Spaces:
Running
on
Zero
Running
on
Zero
new app.py
Browse files
app.py
CHANGED
@@ -28,21 +28,20 @@ if not logger.handlers:
|
|
28 |
handler = logging.StreamHandler()
|
29 |
handler.setFormatter(logging.Formatter("%(asctime)s [%(levelname)s] %(name)s: %(message)s"))
|
30 |
logger.addHandler(handler)
|
31 |
-
logger.warning("here")
|
32 |
|
33 |
# Define repository and local directory
|
34 |
repo_id = "microsoft/OmniParser-v2.0" # HF repo
|
35 |
local_dir = "weights" # Target local directory
|
|
|
|
|
36 |
|
37 |
som_generator = MarkHelper()
|
38 |
magma_som_prompt = "<image>\nIn this view I need to click a button to \"{}\"? Provide the coordinates and the mark index of the containing bounding box if applicable."
|
39 |
magma_qa_prompt = "<image>\n{} Answer the question briefly."
|
40 |
magma_model_id = "microsoft/Magma-8B"
|
41 |
-
magam_model = AutoModelForCausalLM.from_pretrained(magma_model_id, trust_remote_code=True, torch_dtype=
|
42 |
-
magma_processor = AutoProcessor.from_pretrained(magma_model_id, trust_remote_code=True
|
43 |
-
magam_model.to(
|
44 |
-
|
45 |
-
logger.warning(f"The repository is downloading to: {local_dir}")
|
46 |
|
47 |
# Download the entire repository
|
48 |
snapshot_download(repo_id=repo_id, local_dir=local_dir)
|
@@ -58,27 +57,14 @@ MARKDOWN = """
|
|
58 |
<div align="center">
|
59 |
<h2>Magma: A Foundation Model for Multimodal AI Agents</h2>
|
60 |
|
61 |
-
[Jianwei Yang](https://jwyang.github.io/)<sup>*</sup><sup>1</sup><sup>†</sup>
|
62 |
-
[Reuben Tan](https://cs-people.bu.edu/rxtan/)<sup>1</sup><sup>†</sup>
|
63 |
-
[Qianhui Wu](https://qianhuiwu.github.io/)<sup>1</sup><sup>†</sup>
|
64 |
-
[Ruijie Zheng](https://ruijiezheng.com/)<sup>2</sup><sup>‡</sup>
|
65 |
-
[Baolin Peng](https://scholar.google.com/citations?user=u1CNjgwAAAAJ&hl=en&oi=ao)<sup>1</sup><sup>‡</sup>
|
66 |
-
[Yongyuan Liang](https://cheryyunl.github.io)<sup>2</sup><sup>‡</sup>
|
67 |
-
[Yu Gu](https://users.umiacs.umd.edu/~hal/)<sup>1</sup>
|
68 |
-
[Mu Cai](https://pages.cs.wisc.edu/~mucai/)<sup>3</sup>
|
69 |
-
[Seonghyeon Ye](https://seonghyeonye.github.io/)<sup>4</sup>
|
70 |
-
[Joel Jang](https://joeljang.github.io/)<sup>5</sup>
|
71 |
-
[Yuquan Deng](https://scholar.google.com/citations?user=LTC0Q6YAAAAJ&hl=en)<sup>5</sup>
|
72 |
-
[Lars Liden](https://sites.google.com/site/larsliden)<sup>1</sup>
|
73 |
-
[Jianfeng Gao](https://www.microsoft.com/en-us/research/people/jfgao/)<sup>1</sup><sup>▽</sup>
|
74 |
-
|
75 |
-
<sup>1</sup> Microsoft Research; <sup>2</sup> University of Maryland; <sup>3</sup> University of Wisconsin-Madison; <sup>4</sup> KAIST; <sup>5</sup> University of Washington
|
76 |
-
|
77 |
-
<sup>*</sup> Project lead <sup>†</sup> First authors <sup>‡</sup> Second authors <sup>▽</sup> Leadership
|
78 |
-
|
79 |
\[[arXiv Paper](https://www.arxiv.org/pdf/2502.13130)\] \[[Project Page](https://microsoft.github.io/Magma/)\] \[[Github Repo](https://github.com/microsoft/Magma)\] \[[Hugging Face Model](https://huggingface.co/microsoft/Magma-8B)\]
|
80 |
|
81 |
-
This demo is powered by [Gradio](https://gradio.app/) and uses OmniParserv2 to generate Set-of-Mark prompts.
