Try the Qwen-Image 2.1 doodle-in LoRA live on Hugging Face ZeroGPU
Gradio · x · 2026-09-29
Gradio launched a live demo of Qwen-Image-2.1-doodle-in-LoRA on Hugging Face ZeroGPU: sketch an outline on a photo, name the object, and it appears in the photo's style.
The project is fully open-sourced — diffusers and ComfyUI adapters, 6,042 training pairs, all eval outputs, and training/eval scripts — under the non-commercial Qwen Research License.
Related event: Gradio Open-Sources Qwen-Image 2.1 Doodle-In LoRA(3 posts)→
More from Multimodal
- Pruned CTC Cuts ASR Training Memory 5.1x, Enabling Native-LLM-Vocab Speech Recognition — X-LANCE · 2026-09-29
- SCOPD Distillation Keeps 92% of VLM Performance at 10% Visual Tokens — uoft · 2026-09-29
- AnswerMap: Training-Free Black-Box Spatial Rationale Hits 0.85 AUC vs 0.38 for Attention — Mohamed Eltahir · 2026-09-29
- ElevenLabs launches Eleven v4 Turbo in ElevenAgents with ~100ms median latency — lukeharries · 2026-09-29
- Open Edit: open-source agent turns videos (incl. Opus 5.5 outputs) into editable browser projects — Delicious_Baby_3392 · 2026-09-29
- Local MiniMax H3 fight-scene testing on RTX 5090: HyperFlow wins, plus a double-patching trap — glusphere · 2026-09-29