update readme download instructions
Browse files
README.md
CHANGED
|
@@ -20,7 +20,47 @@ pipeline_tag: robotics
|
|
| 20 |
📄 Paper: https://arxiv.org/abs/2602.10556
|
| 21 |
💻 Code: https://github.com/lihzha/lap
|
| 22 |
|
| 23 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 24 |
|
| 25 |
## Model Summary
|
| 26 |
|
|
@@ -53,7 +93,6 @@ Supported manipulation tasks include:
|
|
| 53 |
- container placement
|
| 54 |
- towel manipulation
|
| 55 |
|
| 56 |
-
---
|
| 57 |
|
| 58 |
## Limitations
|
| 59 |
|
|
@@ -67,7 +106,6 @@ Future work includes extending LAP to more complex embodiments and tasks, includ
|
|
| 67 |
- bimanual robots
|
| 68 |
- dexterous hands
|
| 69 |
- mobile manipulation systems
|
| 70 |
-
---
|
| 71 |
|
| 72 |
## Intended Use
|
| 73 |
|
|
@@ -80,7 +118,6 @@ LAP-3B is intended for:
|
|
| 80 |
|
| 81 |
The model is **not intended for safety-critical deployments without additional validation**.
|
| 82 |
|
| 83 |
-
---
|
| 84 |
|
| 85 |
## Citation
|
| 86 |
|
|
|
|
| 20 |
📄 Paper: https://arxiv.org/abs/2602.10556
|
| 21 |
💻 Code: https://github.com/lihzha/lap
|
| 22 |
|
| 23 |
+
|
| 24 |
+
## Download
|
| 25 |
+
|
| 26 |
+
You can download the LAP checkpoint directly from the Hugging Face Hub.
|
| 27 |
+
|
| 28 |
+
### Using the Hugging Face CLI (recommended)
|
| 29 |
+
|
| 30 |
+
Install the Hugging Face Hub CLI:
|
| 31 |
+
|
| 32 |
+
```bash
|
| 33 |
+
pip install -U huggingface_hub
|
| 34 |
+
````
|
| 35 |
+
|
| 36 |
+
Download the checkpoint to the expected directory:
|
| 37 |
+
|
| 38 |
+
```bash
|
| 39 |
+
hf download lihzha/LAP-3B --local-dir ./checkpoint/lap
|
| 40 |
+
```
|
| 41 |
+
|
| 42 |
+
After downloading, the checkpoint will be located at:
|
| 43 |
+
|
| 44 |
+
```
|
| 45 |
+
./checkpoint/lap
|
| 46 |
+
```
|
| 47 |
+
|
| 48 |
+
This matches the default path expected by the LAP codebase.
|
| 49 |
+
|
| 50 |
+
### Alternative: Python API
|
| 51 |
+
|
| 52 |
+
You can also download the checkpoint programmatically:
|
| 53 |
+
|
| 54 |
+
```python
|
| 55 |
+
from huggingface_hub import snapshot_download
|
| 56 |
+
|
| 57 |
+
snapshot_download(
|
| 58 |
+
repo_id="lihzha/LAP-3B",
|
| 59 |
+
local_dir="./checkpoint/lap"
|
| 60 |
+
)
|
| 61 |
+
```
|
| 62 |
+
|
| 63 |
+
|
| 64 |
|
| 65 |
## Model Summary
|
| 66 |
|
|
|
|
| 93 |
- container placement
|
| 94 |
- towel manipulation
|
| 95 |
|
|
|
|
| 96 |
|
| 97 |
## Limitations
|
| 98 |
|
|
|
|
| 106 |
- bimanual robots
|
| 107 |
- dexterous hands
|
| 108 |
- mobile manipulation systems
|
|
|
|
| 109 |
|
| 110 |
## Intended Use
|
| 111 |
|
|
|
|
| 118 |
|
| 119 |
The model is **not intended for safety-critical deployments without additional validation**.
|
| 120 |
|
|
|
|
| 121 |
|
| 122 |
## Citation
|
| 123 |
|