Robotics
Transformers
Safetensors
dm05
text-generation
robot-control
vision-language-action
vla
dm0.5
opendm
Instructions to use Dexmal/DM05 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use Dexmal/DM05 with Transformers:
# Load model directly from transformers import AutoModelForSeq2SeqLM model = AutoModelForSeq2SeqLM.from_pretrained("Dexmal/DM05", dtype="auto") - Notebooks
- Google Colab
- Kaggle
update README
Browse files- README.md +40 -1
- norm_stats.json +136 -0
README.md
CHANGED
|
@@ -1,3 +1,42 @@
|
|
| 1 |
---
|
| 2 |
-
license:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
+
license: mit
|
| 3 |
+
library_name: transformers
|
| 4 |
+
tags:
|
| 5 |
+
- robotics
|
| 6 |
+
- robot-control
|
| 7 |
+
- vision-language-action
|
| 8 |
+
- vla
|
| 9 |
+
- dm05
|
| 10 |
+
- dm0.5
|
| 11 |
+
- opendm
|
| 12 |
---
|
| 13 |
+
|
| 14 |
+
# DM05
|
| 15 |
+
|
| 16 |
+
[](https://github.com/dexmal/opendm) [](https://huggingface.co/Dexmal) [](https://www.dexmal.com/blog/dm0.5/index_en.html)
|
| 17 |
+
|
| 18 |
+
DM05 is the base checkpoint of DM0.5, Dexmal's open-world
|
| 19 |
+
Vision-Language-Action foundation model for embodied intelligence. It uses a
|
| 20 |
+
Gemma3 4B vision-language backbone with a 680M Action Expert to generate
|
| 21 |
+
continuous robot actions, and is designed for natural-language manipulation,
|
| 22 |
+
zero-shot generalization, efficient downstream fine-tuning, long-horizon
|
| 23 |
+
historical context, robust policy behavior, and transfer across robot
|
| 24 |
+
embodiments.
|
| 25 |
+
|
| 26 |
+
## How to Use
|
| 27 |
+
|
| 28 |
+
This model is intended to be used with OpenDM. See the
|
| 29 |
+
[OpenDM README](https://github.com/dexmal/opendm) for installation,
|
| 30 |
+
checkpoint download, inference, fine-tuning, and evaluation commands.
|
| 31 |
+
|
| 32 |
+
## Citation
|
| 33 |
+
|
| 34 |
+
```bibtex
|
| 35 |
+
@misc{dm05,
|
| 36 |
+
title = {{DM0.5}: An Open-World Foundation Model for General-Purpose Embodied Intelligence},
|
| 37 |
+
author = {{Dexmal Team}},
|
| 38 |
+
month = {July},
|
| 39 |
+
year = {2026},
|
| 40 |
+
url = {https://www.dexmal.com/blog/dm0.5/index_en.html}
|
| 41 |
+
}
|
| 42 |
+
```
|
norm_stats.json
ADDED
|
@@ -0,0 +1,136 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"norm_stats": {
|
| 3 |
+
"action": {
|
| 4 |
+
"q01": [
|
| 5 |
+
-0.672650857591629,
|
| 6 |
+
-1.131368241930008,
|
| 7 |
+
-1.0881663817405702,
|
| 8 |
+
-0.9142281610488894,
|
| 9 |
+
-1.0504351261138918,
|
| 10 |
+
-0.8239678527832028,
|
| 11 |
+
-0.0002905806455761194,
|
| 12 |
+
-0.6260258240699768,
|
| 13 |
+
-1.0805172435760497,
|
| 14 |
+
-1.1288272675514222,
|
| 15 |
+
-1.038402559375763,
|
| 16 |
+
-1.0376791499614715,
|
| 17 |
+
-0.841348724555969,
|
| 18 |
+
-0.0003305471634026617
|
| 19 |
+
],
|
| 20 |
+
"q99": [
|
| 21 |
+
0.6728938806533815,
|
| 22 |
+
1.1696967702388763,
|
| 23 |
+
1.1283396813392637,
|
| 24 |
+
0.9106882274627681,
|
| 25 |
+
0.9512754203796385,
|
| 26 |
+
0.7984103721618654,
|
| 27 |
+
0.0705579190496006,
|
| 28 |
+
0.7021891847133639,
|
| 29 |
+
1.2401762067794797,
|
| 30 |
+
1.122687215900421,
