Spaces:
Sleeping
Sleeping
Update my_model/KBVQA.py
Browse files- my_model/KBVQA.py +5 -5
my_model/KBVQA.py
CHANGED
|
@@ -9,7 +9,7 @@ from my_model.utilities.gen_utilities import free_gpu_resources
|
|
| 9 |
from my_model.captioner.image_captioning import ImageCaptioningModel
|
| 10 |
from my_model.object_detection import ObjectDetector
|
| 11 |
import my_model.config.kbvqa_config as config
|
| 12 |
-
|
| 13 |
|
| 14 |
|
| 15 |
class KBVQA(StateManager):
|
|
@@ -50,7 +50,7 @@ class KBVQA(StateManager):
|
|
| 50 |
|
| 51 |
def __init__(self):
|
| 52 |
|
| 53 |
-
|
| 54 |
self.kbvqa_model_name: str = config.KBVQA_MODEL_NAME
|
| 55 |
self.quantization: str = config.QUANTIZATION
|
| 56 |
self.max_context_window: int = config.MAX_CONTEXT_WINDOW
|
|
@@ -241,7 +241,7 @@ class KBVQA(StateManager):
|
|
| 241 |
prompt = self.format_prompt(question, caption=caption, objects=detected_objects_str)
|
| 242 |
num_tokens = len(self.kbvqa_tokenizer.tokenize(prompt))
|
| 243 |
if num_tokens > self.max_context_window:
|
| 244 |
-
|
| 245 |
return
|
| 246 |
|
| 247 |
model_inputs = self.kbvqa_tokenizer(prompt, add_special_tokens=False, return_tensors="pt").to('cuda')
|
|
@@ -268,7 +268,7 @@ def prepare_kbvqa_model(only_reload_detection_model: bool = False) -> KBVQA:
|
|
| 268 |
kbvqa = KBVQA()
|
| 269 |
kbvqa.detection_model = st.session_state.detection_model
|
| 270 |
# Progress bar for model loading
|
| 271 |
-
with
|
| 272 |
|
| 273 |
if not only_reload_detection_model:
|
| 274 |
self.col1.text('this should take no more than a few minutes!')
|
|
@@ -283,7 +283,7 @@ def prepare_kbvqa_model(only_reload_detection_model: bool = False) -> KBVQA:
|
|
| 283 |
free_gpu_resources()
|
| 284 |
progress_bar.progress(100)
|
| 285 |
else:
|
| 286 |
-
progress_bar =
|
| 287 |
kbvqa.load_detector(kbvqa.detection_model)
|
| 288 |
progress_bar.progress(100)
|
| 289 |
|
|
|
|
| 9 |
from my_model.captioner.image_captioning import ImageCaptioningModel
|
| 10 |
from my_model.object_detection import ObjectDetector
|
| 11 |
import my_model.config.kbvqa_config as config
|
| 12 |
+
|
| 13 |
|
| 14 |
|
| 15 |
class KBVQA(StateManager):
|
|
|
|
| 50 |
|
| 51 |
def __init__(self):
|
| 52 |
|
| 53 |
+
self.col1, self.col2, self.col3 = st.columns([0.2, 0.6, 0.2])
|
| 54 |
self.kbvqa_model_name: str = config.KBVQA_MODEL_NAME
|
| 55 |
self.quantization: str = config.QUANTIZATION
|
| 56 |
self.max_context_window: int = config.MAX_CONTEXT_WINDOW
|
|
|
|
| 241 |
prompt = self.format_prompt(question, caption=caption, objects=detected_objects_str)
|
| 242 |
num_tokens = len(self.kbvqa_tokenizer.tokenize(prompt))
|
| 243 |
if num_tokens > self.max_context_window:
|
| 244 |
+
self.col2.write(f"Prompt too long with {num_tokens} tokens, consider increasing the confidence threshold for the object detector")
|
| 245 |
return
|
| 246 |
|
| 247 |
model_inputs = self.kbvqa_tokenizer(prompt, add_special_tokens=False, return_tensors="pt").to('cuda')
|
|
|
|
| 268 |
kbvqa = KBVQA()
|
| 269 |
kbvqa.detection_model = st.session_state.detection_model
|
| 270 |
# Progress bar for model loading
|
| 271 |
+
with self.col1.spinner('Loading model...'):
|
| 272 |
|
| 273 |
if not only_reload_detection_model:
|
| 274 |
self.col1.text('this should take no more than a few minutes!')
|
|
|
|
| 283 |
free_gpu_resources()
|
| 284 |
progress_bar.progress(100)
|
| 285 |
else:
|
| 286 |
+
progress_bar = self.col1.progress(0)
|
| 287 |
kbvqa.load_detector(kbvqa.detection_model)
|
| 288 |
progress_bar.progress(100)
|
| 289 |
|