File size: 2,643 Bytes
f4aead5 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 | import torch
import torchvision.transforms as transforms
_base_ = ['../base.py']
config = dict(
train_config=[
dict(
type='Recognition_frame',
csv_root='/gpfswork/rech/okw/ukw13bv/mmsl/csv/cholec80/csvs',
vid='video%02d.csv'%i,
video_root='/gpfsscratch/rech/okw/ukw13bv/cholec80/frames_output',
transforms=transforms.Compose(
[
transforms.Resize((360, 640)),
transforms.CenterCrop(224),
transforms.ToTensor(),
transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
]
),
) for i in range(1, 41)
],
val_config=[
dict(
type='Recognition_frame',
csv_root='/gpfswork/rech/okw/ukw13bv/mmsl/csv/cholec80/csvs',
vid='video%02d.csv'%i,
video_root='/gpfsscratch/rech/okw/ukw13bv/cholec80/frames_output',
transforms=transforms.Compose(
[
transforms.Resize((360, 640)),
transforms.CenterCrop(224),
transforms.ToTensor(),
transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
]
),
) for i in range(41, 49)
],
test_config=[
dict(
type='Recognition_frame',
csv_root='/gpfswork/rech/okw/ukw13bv/mmsl/csv/cholec80/csvs',
vid='video%02d.csv'%i,
video_root='/gpfsscratch/rech/okw/ukw13bv/cholec80/frames_output',
transforms=transforms.Compose(
[
transforms.Resize((360, 640)),
transforms.CenterCrop(224),
transforms.ToTensor(),
transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
]
),
) for i in range(49, 81)
],
model_config = dict(
type='MVNet_feature_extractor',
backbone_img = dict(
type='img_backbones/ImageEncoder_feature_extractor',
# type='img_backbones/ImageEncoder_CLIPVISUAL',
num_classes=768,
pretrained='imagenet', # imagenet/ssl/random
backbone_name='resnet_50',
# backbone_name='resnet_50_clip'
img_norm=False,
),
backbone_text= dict(
type='text_backbones/BertEncoder',
text_bert_type='/gpfswork/rech/okw/ukw13bv/mmsl/biobert_pretrain_output_all_notes_150000',
text_last_n_layers=4,
text_aggregate_method='sum',
text_norm=False,
text_embedding_dim=768,
text_freeze_bert=False,
text_agg_tokens=True
)
)
)
|