type": "code", - "execution_count": 6, + "execution_count": null, "metadata": {}, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "network info: \n", - "\tlayers: 24\n", - "\tblocks:4\n", - "receptive filed: 505\n", - "dataset loading ...\n", - "dataset file ./dataset.npz loaded\n", - "dataset loaded.\n", - "\tdataset size: 4217\n", - "\tbatch size: 32\n", - "Epoch 1\n", - "---------------------------------------------------------------------------------------\n" - ] - }, - { - "name": "stderr", - "output_type": "stream", - "text": [ - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:30:26.265.495 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/23-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op734] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:30:26.265.757 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/22-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op735] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:30:26.265.903 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/21-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op736] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:30:26.266.043 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/20-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op737] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:30:26.266.171 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/19-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op738] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:30:26.266.298 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/18-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op739] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:30:26.266.427 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/17-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op740] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:30:26.266.552 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/16-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op741] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:30:26.266.679 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/15-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op742] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:30:26.266.808 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/14-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op743] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:30:26.266.933 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/13-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op744] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:30:26.267.066 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/12-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op745] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:30:26.267.192 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/11-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op746] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:30:26.267.323 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/10-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op747] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:30:26.267.457 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/9-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op748] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:30:26.267.581 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/8-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op749] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:30:26.267.703 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/7-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op750] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:30:26.267.825 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/6-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op751] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:30:26.267.950 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/5-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op752] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:30:26.268.076 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/4-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op753] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:30:26.268.199 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/3-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op754] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:30:26.268.332 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/2-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op755] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:30:26.268.454 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/1-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op756] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:30:26.268.584 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/0-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op757] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:30:45.977.461 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/GatherD-op1466] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:00.861.755 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/23-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op3176] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:00.861.965 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/22-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op3177] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:00.862.111 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/21-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op3178] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:00.862.245 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/20-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op3179] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:00.862.378 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/19-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op3180] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:00.862.509 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/18-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op3181] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:00.862.641 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/17-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op3182] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:00.862.771 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/16-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op3183] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:00.862.900 