Open-vocabulary detectors compiled for SiMa.ai Modalix: type what to find, get boxes. Runs on the MLA.
🔄 In a Training Loop
Timothy Do
dotimothy
AI & ML interests
LLMs/VLMs/, Robotics
Recent Activity
updated a collection 1 day ago
SiMa.ai Open-Vocabulary Detection Models updated a model 1 day ago
dotimothy/sima-grounding-dino published a model 1 day ago
dotimothy/sima-grounding-dinoOrganizations
LLiMa LLM models
Language models compiled for SiMa.ai Modalix with the LLiMa runtime. Run them with github.com/dotimothy/neat-genai-studio
SiMa.ai Decision Models
System-1 decision models compiled for SiMa.ai Modalix: one forward pass picks an option, scores, or answers yes/no.
SiMa.ai Change Detection Models
Change detection compiled for SiMa.ai Modalix: CYWS (trained) and zero-shot DINOv2. Two photos in, boxes around what changed out.
LLiMa VLM models
Vision-language models compiled for SiMa.ai Modalix with the LLiMa runtime. Run them with github.com/dotimothy/neat-genai-studio
SO-101 Datasets
Pick-and-place demonstrations on an SO-101 arm, LeRobot v3.0 format: wrist and overhead cameras, 30 fps.
-
dotimothy/BlackMarkerComplete_20260729_142533
SO-101 • Updated • 22 episodes • 50 -
dotimothy/red_cube_20260730_141623
SO-101 • Updated • 30 episodes • 68 -
dotimothy/red_cube_30_20260730_142924
SO-101 • Updated • 46 episodes • 60 -
dotimothy/red_cube_76_20260730_144844
SO-101 • Updated • 77 episodes • 75
SiMa.ai Open-Vocabulary Detection Models
Open-vocabulary detectors compiled for SiMa.ai Modalix: type what to find, get boxes. Runs on the MLA.
SiMa.ai Change Detection Models
Change detection compiled for SiMa.ai Modalix: CYWS (trained) and zero-shot DINOv2. Two photos in, boxes around what changed out.
LLiMa LLM models
Language models compiled for SiMa.ai Modalix with the LLiMa runtime. Run them with github.com/dotimothy/neat-genai-studio
LLiMa VLM models
Vision-language models compiled for SiMa.ai Modalix with the LLiMa runtime. Run them with github.com/dotimothy/neat-genai-studio
SiMa.ai Decision Models
System-1 decision models compiled for SiMa.ai Modalix: one forward pass picks an option, scores, or answers yes/no.
SO-101 Datasets
Pick-and-place demonstrations on an SO-101 arm, LeRobot v3.0 format: wrist and overhead cameras, 30 fps.
-
dotimothy/BlackMarkerComplete_20260729_142533
SO-101 • Updated • 22 episodes • 50 -
dotimothy/red_cube_20260730_141623
SO-101 • Updated • 30 episodes • 68 -
dotimothy/red_cube_30_20260730_142924
SO-101 • Updated • 46 episodes • 60 -
dotimothy/red_cube_76_20260730_144844
SO-101 • Updated • 77 episodes • 75