binmakeswell

[Colossal-AI](https://github.com/hpcaitech/ColossalAI) is a unified deep learning system for the big model era, which integrates many efficient techniques like multi-dimensional tensor parallelism, sequence parallelism, heterogeneous memory management, etc. By using Colossal-AI,...

LAMB: Differences from the paper author's official implementation

3

The LAMB implementation of the PyTorch version you released is different from the official version of TensorFlow released by the paper author. According to the official implementation published in the...

[BUG]: stuck at initialization and no error message

4

### 🐛 Describe the bug When parallel is set to pipeline=4 and tensor=dict(mode='2d', size=4), the program will get stuck on initialization and no error message will be output. ### Environment...

[FEATURE]: example - Diffusion Model

1

### Describe the feature e.g. Latent Diffusion

[FEATURE]: example - stylegan_xl

4

### Describe the feature styleganXL是支持stylegan3,stylegan2ada的通用训练架构，代码也做了简化，用这个做案例会很棒，他家代码刚刚发布几周

### 📚 The doc issue A step-by-step tutorial, eg. model.

### Describe the feature Update/maintenance/usage is not uniform, and the same named model file without description, may cause confusion to users, affecting user experience and development efficiency.

### Describe the feature https://github.com/hpcaitech/ColossalAI-Examples/issues/76

[FEATURE]: require convenient hybrid_parallel example

1

### Describe the feature Most users do not have 64 GPUs to try these config, maybe use less resource? https://github.com/hpcaitech/ColossalAI-Examples/tree/main/image/vision_transformer/hybrid_parallel

### 📚 The doc issue Incomplete: Define your own parallel model