Huaijun Jiang

3.6节train_ch3函数loss计算和sgd问题

1

**bug描述** 3.6节train_ch3函数，如果传入的loss已经是求过平均的，train_l_sum每次只累加一个batch的平均值，最后却除以总样本数，打印的loss结果就会很小，例如3.9节的调用。 如果每个batch不一样大（例如Fashion-MNIST设置batch_size=256时，最后一个batch是96），当optimizer=None时，默认的sgd传入batch_size应该会在最后一个batch造成误差，似乎应该使用y.shape[0]。 另外train_ch5是用batch_count，如果batch大小不一致，最后打印的loss也应该会有微小误差。 ```python def train_ch3(net, train_iter, test_iter, loss, num_epochs, batch_size, params=None, lr=None, optimizer=None): for epoch in range(num_epochs): train_l_sum, train_acc_sum, n = 0.0, 0.0, 0 for X, y in...

`Configuration.get()` always return None if key is not in CS regardless of the provided default value

1

`Configuration.get()` always returns None if key is not in CS regardless of the provided default value. I have to use `config.get_dictionary.get()` to perform the expected behavior. ConfigSpace version: 0.4.20, 0.4.21,...

scikit-optimize version: 0.8.0, 0.8.1, 0.9.0 For all samplers in `skopt/sampler/`, includes `Lhs, Sobol, Halton, Hammersly, Grid`, the return type of `sampler.generate()` is `list`, but should it be `numpy.ndarray` according to...

An implementation to support complex conditions such as 'a < b and a * b < 10 and ...'

3

I implement a `ConditionedConfigurationSpace` that supports complex conditions between hyperparameters (e.g., x1 = 100: return False return True # return config['x1']

Support Markdown links on MyST

9

I love the features you provide. But I'm writing Sphinx docs with markdown (myst-parser). For example, I often write a link to other documents in my docs (markdown): ```markdown [Another...

Support sphinx-design tabs

3

`hoverxref_sphinxtabs = True` seems not work with [sphinx-design tabs](https://sphinx-design.readthedocs.io/en/furo-theme/tabs.html) ([github](https://github.com/executablebooks/sphinx-design)). The reference page in my docs is written in markdown. Can it be supported?

icons not shown correctly in rst toctree

3

### Describe the bug **context** When I write icons in rst toctree: ```rst .. toctree:: :caption: Table of content :maxdepth: 2 :titlesonly: Quick Start :octicon:`rocket;1em` ``` **expectation** I expected icon...

DEPRECATION: pyrfr is being installed using the legacy 'setup.py install' method

1

pyrfr version: 0.8.3 I tried to `pip install pyrfr` on GitHub actions' ["ubuntu-latest", "macos-latest", "windows-latest"]. On macos and windows, I got the deprecation warning: ``` DEPRECATION: pyrfr is being installed...