depth-pro
depthpro-riemann-spring-eval
depthpro-riemann-spring-eval
Resultados de avaliação (v2) do experimento de estimativa de profundidade com regularização Riemanniana aplicada ao fine-tuning do DepthPro, avaliado no domínio Spring.
Não contém dados de treino nem checkpoints — apenas os artefatos de avaliação: figuras qualitativas e tabelas de métrica.
O que foi rodado
Fine-tuning das cabeças do DepthPro com termos de regularização geométrica, comparado contra o modelo zero-shot (sem fine-tune).… See the full description on the dataset page: https://huggingface.co/datasets/AKCITPixel3/depthpro-riemann-spring-eval.srt-depth-probe-artifacts
SRT depth-probe artifacts
Where a frozen multimodal backbone carries cross-modal meaning depends on
what you use to look. A raw-cosine probe and a fitted linear probe rank the
layers of the same model differently, on the same states, in the same run.
Every number here is measured on frozen google/gemma-4-31B-it. Nothing in
this repository is a trained product. These are measurement artifacts and the
scripts that produced them, published so the ranking can be checked instead
of… See the full description on the dataset page: https://huggingface.co/datasets/RiverRider/srt-depth-probe-artifacts.GrapesNet_single_cluster_depth
Grapesnet Single Cluster Depth
This dataset provides real RGB-D imagery of grape clusters in a mixed agricultural environment. Captured using a tripod-mounted smartphone and depth camera system, it offers synchronized color and depth data for natural vineyard conditions to support crop detection research. The dataset contains 696 images with no classification, segmentation, or bounding-box annotations.
This dataset is indexed on https://project-agml.github.io/ as part of the… See the full description on the dataset page: https://huggingface.co/datasets/Project-AgML/GrapesNet_single_cluster_depth.GrapesNet_depth
Grapesnet Depth
This dataset comprises real-world RGB-D imagery captured in a mixed vineyard environment for grape crop detection. Images were collected using tripod-mounted smartphone and depth camera systems, providing synchronized color and depth data from Sonaka grapevine locations in Yelavi, Maharashtra, India. The dataset contains 847 images with no classification, segmentation, or bounding-box annotations.
This dataset is indexed on https://project-agml.github.io/ as part… See the full description on the dataset page: https://huggingface.co/datasets/Project-AgML/GrapesNet_depth.so101_pick_up_cylinder_depth_20260713_171230This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 30,
"features": {
"action": {
"dtype": "float32",
"names": [
"shoulder_pan.pos",
"shoulder_lift.pos",
"elbow_flex.pos",
"wrist_flex.pos",
"wrist_roll.pos",
"gripper.pos"
],
"shape": [
6… See the full description on the dataset page: https://huggingface.co/datasets/mot-prog/so101_pick_up_cylinder_depth_20260713_171230.so101_pick_up_cylinder_depth_20260713_171230_reencoded_reencodedThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 30,
"features": {
"action": {
"dtype": "float32",
"names": [
"shoulder_pan.pos",
"shoulder_lift.pos",
"elbow_flex.pos",
"wrist_flex.pos",
"wrist_roll.pos",
"gripper.pos"
],
"shape": [
6… See the full description on the dataset page: https://huggingface.co/datasets/mot-prog/so101_pick_up_cylinder_depth_20260713_171230_reencoded_reencoded.