Feature Extraction
Transformers
Safetensors
English
remote-sensing
earth-observation
vision
dofa
sentinel-2
multimodal
Instructions to use BiliSakura/DOFA-transformers with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use BiliSakura/DOFA-transformers with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("feature-extraction", model="BiliSakura/DOFA-transformers")# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("BiliSakura/DOFA-transformers", dtype="auto") - Notebooks
- Google Colab
- Kaggle
Sync updated code and configs (no weight re-upload)
Browse files- README.md +20 -16
- dofa-base-patch16-224/__pycache__/modeling_dofa.cpython-311.pyc +0 -0
- dofa-base-patch16-224/__pycache__/pipeline_dofa.cpython-311.pyc +0 -0
- dofa-base-patch16-224/image_processing_dofa.py +1 -1
- dofa-base-patch16-224/preprocessor_config.json +1 -1
- dofa-large-patch16-224/__pycache__/pipeline_dofa.cpython-311.pyc +0 -0
- dofa-large-patch16-224/image_processing_dofa.py +1 -1
- dofa-large-patch16-224/preprocessor_config.json +1 -1
README.md
CHANGED
|
@@ -1,19 +1,3 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: mit
|
| 3 |
-
language:
|
| 4 |
-
- en
|
| 5 |
-
tags:
|
| 6 |
-
- remote-sensing
|
| 7 |
-
- earth-observation
|
| 8 |
-
- vision
|
| 9 |
-
- feature-extraction
|
| 10 |
-
- dofa
|
| 11 |
-
- sentinel-2
|
| 12 |
-
- multimodal
|
| 13 |
-
library_name: transformers
|
| 14 |
-
pipeline_tag: feature-extraction
|
| 15 |
-
---
|
| 16 |
-
|
| 17 |
# DOFA Transformers Models
|
| 18 |
|
| 19 |
Self-contained HuggingFace model checkpoints for [DOFA](https://arxiv.org/abs/2403.15356).
|
|
@@ -31,6 +15,8 @@ Sentinel-2 9-band defaults (`default_wavelengths`, `default_image_mean`, `defaul
|
|
| 31 |
|
| 32 |
## Usage
|
| 33 |
|
|
|
|
|
|
|
| 34 |
```python
|
| 35 |
from transformers import pipeline
|
| 36 |
|
|
@@ -42,9 +28,27 @@ pipe = pipeline(
|
|
| 42 |
trust_remote_code=True,
|
| 43 |
)
|
| 44 |
|
|
|
|
| 45 |
features = pipe(image_array, pool=True, return_tensors=True)
|
| 46 |
```
|
| 47 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 48 |
Override Sentinel-2 defaults for other sensors:
|
| 49 |
|
| 50 |
```python
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
# DOFA Transformers Models
|
| 2 |
|
| 3 |
Self-contained HuggingFace model checkpoints for [DOFA](https://arxiv.org/abs/2403.15356).
|
|
|
|
| 15 |
|
| 16 |
## Usage
|
| 17 |
|
| 18 |
+
Processors default to **`do_resize: false`**. Pass Sentinel-2 stacks at native `(H, W, C)`; the processor rescales values (typically `/255`) without changing spatial size.
