Sierkinhane (Jinheng Xie)

Thank you, Siro. One more question: Do we need to modify the attention interface to use Ring-Attention when we import a built-in Transformer implementation like Qwen2ForCausalLM from the Transformers package, or does Accelerate’s maybe_context_parallel handle this?

updated a dataset 2 months ago

Sierkinhane/Storyboard20K

Updated Sep 24 • 18

commented on Accelerate ND-Parallel: A guide to Efficient Multi-GPU Training 3 months ago

Does this pipeline work for a custom model?

updated a model 3 months ago

Sierkinhane/show-o2-1.5B-test

3B • Updated Sep 8 • 2

published a model 3 months ago

Sierkinhane/show-o2-1.5B-test

3B • Updated Sep 8 • 2

updated 5 models 3 months ago

published 2 models 3 months ago

showlab/show-o2-7B-w-video-und

Any-to-Any • Updated Sep 5 • 32 • 2

showlab/show-o2-1.5B-w-video-und

Any-to-Any • Updated Sep 5 • 7

Jinheng Xie

AI & ML interests

Recent Activity

Organizations

Glance: Accelerating Diffusion Models with 1 Sample

The Image as Its Own Reward: Reinforcement Learning with Adversarial Reward for Image Generation

Computer-Use Agents as Judges for Generative User Interface

Sierkinhane/test

Sierkinhane/test

VCode: a Multimodal Coding Benchmark with SVG as Symbolic Visual Representation

Sierkinhane/ucf101

Sierkinhane/ucf101

Sierkinhane/Storyboard20K

Sierkinhane/show-o2-1.5B-test

Sierkinhane/show-o2-1.5B-test

showlab/show-o2-1.5B-HQ

showlab/show-o2-1.5B

showlab/show-o2-1.5B-w-video-und

showlab/show-o2-7B-w-video-und

showlab/show-o2-7B

showlab/show-o2-7B-w-video-und

showlab/show-o2-1.5B-w-video-und

Jinheng Xie

AI & ML interests

Recent Activity

Organizations

Sierkinhane's activity