Depth Estimation
Transformers
Safetensors
qwen3_vl
image-text-to-text
vision-language-model
3d-vision
multimodal
qwen3-vl
Instructions to use JonnyYu828/DepthVLM-4B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use JonnyYu828/DepthVLM-4B with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("depth-estimation", model="JonnyYu828/DepthVLM-4B")# Load model directly from transformers import AutoProcessor, AutoModelForImageTextToText processor = AutoProcessor.from_pretrained("JonnyYu828/DepthVLM-4B") model = AutoModelForImageTextToText.from_pretrained("JonnyYu828/DepthVLM-4B") - Notebooks
- Google Colab
- Kaggle
Update README.md
Browse files
README.md
CHANGED
|
@@ -15,4 +15,22 @@ tags:
|
|
| 15 |
|
| 16 |
paper:
|
| 17 |
- arxiv: 2605.15876
|
| 18 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 15 |
|
| 16 |
paper:
|
| 17 |
- arxiv: 2605.15876
|
| 18 |
+
---
|
| 19 |
+
|
| 20 |
+
Update 2026-05-18 (v1.0): Initial release
|
| 21 |
+
|
| 22 |
+
# DepthVLM
|
| 23 |
+
|
| 24 |
+
DepthVLM serves as a unified foundation model for both low-level dense geometry prediction and high-level multimodal understanding, while achieving substantially faster inference compared with existing VLM-based approaches such as DepthLM and Youtu-VL.
|
| 25 |
+
|
| 26 |
+
## Highlights
|
| 27 |
+
|
| 28 |
+
- Native dense metric depth estimation in VLMs
|
| 29 |
+
- Unified multimodal understanding and geometry prediction
|
| 30 |
+
- Full-resolution depth prediction with efficient inference
|
| 31 |
+
- Supports both indoor and outdoor metric depth estimation
|
| 32 |
+
- Improved 3D spatial reasoning capability
|
| 33 |
+
|
| 34 |
+
## Paper
|
| 35 |
+
|
| 36 |
+
[Unlocking Dense Metric Depth Estimation in VLMs](https://arxiv.org/abs/2605.15876)
|