Commit
•
2a79ef4
1
Parent(s):
acef01f
Creating captioning pipeline with nucleus sampling
Browse files- pipeline.py +18 -13
pipeline.py
CHANGED
@@ -2,21 +2,28 @@ from typing import Dict, List, Any
|
|
2 |
from PIL import Image
|
3 |
import requests
|
4 |
import torch
|
|
|
5 |
from torchvision import transforms
|
6 |
from torchvision.transforms.functional import InterpolationMode
|
7 |
|
8 |
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
|
9 |
|
10 |
-
from transformers import pipeline, AutoTokenizer
|
11 |
-
|
12 |
|
13 |
class PreTrainedPipeline():
|
14 |
def __init__(self, path=""):
|
15 |
# load the optimized model
|
16 |
-
|
17 |
-
|
18 |
-
|
19 |
-
self.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
20 |
|
21 |
|
22 |
def __call__(self, data: Any) -> List[List[Dict[str, float]]]:
|
@@ -29,13 +36,11 @@ class PreTrainedPipeline():
|
|
29 |
- "label": A string representing what the label/class is. There can be multiple labels.
|
30 |
- "score": A score between 0 and 1 describing how confident the model is for this label/class.
|
31 |
"""
|
32 |
-
|
33 |
parameters = data.pop("parameters", None)
|
34 |
|
35 |
-
|
36 |
-
|
37 |
-
|
38 |
-
else:
|
39 |
-
prediction = self.pipeline(inputs)
|
40 |
# postprocess the prediction
|
41 |
-
return
|
|
|
2 |
from PIL import Image
|
3 |
import requests
|
4 |
import torch
|
5 |
+
from blip import blip_decoder
|
6 |
from torchvision import transforms
|
7 |
from torchvision.transforms.functional import InterpolationMode
|
8 |
|
9 |
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
|
10 |
|
|
|
|
|
11 |
|
12 |
class PreTrainedPipeline():
|
13 |
def __init__(self, path=""):
|
14 |
# load the optimized model
|
15 |
+
self.model_url = 'https://storage.googleapis.com/sfr-vision-language-research/BLIP/models/model_large_caption.pth'
|
16 |
+
self.model = blip_decoder(pretrained=self.model_url, image_size=384, vit='large')
|
17 |
+
self.model.eval()
|
18 |
+
self.model = model.to(device)
|
19 |
+
|
20 |
+
image_size = 384
|
21 |
+
self.transform = transforms.Compose([
|
22 |
+
transforms.Resize((image_size,image_size),interpolation=InterpolationMode.BICUBIC),
|
23 |
+
transforms.ToTensor(),
|
24 |
+
transforms.Normalize((0.48145466, 0.4578275, 0.40821073), (0.26862954, 0.26130258, 0.27577711))
|
25 |
+
])
|
26 |
+
|
27 |
|
28 |
|
29 |
def __call__(self, data: Any) -> List[List[Dict[str, float]]]:
|
|
|
36 |
- "label": A string representing what the label/class is. There can be multiple labels.
|
37 |
- "score": A score between 0 and 1 describing how confident the model is for this label/class.
|
38 |
"""
|
39 |
+
image = data.pop("inputs", data)
|
40 |
parameters = data.pop("parameters", None)
|
41 |
|
42 |
+
image = transform(image).unsqueeze(0).to(device)
|
43 |
+
with torch.no_grad():
|
44 |
+
caption = self.model.generate(image, sample=True, top_p=0.9, max_length=20, min_length=5)
|
|
|
|
|
45 |
# postprocess the prediction
|
46 |
+
return caption
|