Friedrich Marty

Smorty100

https://gitlab.com/users/Marty_Friedrich/projects

AI & ML interests

I'm most interested in content rerouting between LLM and VLLM agens for automation possibilities. Using templates for each agent which is then filled in by another agents inputs seems really useful.

Recent Activity

upvoted a paper 5 days ago

ToonComposer: Streamlining Cartoon Production with Generative Post-Keyframing

upvoted a paper 5 days ago

NextStep-1: Toward Autoregressive Image Generation with Continuous Tokens at Scale

new activity 10 days ago

Qwen/Qwen-Image:It's so unsafe that it can generate inappropriate adult content easily

View all activity

Organizations

None yet

upvoted 2 papers 5 days ago

ToonComposer: Streamlining Cartoon Production with Generative Post-Keyframing

Paper • 2508.10881 • Published 8 days ago • 49

NextStep-1: Toward Autoregressive Image Generation with Continuous Tokens at Scale

Paper • 2508.10711 • Published 8 days ago • 134

New activity in Qwen/Qwen-Image 10 days ago

It's so unsafe that it can generate inappropriate adult content easily

#21 opened 18 days ago by

liougehooa

liked 2 models 11 days ago

zai-org/GLM-4.5V

Image-Text-to-Text • 108B • Updated 5 days ago • 22.4k • • 578

zai-org/GLM-4.5

Text Generation • 358B • Updated 11 days ago • 44k • • 1.24k

upvoted a paper 12 days ago

On the Generalization of SFT: A Reinforcement Learning Perspective with Reward Rectification

Paper • 2508.05629 • Published 15 days ago • 154

New activity in Qwen/Qwen3-30B-A3B-Instruct-2507 13 days ago

Add 1M support

#12 opened 15 days ago by

hzhwcmhf

upvoted a paper 14 days ago

Nexus-Gen: A Unified Model for Image Understanding, Generation, and Editing

Paper • 2504.21356 • Published Apr 30 • 1

liked a Space 18 days ago

706

Qwen Image

🖼

Generate images from text prompts

reacted to chintankp's post with 😔 21 days ago

Post

1593

We’re excited to share that Llama Nemotron Super v1.5 -- our latest open reasoning model -- is leading the Artificial Analysis Intelligence Index - a leaderboard that spans advanced math, science, and agentic tasks, for models running on a single NVIDIA H100.
Super v1.5 is trained with high-quality reasoning synthetic data generated from models like Qwen3-235B and DeepSeek R1. Besides leading accuracy, it also delivers high throughput.
Key features:
- Leading accuracy on multi-step reasoning, math, coding, and function-calling
- Post-trained using RPO, DPO, and RLVR across 26M+ synthetic examples
- Fully transparent training data on HF (Nemotron-Post-Training-Dataset-v1)

Try Super v1.5 on build.nvidia.com or download from Hugging Face