Commit
•
0546112
1
Parent(s):
db4cf02
Modified handler to load BLIP directly from transformers
Browse files- handler.py +21 -26
- requirements.txt +1 -5
handler.py
CHANGED
@@ -1,3 +1,4 @@
|
|
|
|
1 |
from typing import Dict, List, Any
|
2 |
from PIL import Image
|
3 |
import requests
|
@@ -5,32 +6,26 @@ import torch
|
|
5 |
import base64
|
6 |
import os
|
7 |
from io import BytesIO
|
|
|
|
|
8 |
from models.blip_decoder import blip_decoder
|
9 |
from torchvision import transforms
|
10 |
from torchvision.transforms.functional import InterpolationMode
|
|
|
11 |
|
12 |
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
|
13 |
|
14 |
class EndpointHandler():
|
15 |
def __init__(self, path=""):
|
16 |
# load the optimized model
|
17 |
-
|
18 |
-
self.
|
19 |
-
|
20 |
-
|
21 |
-
|
22 |
-
med_config=os.path.join(path, 'configs/med_config.json')
|
23 |
-
)
|
24 |
self.model.eval()
|
25 |
self.model = self.model.to(device)
|
26 |
|
27 |
-
image_size = 384
|
28 |
-
self.transform = transforms.Compose([
|
29 |
-
transforms.Resize((image_size,image_size),interpolation=InterpolationMode.BICUBIC),
|
30 |
-
transforms.ToTensor(),
|
31 |
-
transforms.Normalize((0.48145466, 0.4578275, 0.40821073), (0.26862954, 0.26130258, 0.27577711))
|
32 |
-
])
|
33 |
-
|
34 |
|
35 |
|
36 |
def __call__(self, data: Any) -> Dict[str, Any]:
|
@@ -39,22 +34,22 @@ class EndpointHandler():
|
|
39 |
data (:obj:):
|
40 |
includes the input data and the parameters for the inference.
|
41 |
Return:
|
42 |
-
A :obj:`dict`:. The object returned should be a dict of one list like {"
|
43 |
- "caption": A string corresponding to the generated caption.
|
44 |
"""
|
45 |
inputs = data.pop("inputs", data)
|
46 |
parameters = data.pop("parameters", {})
|
47 |
-
|
48 |
-
|
49 |
-
|
50 |
-
|
|
|
|
|
|
|
51 |
with torch.no_grad():
|
52 |
-
|
53 |
-
|
54 |
-
sample=parameters.get('sample',True),
|
55 |
-
top_p=parameters.get('top_p',0.9),
|
56 |
-
max_length=parameters.get('max_length',20),
|
57 |
-
min_length=parameters.get('min_length',5)
|
58 |
)
|
|
|
59 |
# postprocess the prediction
|
60 |
-
return {"
|
|
|
1 |
+
# +
|
2 |
from typing import Dict, List, Any
|
3 |
from PIL import Image
|
4 |
import requests
|
|
|
6 |
import base64
|
7 |
import os
|
8 |
from io import BytesIO
|
9 |
+
|
10 |
+
from transformers import BlipForConditionalGeneration, BlipProcessor
|
11 |
from models.blip_decoder import blip_decoder
|
12 |
from torchvision import transforms
|
13 |
from torchvision.transforms.functional import InterpolationMode
|
14 |
+
# -
|
15 |
|
16 |
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
|
17 |
|
18 |
class EndpointHandler():
|
19 |
def __init__(self, path=""):
|
20 |
# load the optimized model
|
21 |
+
|
22 |
+
self.processor = BlipProcessor.from_pretrained("Salesforce/blip-image-captioning-base")
|
23 |
+
self.model = BlipForConditionalGeneration.from_pretrained(
|
24 |
+
"Salesforce/blip-image-captioning-base"
|
25 |
+
).to(device)
|
|
|
|
|
26 |
self.model.eval()
|
27 |
self.model = self.model.to(device)
|
28 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
29 |
|
30 |
|
31 |
def __call__(self, data: Any) -> Dict[str, Any]:
|
|
|
34 |
data (:obj:):
|
35 |
includes the input data and the parameters for the inference.
|
36 |
Return:
|
37 |
+
A :obj:`dict`:. The object returned should be a dict of one list like {"captions": ["A hugging face at the office"]} containing :
|
38 |
- "caption": A string corresponding to the generated caption.
|
39 |
"""
|
40 |
inputs = data.pop("inputs", data)
|
41 |
parameters = data.pop("parameters", {})
|
42 |
+
|
43 |
+
raw_images = [Image.open(BytesIO(_img)) for _img in inputs]
|
44 |
+
|
45 |
+
processed_image = self.processor(images=raw_images, return_tensors="pt")
|
46 |
+
processed_image["pixel_values"] = processed_image["pixel_values"].to(device)
|
47 |
+
processed_image = {**processed_image, **parameters}
|
48 |
+
|
49 |
with torch.no_grad():
|
50 |
+
out = self.model.generate(
|
51 |
+
**processed_image
|
|
|
|
|
|
|
|
|
52 |
)
|
53 |
+
captions = self.processor.batch_decode(out, skip_special_tokens=True)
|
54 |
# postprocess the prediction
|
55 |
+
return {"captions": captions}
|
requirements.txt
CHANGED
@@ -1,5 +1 @@
|
|
1 |
-
|
2 |
-
transformers==4.15.0
|
3 |
-
fairscale==0.4.4
|
4 |
-
requests
|
5 |
-
Pillow
|
|
|
1 |
+
git+https://github.com/huggingface/transformers.git@main
|
|
|
|
|
|
|
|