|
|
|
|
|
|
|
|
|
|
|
82 |
</div>
|
83 |
"""
|
84 |
|
@@ -86,7 +72,6 @@ DEVICE = torch.device('cuda')
|
|
86 |
|
87 |
@spaces.GPU
|
88 |
@torch.inference_mode()
|
89 |
-
@torch.autocast(device_type="cuda", dtype=torch.bfloat16)
|
90 |
def get_som_response(instruction, image_som):
|
91 |
prompt = magma_som_prompt.format(instruction)
|
92 |
if magam_model.config.mm_use_image_start_end:
|
@@ -101,23 +86,10 @@ def get_som_response(instruction, image_som):
|
|
101 |
add_generation_prompt=True
|
102 |
)
|
103 |
|
104 |
-
# inputs = magma_processor(images=[image_som], texts=prompt, return_tensors="pt")
|
105 |
-
# # with torch.autocast(device_type="cuda", dtype=torch.bfloat16):
|
106 |
-
# # inputs['pixel_values'] = inputs['pixel_values'].unsqueeze(0).to(torch.bfloat16) # Add .to(torch.bfloat16) here for explicit casting
|
107 |
-
# # inputs['image_sizes'] = inputs['image_sizes'].unsqueeze(0)
|
108 |
-
# # logger.warning(inputs['pixel_values'].dtype)
|
109 |
-
# # # inputs = inputs.to("cuda")
|
110 |
-
# inputs = inputs.to("cuda", dtype=torch.bfloat16)
|
111 |
-
|
112 |
inputs = magma_processor(images=[image_som], texts=prompt, return_tensors="pt")
|
113 |
-
inputs['pixel_values'] = inputs['pixel_values'].
|
114 |
inputs['image_sizes'] = inputs['image_sizes'].unsqueeze(0)
|
115 |
-
inputs
|
116 |
-
|
117 |
-
# 处理其他可能的输入
|
118 |
-
for key in inputs:
|
119 |
-
if key not in ['pixel_values', 'image_sizes'] and torch.is_tensor(inputs[key]):
|
120 |
-
inputs[key] = inputs[key].to("cuda")
|
121 |
|
122 |
magam_model.generation_config.pad_token_id = magma_processor.tokenizer.pad_token_id
|
123 |
with torch.inference_mode():
|
@@ -137,7 +109,6 @@ def get_som_response(instruction, image_som):
|
|
137 |
|
138 |
@spaces.GPU
|
139 |
@torch.inference_mode()
|
140 |
-
@torch.autocast(device_type="cuda", dtype=torch.bfloat16)
|
141 |
def get_qa_response(instruction, image):
|
142 |
prompt = magma_qa_prompt.format(instruction)
|
143 |
if magam_model.config.mm_use_image_start_end:
|
@@ -152,12 +123,10 @@ def get_qa_response(instruction, image):
|
|
152 |
add_generation_prompt=True
|
153 |
)
|
154 |
|
155 |
-
|
156 |
-
|
157 |
-
|
158 |
-
|
159 |
-
# inputs = inputs.to("cuda")
|
160 |
-
inputs = inputs.to("cuda", dtype=torch.bfloat16)
|
161 |
|
162 |
magam_model.generation_config.pad_token_id = magma_processor.tokenizer.pad_token_id
|
163 |
with torch.inference_mode():
|
@@ -177,7 +146,7 @@ def get_qa_response(instruction, image):
|
|
177 |
|
178 |
@spaces.GPU
|
179 |
@torch.inference_mode()
|
180 |
-
@torch.autocast(device_type="cuda", dtype=torch.bfloat16)
|
181 |
def process(
|
182 |
image_input,
|
183 |
box_threshold,
|
@@ -187,107 +156,99 @@ def process(
|
|
187 |
instruction,
|
188 |
) -> Optional[Image.Image]:
|
189 |
|
190 |
-
|
191 |
-
|
192 |
-
|
193 |
-
|
194 |
-
|
195 |
-
|
196 |
-
|
197 |
-
|
198 |
-
|
199 |
-
|
200 |
-
|
201 |
-
|
202 |
-
|
203 |
-
|
204 |
-
|
205 |
-
|
206 |
-
|
207 |
-
|
208 |
-
|
209 |
-
|
210 |
-
|
211 |
-
|
212 |
-
|
213 |
-
|
214 |
-
|
215 |
-
|
216 |
-
|
217 |
-
|
218 |
-
|
219 |
-
|
220 |
-
|
221 |
-
|
222 |
-
|
223 |
-
|
224 |
-
|
225 |
-
|
226 |
-
|
227 |
-
|
228 |
-
|
229 |
-
|
230 |
-
|
231 |
-
|
232 |
-
|
233 |
-
|
234 |
-
|
235 |
-
|
236 |
-
|
237 |
-
|
238 |
-
mark_id
|
239 |
-
|
240 |
-
if str(mark_id) in label_coordinates:
|
241 |
-
bbox_for_mark = label_coordinates[str(mark_id)]
|
242 |
-
else:
|
243 |
-
bbox_for_mark = None
