Anbei

icon indicating a website link https://www.linkedin.com/in/anbei-zhao-12234326/ icon indicating an email [email protected]

icon company AWS icon location Solutions Architect in AWS

Results 2 issues of Anbei

[BUG]: run in Docker can not find parameter `_forward_module.first_stage_model.encoder.conv_in.weight` in the GeminiDDP module

2 comment

### 🐛 Describe the bug Hi I am using docker to run ColossalAI/examples/images/diffusion/ and tried https://hub.docker.com/r/hpcaitech/colossalai/tags 0.2.5, 0.2.4 0.2.2 0.1.10-torch1.11-cu11.3 with train_colossalai_teyvat.yaml and 512-base-ema.ckpt. However, there is an error: Epoch...

bug

multi-host parallel training cannot recognize all gpu

1 comment

Hi I am following https://github.com/hpcaitech/ColossalAI/tree/main/examples/images/diffusion and trying to use 2 hosts with each 1 T4 GPU. I changed lightning: trainer: accelerator: 'gpu' devices: 2 And had error: lightning.fabric.utilities.exceptions.MisconfigurationException: You requested...