Spaces:
Running
Running
Org card: three-layer framing (embedding, perception, memory); add Engram + live site
Browse files
README.md
CHANGED
|
@@ -9,11 +9,11 @@ pinned: false
|
|
| 9 |
|
| 10 |

|
| 11 |
|
| 12 |
-
Eximius Labs builds open-weight multimodal models that run on your own hardware: the
|
| 13 |
|
| 14 |
We freeze a strong base model and add new senses from the outside, so existing embeddings stay valid and gaining a modality never forces a re-index. One shared vector space spans text, image, video, audio, thermal, and motion (IMU).
|
| 15 |
|
| 16 |
-
## Fusion Embedding, the
|
| 17 |
|
| 18 |
One vector space for cross-modal retrieval, RAG, clustering, and semantic search.
|
| 19 |
|
|
@@ -27,8 +27,19 @@ One vector space for cross-modal retrieval, RAG, clustering, and semantic search
|
|
| 27 |
|
| 28 |
Dense scene understanding and geometric place recognition on a frozen vision backbone, projected into the Fusion Embedding space. Code: https://github.com/Eximius-Labs/fusion-perception.
|
| 29 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 30 |
## Build with it
|
| 31 |
|
|
|
|
| 32 |
- Cookbook: https://github.com/Eximius-Labs/cookbook
|
| 33 |
- Code: https://github.com/Eximius-Labs
|
| 34 |
- Paper: https://arxiv.org/abs/2607.18666
|
|
|
|
| 9 |
|
| 10 |

|
| 11 |
|
| 12 |
+
Eximius Labs builds open-weight multimodal models that run on your own hardware: the embedding, perception, and memory layers for physical AI. The technical report is on arXiv: https://arxiv.org/abs/2607.18666. Try a real robot's memory in the live playground: https://www.eximiuslabs.com/playground.
|
| 13 |
|
| 14 |
We freeze a strong base model and add new senses from the outside, so existing embeddings stay valid and gaining a modality never forces a re-index. One shared vector space spans text, image, video, audio, thermal, and motion (IMU).
|
| 15 |
|
| 16 |
+
## Fusion Embedding, the embedding layer
|
| 17 |
|
| 18 |
One vector space for cross-modal retrieval, RAG, clustering, and semantic search.
|
| 19 |
|
|
|
|
| 27 |
|
| 28 |
Dense scene understanding and geometric place recognition on a frozen vision backbone, projected into the Fusion Embedding space. Code: https://github.com/Eximius-Labs/fusion-perception.
|
| 29 |
|
| 30 |
+
## Engram, the memory layer
|
| 31 |
+
|
| 32 |
+
The open cross-modal memory layer for physical AI, built on Fusion Embedding. It indexes a robot's video, audio, and motion into one embedding space on a shared clock and answers questions about it in plain language, including temporal reasoning that retrieval alone cannot do.
|
| 33 |
+
|
| 34 |
+
```
|
| 35 |
+
pip install engram-robomem
|
| 36 |
+
```
|
| 37 |
+
|
| 38 |
+
Code: https://github.com/Eximius-Labs/engram 路 PyPI: https://pypi.org/project/engram-robomem 路 Playground: https://www.eximiuslabs.com/playground
|
| 39 |
+
|
| 40 |
## Build with it
|
| 41 |
|
| 42 |
+
- Site: https://www.eximiuslabs.com
|
| 43 |
- Cookbook: https://github.com/Eximius-Labs/cookbook
|
| 44 |
- Code: https://github.com/Eximius-Labs
|
| 45 |
- Paper: https://arxiv.org/abs/2607.18666
|