|
244 |
else:
|
245 |
bbox_for_mark = None
|
246 |
-
|
247 |
-
|
248 |
-
|
249 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
250 |
image_input,
|
251 |
-
[
|
252 |
-
som_generator,
|
253 |
edgecolor=(255,127,111),
|
254 |
-
|
255 |
-
fn_save=None,
|
256 |
normalized_to_pixel=False,
|
257 |
add_mark=False
|
258 |
)
|
259 |
-
|
260 |
-
|
261 |
-
|
262 |
-
|
263 |
-
|
264 |
-
click_point = [item / 1000 for item in click_point]
|
265 |
-
else:
|
266 |
-
click_point = pred_2_point(magma_response)
|
267 |
-
# de-normalize click_point (width, height)
|
268 |
-
click_point = [click_point[0] * image_input.size[0], click_point[1] * image_input.size[1]]
|
269 |
-
|
270 |
-
image_som = plot_circles_with_marks(
|
271 |
-
image_input,
|
272 |
-
[click_point],
|
273 |
-
som_generator,
|
274 |
-
edgecolor=(255,127,111),
|
275 |
-
linewidth=3,
|
276 |
-
fn_save=None,
|
277 |
-
normalized_to_pixel=False,
|
278 |
-
add_mark=False
|
279 |
-
)
|
280 |
-
except:
|
281 |
-
image_som = image_input
|
282 |
-
|
283 |
-
logger.warning("finish processing")
|
284 |
-
return image_som, str(parsed_content_list)
|
285 |
-
except Exception as e:
|
286 |
-
error_message = traceback.format_exc()
|
287 |
-
logger.warning(error_message)
|
288 |
-
return image_input, error_message
|
289 |
-
|
290 |
-
logger.warning("Starting App.")
|
291 |
with gr.Blocks() as demo:
|
292 |
gr.Markdown(MARKDOWN)
|
293 |
with gr.Row():
|
@@ -326,6 +287,6 @@ with gr.Blocks() as demo:
|
|
326 |
outputs=[image_output_component, text_output_component]
|
327 |
)
|
328 |
|
329 |
-
# demo.launch(debug=
|
330 |
# demo.launch(share=True, server_port=7861, server_name='0.0.0.0')
|
331 |
-
demo.queue().launch(share=False)
|
|
|
28 |
handler = logging.StreamHandler()
|
29 |
handler.setFormatter(logging.Formatter("%(asctime)s [%(levelname)s] %(name)s: %(message)s"))
|
30 |
logger.addHandler(handler)
|
|
|
31 |
|
32 |
# Define repository and local directory
|
33 |
repo_id = "microsoft/OmniParser-v2.0" # HF repo
|
34 |
local_dir = "weights" # Target local directory
|
35 |
+
dtype = torch.bfloat16
|
36 |
+
DEVICE = torch.device('cuda')
|
37 |
|
38 |
som_generator = MarkHelper()
|
39 |
magma_som_prompt = "<image>\nIn this view I need to click a button to \"{}\"? Provide the coordinates and the mark index of the containing bounding box if applicable."
|
40 |
magma_qa_prompt = "<image>\n{} Answer the question briefly."
|
41 |
magma_model_id = "microsoft/Magma-8B"
|
42 |
+
magam_model = AutoModelForCausalLM.from_pretrained(magma_model_id, trust_remote_code=True, torch_dtype=dtype)
|
43 |
+
magma_processor = AutoProcessor.from_pretrained(magma_model_id, trust_remote_code=True)
|
44 |
+
magam_model.to(DEVICE)
|
|
|
|
|
45 |
|
46 |
# Download the entire repository
|
47 |
snapshot_download(repo_id=repo_id, local_dir=local_dir)
|
|
|
57 |
<div align="center">
|
58 |
<h2>Magma: A Foundation Model for Multimodal AI Agents</h2>
|
59 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
60 |
\[[arXiv Paper](https://www.arxiv.org/pdf/2502.13130)\] \[[Project Page](https://microsoft.github.io/Magma/)\] \[[Github Repo](https://github.com/microsoft/Magma)\] \[[Hugging Face Model](https://huggingface.co/microsoft/Magma-8B)\]
|
61 |
|
62 |
+
This demo is powered by [Gradio](https://gradio.app/) and uses [OmniParserv2](https://github.com/microsoft/OmniParser) to generate [Set-of-Mark prompts](https://github.com/microsoft/SoM).
