Update prediction.py
Browse files- prediction.py +1 -23
prediction.py
CHANGED
@@ -1,15 +1,9 @@
|
|
1 |
-
import os
|
2 |
-
os.chdir('..')
|
3 |
-
base_dir = os.getcwd()
|
4 |
from dataloader import CellLoader
|
5 |
-
from celle_main import instantiate_from_config
|
6 |
-
from omegaconf import OmegaConf
|
7 |
|
8 |
def run_image_prediction(
|
9 |
sequence_input,
|
10 |
nucleus_image,
|
11 |
-
|
12 |
-
model_config_path,
|
13 |
device
|
14 |
):
|
15 |
"""
|
@@ -37,22 +31,6 @@ def run_image_prediction(
|
|
37 |
# Convert SEQUENCE to sequence using dataset.tokenize_sequence()
|
38 |
sequence = dataset.tokenize_sequence(sequence_input)
|
39 |
|
40 |
-
# Load model config and set ckpt_path if not provided in config
|
41 |
-
config = OmegaConf.load(model_config_path)
|
42 |
-
if config["model"]["params"]["ckpt_path"] is None:
|
43 |
-
config["model"]["params"]["ckpt_path"] = model_ckpt_path
|
44 |
-
|
45 |
-
# Set condition_model_path and vqgan_model_path to None
|
46 |
-
config["model"]["params"]["condition_model_path"] = None
|
47 |
-
config["model"]["params"]["vqgan_model_path"] = None
|
48 |
-
|
49 |
-
os.chdir(os.path.dirname(model_ckpt_path))
|
50 |
-
|
51 |
-
# Instantiate model from config and move to device
|
52 |
-
model = instantiate_from_config(config.model).to(device)
|
53 |
-
|
54 |
-
os.chdir(base_dir)
|
55 |
-
|
56 |
# Sample from model using provided sequence and nucleus image
|
57 |
_, _, _, predicted_threshold, predicted_heatmap = model.celle.sample(
|
58 |
text=sequence.to(device),
|
|
|
|
|
|
|
|
|
1 |
from dataloader import CellLoader
|
|
|
|
|
2 |
|
3 |
def run_image_prediction(
|
4 |
sequence_input,
|
5 |
nucleus_image,
|
6 |
+
model,
|
|
|
7 |
device
|
8 |
):
|
9 |
"""
|
|
|
31 |
# Convert SEQUENCE to sequence using dataset.tokenize_sequence()
|
32 |
sequence = dataset.tokenize_sequence(sequence_input)
|
33 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
34 |
# Sample from model using provided sequence and nucleus image
|
35 |
_, _, _, predicted_threshold, predicted_heatmap = model.celle.sample(
|
36 |
text=sequence.to(device),
|