Mr.Li

resource.setrlimit(resource.RLIMIT_NOFILE, (20480, rlimit[1])) 20480 is too larger to run it.when I change it to 4096,I can run it.

About the pre-trained model

14

Thank you for this work. I want to use the pre-trained model of AttentionNet-IRSE-56/92 from the MODEL_ZOO.md for fine-tuning. Where can I get the pre-trained model?

About reduction ratios R

4

Thanks for your working! I have some questions: 1. That reduction ratios R=8 is the best trade-off in the paper , but the code defaults to 4, so why? 2....

What is the difference between these two documents?

有关loss的问题

6

@xialuxi 非常感谢你的工作！同时我存在一些疑问： 1、在[https://github.com/TinyZeaMays/CircleLoss/blob/master/circle_loss.py](url) 的第37、39行与本文有所差异，可以解释一下原因吗? `dim=1 or dim=0` 和 `loss or loss.mean()` 2、我试着用你的代码跑人证识别，75人的200多万照片，class_num 为75万。参数设置为 margin=0.25 gamma=256 or 128，batch_size=256, 训练过程中损失非常大，并且无法收敛。麻烦问一下可能是什么原因呢？

[DOC]: 官方示例报错

3

### 📚 The doc issue 官方文档的示例代码报错，原因是文档过旧吗？ ``` File "pp_plus.py", line 9, in from colossalai.builder import build_pipeline_model ImportError: cannot import name 'build_pipeline_model' from 'colossalai.builder' (/opt/conda/lib/python3.8/site-packages/colossalai/builder/__init__.py) ``` 

[DOC]: 公式乱码

3

### 📚 The doc issue 当前官方最新文档的公式乱码、格式混乱。对阅读造成很大的困扰，期待修复~ 

Why is the API document removed from the official website page

1

### Discussed in https://github.com/hpcaitech/ColossalAI/discussions/2937 Originally posted by **bobo0810** February 28, 2023 Why is the API document removed from the official website page？

数据并行数 × epoch数 = 真实的epoch数?

1

### Discussed in https://github.com/hpcaitech/ColossalAI/discussions/2961 Originally posted by **bobo0810** March 1, 2023 每个gpu上的dataloader都是完整的数据集，未做拆分。 即epoch=3 gpu=2时仅数据并行，模型实际上过了6遍数据集。 ``` engine, train_dataloader, val_dataloader, _ = colossalai.initialize( model, optimizer, criterion, train_dataloader, val_dataloader, ) ```

[BUG]: ZeRO无法使用预训练权重

3

### 🐛 Describe the bug 示例代码 https://github.com/hpcaitech/ColossalAI/blob/v0.2.5/tests/test_zero/test_zero_engine.py resnet18(num_classes=10)✅ resnet18( num_classes=10,pretrained=True)❌ ```python File "/opt/conda/lib/python3.8/site-packages/torchvision/models/resnet.py", line 309, in resnet18 return _resnet("resnet18", BasicBlock, [2, 2, 2, 2], pretrained, progress, **kwargs) File "/opt/conda/lib/python3.8/site-packages/torchvision/models/resnet.py", line...