|
63 |
+
|
64 |
+
The demo supports three modes:
|
65 |
+
1. Empty text inut: it downgrades to an OmniParser demo.
|
66 |
+
2. Text input starting with "Q:": it leads to a visual question answering demo.
|
67 |
+
3. Text input for UI navigation: it leads to a UI navigation demo.
|
68 |
</div>
|
69 |
"""
|
70 |
|
|
|
72 |
|
73 |
@spaces.GPU
|
74 |
@torch.inference_mode()
|
|
|
75 |
def get_som_response(instruction, image_som):
|
76 |
prompt = magma_som_prompt.format(instruction)
|
77 |
if magam_model.config.mm_use_image_start_end:
|
|
|
86 |
add_generation_prompt=True
|
87 |
)
|
88 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
89 |
inputs = magma_processor(images=[image_som], texts=prompt, return_tensors="pt")
|
90 |
+
inputs['pixel_values'] = inputs['pixel_values'].unsqueeze(0)
|
91 |
inputs['image_sizes'] = inputs['image_sizes'].unsqueeze(0)
|
92 |
+
inputs = inputs.to(dtype).to(DEVICE)
|
|
|
|
|
|
|
|
|
|
|
93 |
|
94 |
magam_model.generation_config.pad_token_id = magma_processor.tokenizer.pad_token_id
|
95 |
with torch.inference_mode():
|
|
|
109 |
|
110 |
@spaces.GPU
|
111 |
@torch.inference_mode()
|
|
|
112 |
def get_qa_response(instruction, image):
|
113 |
prompt = magma_qa_prompt.format(instruction)
|
114 |
if magam_model.config.mm_use_image_start_end:
|
|
|
123 |
add_generation_prompt=True
|
124 |
)
|
125 |
|
126 |
+
inputs = magma_processor(images=[image], texts=prompt, return_tensors="pt")
|
127 |
+
inputs['pixel_values'] = inputs['pixel_values'].unsqueeze(0)
|
128 |
+
inputs['image_sizes'] = inputs['image_sizes'].unsqueeze(0)
|
129 |
+
inputs = inputs.to(dtype).to(DEVICE)
|
|
|
|
|
130 |
|
131 |
magam_model.generation_config.pad_token_id = magma_processor.tokenizer.pad_token_id
|
132 |
with torch.inference_mode():
|
|
|
146 |
|
147 |
@spaces.GPU
|
148 |
@torch.inference_mode()
|
149 |
+
# @torch.autocast(device_type="cuda", dtype=torch.bfloat16)
|
150 |
def process(
|
151 |
image_input,
|
152 |
box_threshold,
|
|
|
156 |
instruction,
|
157 |
) -> Optional[Image.Image]:
|
158 |
|
159 |
+
# image_save_path = 'imgs/saved_image_demo.png'
|
160 |
+
# image_input.save(image_save_path)
|
161 |
+
# image = Image.open(image_save_path)
|
162 |
+
box_overlay_ratio = image_input.size[0] / 3200
|
163 |
+
draw_bbox_config = {
|
164 |
+
'text_scale': 0.8 * box_overlay_ratio,
|
165 |
+
'text_thickness': max(int(2 * box_overlay_ratio), 1),
|
166 |
+
'text_padding': max(int(3 * box_overlay_ratio), 1),
|
167 |
+
'thickness': max(int(3 * box_overlay_ratio), 1),
|
168 |
+
}
|
169 |
+
|
170 |
+
ocr_bbox_rslt, is_goal_filtered = check_ocr_box(image_input, display_img = False, output_bb_format='xyxy', goal_filtering=None, easyocr_args={'paragraph': False, 'text_threshold':0.9}, use_paddleocr=use_paddleocr)
|
171 |
+
text, ocr_bbox = ocr_bbox_rslt
|
172 |
+
dino_labled_img, label_coordinates, parsed_content_list = get_som_labeled_img(image_input, yolo_model, BOX_TRESHOLD = box_threshold, output_coord_in_ratio=False, ocr_bbox=ocr_bbox,draw_bbox_config=draw_bbox_config, caption_model_processor=caption_model_processor, ocr_text=text,iou_threshold=iou_threshold, imgsz=imgsz,)
|
173 |
+
parsed_content_list = '\n'.join([f'icon {i}: ' + str(v) for i,v in enumerate(parsed_content_list)])
|
174 |
+
|
175 |
+
if len(instruction) == 0:
|
176 |
+
logger.warning('finish processing')
|
177 |
+
image = Image.open(io.BytesIO(base64.b64decode(dino_labled_img)))