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/15-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op3184] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:00.863.031 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/14-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op3185] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:00.863.160 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/13-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op3186] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:00.863.288 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/12-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op3187] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:00.863.418 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/11-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op3188] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:00.863.548 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/10-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op3189] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:00.863.677 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/9-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op3190] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:00.863.805 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/8-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op3191] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:00.863.933 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/7-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op3192] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:00.864.061 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/6-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op3193] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:00.864.188 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/5-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op3194] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:00.864.315 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/4-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op3195] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:00.864.439 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/3-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op3196] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:00.864.563 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/2-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op3197] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:00.864.706 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/1-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op3198] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:00.864.837 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/conv_layers-CellList/0-ResidualConv1dGLU/dropout_op-Dropout/DropoutGenMask-op3199] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:01.345.199 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Default/GatherD-op3253] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:01.484.191 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Gradients/Default/conv_layers-CellList/0-ResidualConv1dGLU/gradStridedSlice/StridedSliceGrad-op3316] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:01.503.914 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Gradients/Default/conv_layers-CellList/0-ResidualConv1dGLU/gradStridedSlice/StridedSliceGrad-op3324] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:01.525.403 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Gradients/Default/conv_layers-CellList/0-ResidualConv1dGLU/gradStridedSlice/StridedSliceGrad-op3333] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:01.546.652 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Gradients/Default/conv_layers-CellList/0-ResidualConv1dGLU/gradStridedSlice/StridedSliceGrad-op3342] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:01.567.835 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Gradients/Default/conv_layers-CellList/0-ResidualConv1dGLU/gradStridedSlice/StridedSliceGrad-op3351] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:01.589.048 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Gradients/Default/conv_layers-CellList/0-ResidualConv1dGLU/gradStridedSlice/StridedSliceGrad-op3360] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:01.610.138 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Gradients/Default/conv_layers-CellList/0-ResidualConv1dGLU/gradStridedSlice/StridedSliceGrad-op3369] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:01.630.849 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Gradients/Default/conv_layers-CellList/0-ResidualConv1dGLU/gradStridedSlice/StridedSliceGrad-op3378] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:01.651.860 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Gradients/Default/conv_layers-CellList/0-ResidualConv1dGLU/gradStridedSlice/StridedSliceGrad-op3387] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:01.672.647 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Gradients/Default/conv_layers-CellList/0-ResidualConv1dGLU/gradStridedSlice/StridedSliceGrad-op3396] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:01.693.632 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Gradients/Default/conv_layers-CellList/0-ResidualConv1dGLU/gradStridedSlice/StridedSliceGrad-op3405] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:01.714.772 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Gradients/Default/conv_layers-CellList/0-ResidualConv1dGLU/gradStridedSlice/StridedSliceGrad-op3414] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:01.735.662 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Gradients/Default/conv_layers-CellList/0-ResidualConv1dGLU/gradStridedSlice/StridedSliceGrad-op3423] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:01.756.569 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Gradients/Default/conv_layers-CellList/0-ResidualConv1dGLU/gradStridedSlice/StridedSliceGrad-op3432] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:01.777.597 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Gradients/Default/conv_layers-CellList/0-ResidualConv1dGLU/gradStridedSlice/StridedSliceGrad-op3441] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:01.798.653 