Update README.md
Browse files
README.md
CHANGED
@@ -41,7 +41,7 @@ For the SDO team: this model is the first version for demonstration purposes. It
|
|
41 |
We will include more technical details here soon.
|
42 |
|
43 |
## Example Images
|
44 |
-
|
45 |
|
46 |
#### High_Energy_Ion_Fe_Nuclei
|
47 |
|
@@ -58,26 +58,75 @@ The ViT model was pretrained on a dataset consisting of 14 million images and 21
|
|
58 |
More information on the base model used can be found here: (https://huggingface.co/google/vit-base-patch16-224-in21k)
|
59 |
|
60 |
## How to use this Model
|
61 |
-
(quick
|
62 |
|
|
|
63 |
```python
|
64 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
65 |
from transformers import AutoFeatureExtractor, AutoModelForImageClassification
|
66 |
from PIL import Image
|
67 |
-
import requests
|
68 |
|
69 |
-
|
70 |
-
image = Image.open(
|
|
|
71 |
|
|
|
|
|
|
|
|
|
72 |
feature_extractor = AutoFeatureExtractor.from_pretrained("kenobi/NASA_GeneLab_MBT")
|
73 |
model = AutoModelForImageClassification.from_pretrained("kenobi/NASA_GeneLab_MBT")
|
74 |
-
inputs = feature_extractor(images=image, return_tensors="pt")
|
75 |
|
|
|
|
|
|
|
|
|
76 |
outputs = model(**inputs)
|
77 |
logits = outputs.logits
|
78 |
-
|
|
|
79 |
predicted_class_idx = logits.argmax(-1).item()
|
80 |
-
|
|
|
|
|
81 |
```
|
82 |
|
83 |
## BibTeX & References
|
|
|
41 |
We will include more technical details here soon.
|
42 |
|
43 |
## Example Images
|
44 |
+
>>> Use one of the images below for the inference API field on the upper right.
|
45 |
|
46 |
#### High_Energy_Ion_Fe_Nuclei
|
47 |
|
|
|
58 |
More information on the base model used can be found here: (https://huggingface.co/google/vit-base-patch16-224-in21k)
|
59 |
|
60 |
## How to use this Model
|
61 |
+
(quick snippets to work on Google Colab)
|
62 |
|
63 |
+
First a snippet to downnload test images from an online repository:
|
64 |
```python
|
65 |
+
import urllib.request
|
66 |
+
|
67 |
+
def download_image(url, filename):
|
68 |
+
try:
|
69 |
+
# Define custom headers
|
70 |
+
headers = {
|
71 |
+
'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/58.0.3029.110 Safari/537.3'
|
72 |
+
}
|
73 |
+
|
74 |
+
# Create a request with custom headers
|
75 |
+
req = urllib.request.Request(url, headers=headers)
|
76 |
+
|
77 |
+
# Open the URL and read the content
|
78 |
+
with urllib.request.urlopen(req) as response:
|
79 |
+
img_data = response.read()
|
80 |
+
|
81 |
+
# Write the content to a file
|
82 |
+
with open(filename, 'wb') as handler:
|
83 |
+
handler.write(img_data)
|
84 |
+
|
85 |
+
print(f"Image '{filename}' downloaded successfully")
|
86 |
+
except Exception as e:
|
87 |
+
print(f"Error downloading the image '{filename}':", e)
|
88 |
+
|
89 |
+
# List of URLs and corresponding filenames
|
90 |
+
urls = [
|
91 |
+
('https://roosevelt.devron-systems.com/HF/P242_73665006707-A6_002_008_proj.tif', 'P242_73665006707-A6_002_008_proj.tif'),
|
92 |
+
('https://roosevelt.devron-systems.com/HF/P278_73668090728-A7_003_027_proj.tif', 'P278_73668090728-A7_003_027_proj.tif')
|
93 |
+
]
|
94 |
+
|
95 |
+
# Download each image
|
96 |
+
for url, filename in urls:
|
97 |
+
download_image(url, filename)
|
98 |
+
```
|
99 |
+
|
100 |
+
Then use the images for inference:
|
101 |
+
|
102 |
+
```python
|
103 |
+
#!pip install transformers --quiet # uncomment this pip install for local use if you do not have transformers installed
|
104 |
from transformers import AutoFeatureExtractor, AutoModelForImageClassification
|
105 |
from PIL import Image
|
|
|
106 |
|
107 |
+
# Load the image
|
108 |
+
#image = Image.open('P242_73665006707-A6_002_008_proj.tif') #First Image
|
109 |
+
image = Image.open('P278_73668090728-A7_003_027_proj.tif') #Second Image
|
110 |
|
111 |
+
# Convert grayscale image to RGB
|
112 |
+
image_rgb = image.convert("RGB")
|
113 |
+
|
114 |
+
# Load the pre-trained feature extractor and classification model
|
115 |
feature_extractor = AutoFeatureExtractor.from_pretrained("kenobi/NASA_GeneLab_MBT")
|
116 |
model = AutoModelForImageClassification.from_pretrained("kenobi/NASA_GeneLab_MBT")
|
|
|
117 |
|
118 |
+
# Extract features from the image
|
119 |
+
inputs = feature_extractor(images=image_rgb, return_tensors="pt")
|
120 |
+
|
121 |
+
# Perform classification
|
122 |
outputs = model(**inputs)
|
123 |
logits = outputs.logits
|
124 |
+
|
125 |
+
# Obtain the predicted class index and label
|
126 |
predicted_class_idx = logits.argmax(-1).item()
|
127 |
+
predicted_class_label = model.config.id2label[predicted_class_idx]
|
128 |
+
|
129 |
+
print("Predicted class:", predicted_class_label)
|
130 |
```
|
131 |
|
132 |
## BibTeX & References
|