|
178 |
+
return image, str(parsed_content_list)
|
179 |
+
|
180 |
+
elif instruction.startswith('Q:'):
|
181 |
+
response = get_qa_response(instruction, image_input)
|
182 |
+
return image_input, response
|
183 |
+
|
184 |
+
# parsed_content_list = str(parsed_content_list)
|
185 |
+
# convert xywh to yxhw
|
186 |
+
label_coordinates_yxhw = {}
|
187 |
+
for key, val in label_coordinates.items():
|
188 |
+
if val[2] < 0 or val[3] < 0:
|
189 |
+
continue
|
190 |
+
label_coordinates_yxhw[key] = [val[1], val[0], val[3], val[2]]
|
191 |
+
image_som = plot_boxes_with_marks(image_input.copy(), [val for key, val in label_coordinates_yxhw.items()], som_generator, edgecolor=(255,0,0), fn_save=None, normalized_to_pixel=False)
|
192 |
+
|
193 |
+
# convert xywh to xyxy
|
194 |
+
for key, val in label_coordinates.items():
|
195 |
+
label_coordinates[key] = [val[0], val[1], val[0] + val[2], val[1] + val[3]]
|
196 |
+
|
197 |
+
# normalize label_coordinates
|
198 |
+
for key, val in label_coordinates.items():
|
199 |
+
label_coordinates[key] = [val[0] / image_input.size[0], val[1] / image_input.size[1], val[2] / image_input.size[0], val[3] / image_input.size[1]]
|
200 |
+
|
201 |
+
magma_response = get_som_response(instruction, image_som)
|
202 |
+
logger.warning("magma repsonse: ", magma_response)
|
203 |
+
|
204 |
+
# map magma_response into the mark id
|
205 |
+
mark_id = extract_mark_id(magma_response)
|
206 |
+
if mark_id is not None:
|
207 |
+
if str(mark_id) in label_coordinates:
|
208 |
+
bbox_for_mark = label_coordinates[str(mark_id)]
|
|
|
|
|
|
|
|
|
209 |
else:
|
210 |
bbox_for_mark = None
|
211 |
+
else:
|
212 |
+
bbox_for_mark = None
|
213 |
+
|
214 |
+
if bbox_for_mark:
|
215 |
+
# draw bbox_for_mark on the image
|
216 |
+
image_som = plot_boxes_with_marks(
|
217 |
+
image_input,
|
218 |
+
[label_coordinates_yxhw[str(mark_id)]],
|
219 |
+
som_generator,
|
220 |
+
edgecolor=(255,127,111),
|
221 |
+
alpha=30,
|
222 |
+
fn_save=None,
|
223 |
+
normalized_to_pixel=False,
|
224 |
+
add_mark=False
|
225 |
+
)
|
226 |
+
else:
|
227 |
+
try:
|
228 |
+
if 'box' in magma_response:
|
229 |
+
pred_bbox = extract_bbox(magma_response)
|
230 |
+
click_point = [(pred_bbox[0][0] + pred_bbox[1][0]) / 2, (pred_bbox[0][1] + pred_bbox[1][1]) / 2]
|
231 |
+
click_point = [item / 1000 for item in click_point]
|
232 |
+
else:
|
233 |
+
click_point = pred_2_point(magma_response)
|
234 |
+
# de-normalize click_point (width, height)
|
235 |
+
click_point = [click_point[0] * image_input.size[0], click_point[1] * image_input.size[1]]
|
236 |
+
|
237 |
+
image_som = plot_circles_with_marks(
|
238 |
image_input,
|
239 |
+
[click_point],
|
240 |
+
som_generator,
|
241 |
edgecolor=(255,127,111),
|
242 |
+
linewidth=3,
|
243 |
+
fn_save=None,
|
244 |
normalized_to_pixel=False,
|
245 |
add_mark=False
|
246 |
)
|
247 |
+
except:
|
248 |
+
image_som = image_input
|
249 |
+
|
250 |
+
return image_som, str(parsed_content_list)
|
251 |
+
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
252 |
with gr.Blocks() as demo:
|
253 |
gr.Markdown(MARKDOWN)
|
254 |
with gr.Row():
|
|
|
287 |
outputs=[image_output_component, text_output_component]
|
288 |
)
|
289 |
|
290 |
+
# demo.launch(debug=False, show_error=True, share=True)
|
291 |
# demo.launch(share=True, server_port=7861, server_name='0.0.0.0')
|
292 |
+
demo.queue().launch(share=False)
|