xcharleslin

Loss of log probability precision in PageRank implementation

1

The reference implementation of PageRank throws away accuracy when calculating missing mass (RunPageRankBasic.java:456) by bringing the log-probability back into linear space to compute the missing mass. Obviously, this introduces error...

### Polars version checks - [X] I have checked that this issue has not already been reported. - [X] I have confirmed this bug exists on the [latest version](https://pypi.org/project/polars/) of...

Negative scale for Decimal and Decimal256

1

Analogous to https://github.com/apache/arrow-rs/issues/1785. Currently, Arrow2 Decimal and Decimal256 have [scale typed as `usize`](https://github.com/jorgecarleitao/arrow2/blob/8ee5ad8c774e7355218376caf73d94bce2a769a6/src/datatypes/mod.rs#L156), which would not support negative scale. As a driveby, precision is also currently typed as `usize` but...

Minimum repro: ```python import torch from min_dalle import MinDalle from concurrent.futures import ThreadPoolExecutor USE_GPU = True def f(text: str, root: str): return MinDalle( models_root=f'./{root}', dtype=torch.float32, device="cuda", is_mega=False, is_reusable=True, ).generate_image( text,...