|
| 31 |
+
0.9586501039505002,
|
| 32 |
+
1.1267098831653595,
|
| 33 |
+
0.8427464372634894,
|
| 34 |
+
0.07060984028726816
|
| 35 |
+
],
|
| 36 |
+
"mean": [
|
| 37 |
+
-0.00261860090540722,
|
| 38 |
+
0.005378620466217399,
|
| 39 |
+
-0.006516445428133011,
|
| 40 |
+
0.00071163714164868,
|
| 41 |
+
0.005850601242855191,
|
| 42 |
+
0.00016266797319985926,
|
| 43 |
+
0.011993872933089733,
|
| 44 |
+
0.002125306404195726,
|
| 45 |
+
0.004999904544092715,
|
| 46 |
+
-0.006896577542647719,
|
| 47 |
+
0.00016130938092828728,
|
| 48 |
+
-0.0071029057726264,
|
| 49 |
+
-0.0004772193351527676,
|
| 50 |
+
0.012536692433059216
|
| 51 |
+
],
|
| 52 |
+
"std": [
|
| 53 |
+
0.1697254404425621,
|
| 54 |
+
0.30742591619491577,
|
| 55 |
+
0.2893647998571396,
|
| 56 |
+
0.22407980263233185,
|
| 57 |
+
0.24888065457344055,
|
| 58 |
+
0.21918441355228424,
|
| 59 |
+
0.021581493318080902,
|
| 60 |
+
0.18644987791776657,
|
| 61 |
+
0.3061055839061737,
|
| 62 |
+
0.2909375876188278,
|
| 63 |
+
0.24517737329006195,
|
| 64 |
+
0.2714238241314888,
|
| 65 |
+
0.2321367785334587,
|
| 66 |
+
0.02121920231729746
|
| 67 |
+
]
|
| 68 |
+
},
|
| 69 |
+
"state": {
|
| 70 |
+
"q01": [
|
| 71 |
+
-1.3522899922132492,
|
| 72 |
+
-2.1652386215180157,
|
| 73 |
+
-0.08790020607709885,
|
| 74 |
+
-1.090997260761261,
|
| 75 |
+
-1.7576654033660888,
|
| 76 |
+
-2.9801979890823365,
|
| 77 |
+
-0.0002905806455761194,
|
| 78 |
+
-1.2919814363002777,
|
| 79 |
+
-2.1020697339057923,
|
| 80 |
+
-0.07839894692748785,
|
| 81 |
+
-2.9396027744293214,
|
| 82 |
+
-0.8631593664884568,
|
| 83 |
+
-1.1610263860702514,
|
| 84 |
+
-0.00034549142029136423
|
| 85 |
+
],
|
| 86 |
+
"q99": [
|
| 87 |
+
1.3684158971548084,
|
| 88 |
+
0.17555581547319887,
|
| 89 |
+
2.1748491822361946,
|
| 90 |
+
2.910172110319138,
|
| 91 |
+
1.0934281175613405,
|
| 92 |
+
0.7161070912361147,
|
| 93 |
+
0.07054356536138803,
|
| 94 |
+
1.4658485317230223,
|
| 95 |
+
0.12374066639840642,
|
| 96 |
+
2.0903009934291243,
|
| 97 |
+
1.7147838599681853,
|
| 98 |
+
1.6415026142120364,
|
| 99 |
+
2.8933019072532655,
|
| 100 |
+
0.07059489603037945
|
| 101 |
+
],
|
| 102 |
+
"mean": [
|
| 103 |
+
-0.476219080388546,
|
| 104 |
+
-0.6785732209682465,
|
| 105 |
+
0.5610137432813644,
|
| 106 |
+
1.37710839509964,
|
| 107 |
+
-0.3851655125617981,
|
| 108 |
+
-1.4545539617538452,
|
| 109 |
+
0.012111684307456017,
|
| 110 |
+
0.43913714587688446,
|
| 111 |
+
-0.7300353348255157,
|
| 112 |
+
0.6235602796077728,
|
| 113 |
+
-1.3437719941139221,
|
| 114 |
+
0.4893051087856293,
|
| 115 |
+
1.401452362537384,
|
| 116 |
+
0.012644956354051828
|
| 117 |
+
],
|
| 118 |
+
"std": [
|
| 119 |
+
0.49391762912273407,
|
| 120 |
+
0.6390611827373505,
|
| 121 |
+
0.6070653796195984,
|
| 122 |
+
0.6070861220359802,
|
| 123 |
+
0.601529985666275,
|
| 124 |
+
0.5551068037748337,
|
| 125 |
+
0.02163003385066986,
|
| 126 |
+
0.4867953509092331,
|
| 127 |
+
0.623783141374588,
|
| 128 |
+
0.6057631075382233,
|
| 129 |
+
0.6544217765331268,
|
| 130 |
+
0.6064487993717194,
|
| 131 |
+
0.6204572021961212,
|
| 132 |
+
0.021267084404826164
|
| 133 |
+
]
|
| 134 |
+
}
|
| 135 |
+
}
|
| 136 |
+
}
|