Commit
·
89e64c9
1
Parent(s):
6640fa0
Update app.py
Browse files
app.py
CHANGED
|
@@ -259,7 +259,7 @@ def draw_text(
|
|
| 259 |
if 'model' not in st.session_state:
|
| 260 |
#with st.spinner('We are orginizing your traks...'):
|
| 261 |
text_encoder = AutoModel.from_pretrained(CLIP_TEXT_MODEL_PATH, local_files_only=True)
|
| 262 |
-
vision_encoder = CLIPVisionModel.from_pretrained(CLIP_VISION_MODEL_PATH, local_files_only=True)
|
| 263 |
tokenizer = AutoTokenizer.from_pretrained(TEXT_MODEL)
|
| 264 |
model = CLIPDemo(vision_encoder=vision_encoder, text_encoder=text_encoder, tokenizer=tokenizer, device=device)
|
| 265 |
model.compute_image_embeddings(glob.glob(SPECTROGRAMS_PATH + "/*.jpeg")[:1000])
|
|
@@ -313,7 +313,7 @@ def draw_audio(
|
|
| 313 |
if 'model' not in st.session_state:
|
| 314 |
#with st.spinner('We are orginizing your traks...'):
|
| 315 |
text_encoder = AutoModel.from_pretrained(CLIP_TEXT_MODEL_PATH, local_files_only=True)
|
| 316 |
-
vision_encoder = CLIPVisionModel.from_pretrained(CLIP_VISION_MODEL_PATH, local_files_only=True)
|
| 317 |
tokenizer = AutoTokenizer.from_pretrained(TEXT_MODEL)
|
| 318 |
model = CLIPDemo(vision_encoder=vision_encoder, text_encoder=text_encoder, tokenizer=tokenizer, device=device)
|
| 319 |
model.compute_image_embeddings(glob.glob(SPECTROGRAMS_PATH+"/*.jpeg")[:5000])
|
|
@@ -381,7 +381,7 @@ def draw_camera(
|
|
| 381 |
if 'model' not in st.session_state:
|
| 382 |
#with st.spinner('We are orginizing your traks...'):
|
| 383 |
text_encoder = AutoModel.from_pretrained(CLIP_TEXT_MODEL_PATH, local_files_only=True)
|
| 384 |
-
vision_encoder = CLIPVisionModel.from_pretrained(CLIP_VISION_MODEL_PATH, local_files_only=True)
|
| 385 |
tokenizer = AutoTokenizer.from_pretrained(TEXT_MODEL)
|
| 386 |
model = CLIPDemo(vision_encoder=vision_encoder, text_encoder=text_encoder, tokenizer=tokenizer, device=device)
|
| 387 |
model.compute_image_embeddings(glob.glob(SPECTROGRAMS_PATH + "/*.jpeg")[:5000])
|
|
|
|
| 259 |
if 'model' not in st.session_state:
|
| 260 |
#with st.spinner('We are orginizing your traks...'):
|
| 261 |
text_encoder = AutoModel.from_pretrained(CLIP_TEXT_MODEL_PATH, local_files_only=True)
|
| 262 |
+
vision_encoder = CLIPVisionModel.from_pretrained(CLIP_VISION_MODEL_PATH, local_files_only=True).to(device)
|
| 263 |
tokenizer = AutoTokenizer.from_pretrained(TEXT_MODEL)
|
| 264 |
model = CLIPDemo(vision_encoder=vision_encoder, text_encoder=text_encoder, tokenizer=tokenizer, device=device)
|
| 265 |
model.compute_image_embeddings(glob.glob(SPECTROGRAMS_PATH + "/*.jpeg")[:1000])
|
|
|
|
| 313 |
if 'model' not in st.session_state:
|
| 314 |
#with st.spinner('We are orginizing your traks...'):
|
| 315 |
text_encoder = AutoModel.from_pretrained(CLIP_TEXT_MODEL_PATH, local_files_only=True)
|
| 316 |
+
vision_encoder = CLIPVisionModel.from_pretrained(CLIP_VISION_MODEL_PATH, local_files_only=True).to(device)
|
| 317 |
tokenizer = AutoTokenizer.from_pretrained(TEXT_MODEL)
|
| 318 |
model = CLIPDemo(vision_encoder=vision_encoder, text_encoder=text_encoder, tokenizer=tokenizer, device=device)
|
| 319 |
model.compute_image_embeddings(glob.glob(SPECTROGRAMS_PATH+"/*.jpeg")[:5000])
|
|
|
|
| 381 |
if 'model' not in st.session_state:
|
| 382 |
#with st.spinner('We are orginizing your traks...'):
|
| 383 |
text_encoder = AutoModel.from_pretrained(CLIP_TEXT_MODEL_PATH, local_files_only=True)
|
| 384 |
+
vision_encoder = CLIPVisionModel.from_pretrained(CLIP_VISION_MODEL_PATH, local_files_only=True).to(device)
|
| 385 |
tokenizer = AutoTokenizer.from_pretrained(TEXT_MODEL)
|
| 386 |
model = CLIPDemo(vision_encoder=vision_encoder, text_encoder=text_encoder, tokenizer=tokenizer, device=device)
|
| 387 |
model.compute_image_embeddings(glob.glob(SPECTROGRAMS_PATH + "/*.jpeg")[:5000])
|