|
| 19 |
+
|
| 20 |
```python
|
| 21 |
from transformers import pipeline
|
| 22 |
|
|
|
|
| 28 |
trust_remote_code=True,
|
| 29 |
)
|
| 30 |
|
| 31 |
+
# Native-resolution patch, e.g. 512×512×9 bands (uint8 or float)
|
| 32 |
features = pipe(image_array, pool=True, return_tensors=True)
|
| 33 |
```
|
| 34 |
|
| 35 |
+
Dense features:
|
| 36 |
+
|
| 37 |
+
```python
|
| 38 |
+
tokens = pipe(image_array, pool=False, return_tensors=True)
|
| 39 |
+
```
|
| 40 |
+
|
| 41 |
+
Opt in to 224×224 resize (original pretraining size):
|
| 42 |
+
|
| 43 |
+
```python
|
| 44 |
+
features = pipe(
|
| 45 |
+
image_array,
|
| 46 |
+
pool=True,
|
| 47 |
+
return_tensors=True,
|
| 48 |
+
image_processor_kwargs={"do_resize": True},
|
| 49 |
+
)
|
| 50 |
+
```
|
| 51 |
+
|
| 52 |
Override Sentinel-2 defaults for other sensors:
|
| 53 |
|
| 54 |
```python
|
dofa-base-patch16-224/__pycache__/modeling_dofa.cpython-311.pyc
ADDED
|
Binary file (21.1 kB). View file
|
|
|
dofa-base-patch16-224/__pycache__/pipeline_dofa.cpython-311.pyc
ADDED
|
Binary file (2.76 kB). View file
|
|
|
dofa-base-patch16-224/image_processing_dofa.py
CHANGED
|
@@ -51,7 +51,7 @@ class DOFAImageProcessor(BaseImageProcessor):
|
|
| 51 |
|
| 52 |
def __init__(
|
| 53 |
self,
|
| 54 |
-
do_resize: bool =
|
| 55 |
size: Optional[dict[str, int]] = None,
|
| 56 |
resample: PILImageResampling = PILImageResampling.BILINEAR,
|
| 57 |
do_rescale: bool = True,
|
|
|
|
| 51 |
|
| 52 |
def __init__(
|
| 53 |
self,
|
| 54 |
+
do_resize: bool = False,
|
| 55 |
size: Optional[dict[str, int]] = None,
|
| 56 |
resample: PILImageResampling = PILImageResampling.BILINEAR,
|
| 57 |
do_rescale: bool = True,
|
dofa-base-patch16-224/preprocessor_config.json
CHANGED
|
@@ -4,7 +4,7 @@
|
|
| 4 |
"height": 224,
|
| 5 |
"width": 224
|
| 6 |
},
|
| 7 |
-
"do_resize":
|
| 8 |
"do_rescale": true,
|
| 9 |
"do_normalize": true,
|
| 10 |
"do_convert_rgb": false,
|
|
|
|
| 4 |
"height": 224,
|
| 5 |
"width": 224
|
| 6 |
},
|
| 7 |
+
"do_resize": false,
|
| 8 |
"do_rescale": true,
|
| 9 |
"do_normalize": true,
|
| 10 |
"do_convert_rgb": false,
|
dofa-large-patch16-224/__pycache__/pipeline_dofa.cpython-311.pyc
ADDED
|
Binary file (2.76 kB). View file
|
|
|
dofa-large-patch16-224/image_processing_dofa.py
CHANGED
|
@@ -51,7 +51,7 @@ class DOFAImageProcessor(BaseImageProcessor):
|
|
| 51 |
|
| 52 |
def __init__(
|
| 53 |
self,
|
| 54 |
-
do_resize: bool =
|
| 55 |
size: Optional[dict[str, int]] = None,
|
| 56 |
resample: PILImageResampling = PILImageResampling.BILINEAR,
|
| 57 |
do_rescale: bool = True,
|
|
|
|
| 51 |
|
| 52 |
def __init__(
|
| 53 |
self,
|
| 54 |
+
do_resize: bool = False,
|
| 55 |
size: Optional[dict[str, int]] = None,
|
| 56 |
resample: PILImageResampling = PILImageResampling.BILINEAR,
|
| 57 |
do_rescale: bool = True,
|
dofa-large-patch16-224/preprocessor_config.json
CHANGED
|
@@ -4,7 +4,7 @@
|
|
| 4 |
"height": 224,
|
| 5 |
"width": 224
|
| 6 |
},
|
| 7 |
-
"do_resize":
|
| 8 |
"do_rescale": true,
|
| 9 |
"do_normalize": true,
|
| 10 |
"do_convert_rgb": false,
|
|
|
|
| 4 |
"height": 224,
|
| 5 |
"width": 224
|
| 6 |
},
|
| 7 |
+
"do_resize": false,
|
| 8 |
"do_rescale": true,
|
| 9 |
"do_normalize": true,
|
| 10 |
"do_convert_rgb": false,
|