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Gradients/Default/conv_layers-CellList/0-ResidualConv1dGLU/gradStridedSlice/StridedSliceGrad-op3450] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:01.819.796 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Gradients/Default/conv_layers-CellList/0-ResidualConv1dGLU/gradStridedSlice/StridedSliceGrad-op3459] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:01.840.969 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Gradients/Default/conv_layers-CellList/0-ResidualConv1dGLU/gradStridedSlice/StridedSliceGrad-op3468] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:01.862.029 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Gradients/Default/conv_layers-CellList/0-ResidualConv1dGLU/gradStridedSlice/StridedSliceGrad-op3477] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:01.883.212 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Gradients/Default/conv_layers-CellList/0-ResidualConv1dGLU/gradStridedSlice/StridedSliceGrad-op3486] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:01.904.353 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Gradients/Default/conv_layers-CellList/0-ResidualConv1dGLU/gradStridedSlice/StridedSliceGrad-op3495] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:01.925.578 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Gradients/Default/conv_layers-CellList/0-ResidualConv1dGLU/gradStridedSlice/StridedSliceGrad-op3504] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:01.946.788 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Gradients/Default/conv_layers-CellList/0-ResidualConv1dGLU/gradStridedSlice/StridedSliceGrad-op3513] don't support int64, reduce precision from int64 to int32.\n", - "[WARNING] DEVICE(23790,ffff904c9780,python):2022-11-10-14:31:01.967.784 [mindspore/ccsrc/plugin/device/ascend/hal/device/kernel_select_ascend.cc:330] FilterRaisedOrReducePrecisionMatchedKernelInfo] Operator:[Gradients/Default/conv_layers-CellList/0-ResidualConv1dGLU/gradStridedSlice/StridedSliceGrad-op3522] don't support int64, reduce precision from int64 to int32.\n" - ] - }, - { - "name": "stdout", - "output_type": "stream", - "text": [ - "loss: 5.545\t\t 0/4217\n", - "loss: 5.127\t\t 20/4217\n", - "loss: 5.352\t\t 40/4217\n", - "loss: 4.907\t\t 60/4217\n", - "loss: 4.723\t\t 80/4217\n", - "loss: 4.418\t\t 100/4217\n", - "loss: 4.196\t\t 120/4217\n", - "loss: 4.028\t\t 140/4217\n", - "loss: 3.891\t\t 160/4217\n", - "loss: 3.751\t\t 180/4217\n", - "loss: 3.662\t\t 200/4217\n", - "loss: 3.766\t\t 220/4217\n", - "loss: 3.652\t\t 240/4217\n", - "loss: 3.498\t\t 260/4217\n", - "loss: 3.604\t\t 280/4217\n", - "loss: 3.656\t\t 300/4217\n", - "loss: 3.427\t\t 320/4217\n", - "loss: 3.288\t\t 340/4217\n", - "loss: 3.388\t\t 360/4217\n", - "loss: 3.370\t\t 380/4217\n", - "loss: 3.483\t\t 400/4217\n", - "loss: 3.350\t\t 420/4217\n", - "loss: 3.363\t\t 440/4217\n", - "loss: 3.350\t\t 460/4217\n", - "loss: 3.275\t\t 480/4217\n", - "loss: 3.510\t\t 500/4217\n", - "loss: 3.328\t\t 520/4217\n", - "loss: 3.381\t\t 540/4217\n", - "loss: 3.267\t\t 560/4217\n", - "loss: 3.473\t\t 580/4217\n", - "loss: 3.411\t\t 600/4217\n", - "loss: 3.316\t\t 620/4217\n", - "loss: 3.433\t\t 640/4217\n", - "loss: 3.348\t\t 660/4217\n", - "loss: 3.236\t\t 680/4217\n", - "loss: 3.154\t\t 700/4217\n", - "loss: 3.193\t\t 720/4217\n", - "loss: 3.409\t\t 740/4217\n", - "loss: 3.447\t\t 760/4217\n", - "loss: 3.242\t\t 780/4217\n", - "loss: 3.294\t\t 800/4217\n", - "loss: 3.258\t\t 820/4217\n", - "loss: 3.242\t\t 840/4217\n", - "loss: 3.227\t\t 860/4217\n", - "loss: 3.342\t\t 880/4217\n", - "loss: 3.431\t\t 900/4217\n", - "loss: 3.336\t\t 920/4217\n", - "loss: 3.375\t\t 940/4217\n", - "loss: 3.272\t\t 960/4217\n", - "loss: 3.360\t\t 980/4217\n", - "loss: 3.304\t\t 1000/4217\n" - ] - } - ], + "outputs": [], "source": [ "# 训练模型\n", "import numpy as np\n", "\n", - "ms.set_context(mode=ms.GRAPH_MODE)\n", - "\n", "dataset_file = \"./dataset.npz\"\n", "wave_location = \"./dataset/\"\n", "\n", @@ -723,22 +569,9 @@ }, { "cell_type": "code", - "execution_count": 8, + "execution_count": null, "metadata": {}, - "outputs": [ - { - "data": { - "image/png": "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", - "text/plain": [ - "
" - ] - }, - "metadata": { - "needs_background": "light" - }, - "output_type": "display_data" - } - ], + "outputs": [], "source": [ "import matplotlib.pyplot as plt\n", "\n", @@ -782,19 +615,32 @@ " head = head_file[random_start: random_start + head_length]\n", " total_length = int(gen_time_length * 16000 * 60)\n", " for _ in tqdm(range(total_length), ncols=60):\n", - " pred = pred_one(model, head[-1024:]).asnumpy()\n", + " # pred = pred_one(model, head[-1024:]).asnumpy()\n", + " current_input = head[-head_length:]\n", + " pred = pred_one(model, current_input).asnumpy()\n", " head = np.append(head, pred)\n", " return head\n", "\n", + "\n", "def pred_one(model, x):\n", + " # 构造 onehot: [Batch, Channels, Time]\n", " onehot = np.eye(256)[x].transpose()\n", - " pred = model(ms.Tensor(onehot).expand_dims(0).astype(ms.float32))\n", + " \n", + " # 转换类型\n", + " input_tensor = ms.Tensor(onehot).astype(ms.float32)\n", + " \n", + " input_tensor = mint.unsqueeze(input_tensor, 0)\n", + " \n", + " # 推理\n", + " pred = model(input_tensor)\n", " pred_sample = pred[0, :, -1]\n", - " return pred_sample.argmax()\n", + " \n", + " return mint.argmax(pred_sample)\n", "\n", "\n", "model = WaveNet(out_channels=256, layers=24, blocks=4)\n", "ms.load_checkpoint(\"wavenet_1.ckpt\", model)\n", + "model.set_train(False)\n", "output = gen_music(model, gen_time_length=1/6, head_location=\"./pred_head.npz\") # 生成一个10s（1/6分钟）的片段\n", "output = pre.inv_mulaw_quantize(output, 256)\n", "sf.write(\"gen.wav\", output, 16000, subtype='PCM_24')\n", @@ -802,18 +648,19 @@ ] }, { - "cell_type": "code", - "execution_count": null, + "cell_type": "markdown", "metadata": {}, - "outputs": [], - "source": [] + "source": [ + "### **参考文献**\n", + "[1] van den Oord, A., Dieleman, S., Zen, H., Simonyan, K., Vinyals, O., Graves, A., Kalchbrenner, N., Senior, A., & Kavukcuoglu, K. (2016). WaveNet: A generative model for raw audio. In Proceedings of the 9th ISCA Speech Synthesis Workshop (SSW 2016)." + ] } ], "metadata": { "kernelspec": { - "display_name": "Python 3.9.13 64-bit", + "display_name": "MindSpore Environment", "language": "python", - "name": "python3" + "name": "mindspore_env" }, "language_info": { "codemirror_mode": { @@ -825,7 +672,7 @@ "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", - "version": "3.9.13" + "version": "3.8.20" }, "vscode": { "interpreter": {