Live Human Feedback in the Training Loop: Aligning Diffusion Models with Real People
• 11
AI & ML interests
RLHF, Model Evaluation, Benchmarks, Data Labeling, Human Feedback, Computer Vision, Image Generation, Video Generation, LLMs, Translations
Recent Activity
View all activity
Articles
Organization Card
Fast human data labeling at scale — rapidata.ai for questions or data inquiries just write to us: info@rapidata.ai
Human judgements of LLM output: translation quality across 10+ languages and multilingual humour.
-
Rapidata/Translation-gpt4o_mini-v-gpt4o-v-deepl
Viewer • Updated • 373 • 97 • 16 -
Rapidata/Translation-deepseek-llama-mixtral-v-deepl
Viewer • Updated • 845 • 138 • 16 -
Rapidata/multilingual-llm-jokes-4o-claude-gemini
Viewer • Updated • 9.98k • 195 • 15 -
Rapidata/vlm-vs-crowd
Viewer • Updated • 325 • 133
Human judgements of LLM output: translation quality across 10+ languages and multilingual humour.
-
Rapidata/Translation-gpt4o_mini-v-gpt4o-v-deepl
Viewer • Updated • 373 • 97 • 16 -
Rapidata/Translation-deepseek-llama-mixtral-v-deepl
Viewer • Updated • 845 • 138 • 16 -
Rapidata/multilingual-llm-jokes-4o-claude-gemini
Viewer • Updated • 9.98k • 195 • 15 -
Rapidata/vlm-vs-crowd
Viewer • Updated • 325 • 133
Text-to-audio / TTS preference judgements (naturalness, friendliness) and audio association studies.
spaces 9
Running
9
Text to Image Benchmark
🖼
Text to Image model benchmark, ranked by humans
Running
10
SVG Generation Benchmark
🔺
SVG generation benchmark, ranked by humans
Running
7
Relative Camera Movement Benchmark
🧭
World models relative camera benchmark, ranked by humans
Running
7
Video Realism Benchmark
📷
World models realism benchmark, ranked by humans
Running
10
Video Physics Benchmark
🧪
World models physics benchmark, ranked by humans
Running
9
Image to 3D Benchmark
🧊
Image to 3D model benchmark, ranked by humans
models 0
None public yet
datasets 81
Rapidata/image-to-3d-single-object-human-feedback-benchmark
Viewer • Updated • 49.3k • 648 • 3
Rapidata/xAI_Aurora_t2i_human_preferences
Viewer • Updated • 26.2k • 214 • 15
Rapidata/vlm-vs-crowd
Viewer • Updated • 325 • 133
Rapidata/text-to-svg-human-feedback-benchmark
Viewer • Updated • 518k • 5.06k • 39
Rapidata/text-to-3d-with-llm-code
Viewer • Updated • 44.5k • 5.06k • 47
Rapidata/text-2-video-human-preferences
Viewer • Updated • 2.57k • 2.33k • 22
Rapidata/text-2-video-human-preferences-wan2.1
Viewer • Updated • 787 • 564 • 21
Rapidata/text-2-video-human-preferences-veo3
Viewer • Updated • 1.02k • 1.28k • 21
Rapidata/text-2-video-human-preferences-veo3.1
Updated • 976 • 9
Rapidata/text-2-video-human-preferences-veo2
Viewer • Updated • 760 • 431 • 15