Wulx2050

Results 16 comments of Wulx2050

> 函数 `pde(t, x)` 写错了,dx/dt的符号应该是为正。已经完成了,见: The function `pde(t, x)` is wrong, the sign of dx/dt should be positive. See: https://github.com/Wulx2050/DeepXDE-and-PINN/blob/main/2%E4%BB%80%E4%B9%88%E6%98%AFPINN.ipynb

`DeepXDE`好像并不支持数据生成器(data generator),我要如何修改`DeepXDE`源码使得`batch_size` 参数真正生效从而可以训练几百万条数据? `DeepXDE` does not seem to support data generators. How do I modify the `DeepXDE` source code to make the `batch_size` parameter take effect so that I can...

> What "data generator" do you mean? The one in TensorFlow or PyTorch? `data generator` 是`keras` 的说法,现在统一叫`DataLoader`,因为CV和NLP中一次性加载所有数据集几乎不可能,默认`data generator` 。 现在DeepXDE是这种一次性加载所有数据集的方式: ```python # Load entire dataset X, y = np.load('data.npy') #...

> The error is due to `nan`. You may check why you get `nan`. hi,lulu,可以在 `dde.saveplot()` 中加入 `plt.figure(figsize=figsize)` 中的 figsize 参数吗, 像这样`dde.saveplot(losshistory, train_state, issave=True, isplot=True, figsize=(10,8))` 去使用它。这样方便在 jupyter notebook 中使用...

> If we have enough time, we will try to migrate. However, I hope that Baidu official can release an open source model of text to image on paddlepaddle. I...

Another project with code and models - ERNIE-SAT 类别文心·跨模态大模型 应用语音编辑、语音生成、语音克隆、带语音克隆的语音到语音翻译 ERNIE-SAT 采用语音-文本联合训练的方式在中文和英文数据集上进行预训练。使得模型学到了语音和文本的对齐关系,并且生成频谱的精度更高,合成声音的质量更高。 https://wenxin.baidu.com/wenxin/modelbasedetail/ernie_sat/