Yiming Lin

Update prompt pure to 1.9.0

2

Please could you update Pure to version 1.9.0, which fixed a bug showing conda environment https://github.com/sindresorhus/pure/pull/440. Thank you very much.

Why the pool_size is set to (20, 12) in the SPHead?

3

Hi, Thank you for open-sourcing the amazing work. In the `SPHead` module, [the pool_size parameter er is set to 20, 12.](https://github.com/Andrew-Qibin/SPNet/blob/ac3a6429509dffedd18e0acef3d6ead3ad3d9991/models/spnet.py#L53). Can I ask why these two values? If the...

Hi @Eskender-B , Thanks for your work. Have you managed to reproduce the results (say F1 score) on the helen or lfw-pl dataset?

[resolved] When will a new official release version be out and how to build on Mac M1?

19

Hi @ahrm Thank you for the project. The latest release was from 2022 but there have been many commits. When do you expect to create a new official release so...

How to train SphereFace2 on >1m identities, e.g. WebFace24M?

1

Hi @wy1iu @ydwen Thank you for open-sourcing this repo. Section 2.4 of the Sphereface2 paper says > the gradient computations in SphereFace2 are class-independent and can be performed locally within...

How to check `is_last_batch` in torchtnt==0.1.0?

2

### 🐛 Describe the bug In the `train_step` method of a Callback class: ```python def train_step(self, state: State, data: TrainBatch) -> None: global_step = state.train_state.progress.num_steps_completed is_last_batch = state.train_state.is_last_batch # error...

About mask2 in cifp_margin

4

Hello, why in CIFP margin did you consider only `neg_logits` that are smaller than `pos_logits - self.m`? Why those larger`neg_logits` should not be penalised? Thanks. https://github.com/fuenwang/MixFairFace/blob/11691d330a3cde2e1c4eeb76211ec6d5b7752547/network.py#L29-L30