Upload folder using huggingface_hub
Browse files
api.py
CHANGED
@@ -3,25 +3,34 @@ from flask import Flask, request, jsonify
|
|
3 |
from PIL import Image
|
4 |
from io import BytesIO
|
5 |
import base64
|
|
|
6 |
from transformers import AutoProcessor, AutoModelForCausalLM
|
7 |
import threading
|
8 |
-
from unittest.mock import patch
|
9 |
-
from transformers.dynamic_module_utils import get_imports
|
10 |
|
11 |
app = Flask(__name__)
|
12 |
|
13 |
# Parse command line arguments
|
14 |
parser = argparse.ArgumentParser(description='Start the Flask server with specified model and device.')
|
15 |
-
parser.add_argument('--model-path', type=str,
|
16 |
parser.add_argument('--device', type=str, choices=['cpu', 'gpu'], default='auto', help='Device to use: "cpu", "gpu", or "auto"')
|
17 |
args = parser.parse_args()
|
18 |
|
19 |
# Determine the device
|
20 |
-
device
|
21 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
22 |
|
23 |
-
|
24 |
-
|
|
|
|
|
|
|
25 |
|
26 |
def fixed_get_imports(filename: str | os.PathLike) -> list[str]:
|
27 |
if not str(filename).endswith("modeling_florence2.py"):
|
@@ -30,9 +39,10 @@ def fixed_get_imports(filename: str | os.PathLike) -> list[str]:
|
|
30 |
imports.remove("flash_attn")
|
31 |
return imports
|
32 |
|
33 |
-
|
34 |
-
|
35 |
-
|
|
|
36 |
|
37 |
lock = threading.Lock() # Use a lock to ensure thread safety when accessing the model
|
38 |
|
@@ -40,7 +50,7 @@ def predict_image(image, task: str = "<OD>", prompt: str = None):
|
|
40 |
prompt = task + " " + prompt if prompt else task
|
41 |
print(f"Prompt: {prompt}")
|
42 |
with lock:
|
43 |
-
inputs = processor(text=prompt, images=image, return_tensors="pt").to(device)
|
44 |
generated_ids = model.generate(
|
45 |
input_ids=inputs["input_ids"],
|
46 |
pixel_values=inputs["pixel_values"],
|
|
|
3 |
from PIL import Image
|
4 |
from io import BytesIO
|
5 |
import base64
|
6 |
+
import torch
|
7 |
from transformers import AutoProcessor, AutoModelForCausalLM
|
8 |
import threading
|
|
|
|
|
9 |
|
10 |
app = Flask(__name__)
|
11 |
|
12 |
# Parse command line arguments
|
13 |
parser = argparse.ArgumentParser(description='Start the Flask server with specified model and device.')
|
14 |
+
parser.add_argument('--model-path', type=str, required=True, help='Path to the pretrained model')
|
15 |
parser.add_argument('--device', type=str, choices=['cpu', 'gpu'], default='auto', help='Device to use: "cpu", "gpu", or "auto"')
|
16 |
args = parser.parse_args()
|
17 |
|
18 |
# Determine the device
|
19 |
+
if args.device == 'auto':
|
20 |
+
device = "cuda:0" if torch.cuda.is_available() else "cpu"
|
21 |
+
elif args.device == 'gpu':
|
22 |
+
if torch.cuda.is_available():
|
23 |
+
device = "cuda:0"
|
24 |
+
else:
|
25 |
+
raise ValueError("GPU option specified but no GPU is available.")
|
26 |
+
else:
|
27 |
+
device = "cpu"
|
28 |
|
29 |
+
torch_dtype = torch.float16 if device.startswith("cuda") else torch.float32
|
30 |
+
|
31 |
+
from unittest.mock import patch
|
32 |
+
from transformers.dynamic_module_utils import get_imports
|
33 |
+
import os
|
34 |
|
35 |
def fixed_get_imports(filename: str | os.PathLike) -> list[str]:
|
36 |
if not str(filename).endswith("modeling_florence2.py"):
|
|
|
39 |
imports.remove("flash_attn")
|
40 |
return imports
|
41 |
|
42 |
+
# Initialize the model and processor
|
43 |
+
with patch("transformers.dynamic_module_utils.get_imports", fixed_get_imports): #workaround for unnecessary flash_attn requirement
|
44 |
+
model = AutoModelForCausalLM.from_pretrained(args.model_path, attn_implementation="sdpa", torch_dtype=torch_dtype,trust_remote_code=True).to(device)
|
45 |
+
processor = AutoProcessor.from_pretrained(args.model_path, trust_remote_code=True)
|
46 |
|
47 |
lock = threading.Lock() # Use a lock to ensure thread safety when accessing the model
|
48 |
|
|
|
50 |
prompt = task + " " + prompt if prompt else task
|
51 |
print(f"Prompt: {prompt}")
|
52 |
with lock:
|
53 |
+
inputs = processor(text=prompt, images=image, return_tensors="pt").to(device, torch_dtype)
|
54 |
generated_ids = model.generate(
|
55 |
input_ids=inputs["input_ids"],
|
56 |
pixel_values=inputs["pixel_values"],
|