soldni commited on
Commit
46fb535
1 Parent(s): 46afbb1

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +50 -0
README.md CHANGED
@@ -127,6 +127,56 @@ print(generated_text)
127
 
128
  *Benchmarks: AI2D test, ChartQA test, VQA v2.0 test, DocQA test, InfographicVQA test, TextVQA val, RealWorldQA, MMMU val, MathVista testmini, CountBenchQA, Flickr Count (we collected this new dataset that is significantly harder than CountBenchQA).*
129
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
130
  ## License and Use
131
 
132
  This model is licensed under Apache 2.0. It is intended for research and educational use.
 
127
 
128
  *Benchmarks: AI2D test, ChartQA test, VQA v2.0 test, DocQA test, InfographicVQA test, TextVQA val, RealWorldQA, MMMU val, MathVista testmini, CountBenchQA, Flickr Count (we collected this new dataset that is significantly harder than CountBenchQA).*
129
 
130
+
131
+ ### I'm getting an error a broadcast error when processing images!
132
+
133
+ Your image might not be in RGB format. You can convert it using the following code snippet:
134
+
135
+ ```python
136
+ from PIL import Image
137
+
138
+ image = Image.open(...)
139
+
140
+ if image.mode != "RGB":
141
+ image = image.convert("RGB")
142
+ ```
143
+
144
+ ### Molmo doesn't work great with transparent images!
145
+
146
+ We received reports that Molmo models might struggle with transparent images.
147
+ For the time being, we recommend adding a white or dark background to your images before passing them to the model. The code snippet below shows how to do this using the Python Imaging Library (PIL):
148
+
149
+ ```python
150
+
151
+ # Load the image
152
+ url = "..."
153
+ image = Image.open(requests.get(url, stream=True).raw)
154
+
155
+ # Convert the image to grayscale to calculate brightness
156
+ gray_image = image.convert('L') # Convert to grayscale
157
+
158
+ # Calculate the average brightness
159
+ stat = ImageStat.Stat(gray_image)
160
+ average_brightness = stat.mean[0] # Get the average value
161
+
162
+ # Define background color based on brightness (threshold can be adjusted)
163
+ bg_color = (0, 0, 0) if average_brightness > 127 else (255, 255, 255)
164
+
165
+ # Create a new image with the same size as the original, filled with the background color
166
+ new_image = Image.new('RGB', image.size, bg_color)
167
+
168
+ # Paste the original image on top of the background (use image as a mask if needed)
169
+ new_image.paste(image, (0, 0), image if image.mode == 'RGBA' else None)
170
+
171
+ # Now you can pass the new_image to Molmo
172
+ processor = AutoProcessor.from_pretrained(
173
+ 'allenai/Molmo-7B-D-0924',
174
+ trust_remote_code=True,
175
+ torch_dtype='auto',
176
+ device_map='auto'
177
+ )
178
+ ```
179
+
180
  ## License and Use
181
 
182
  This model is licensed under Apache 2.0. It is intended for research and educational use.