-
Notifications
You must be signed in to change notification settings - Fork 12
/
model_loader.py
95 lines (79 loc) · 2.84 KB
/
model_loader.py
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
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
from pathlib import Path
from typing import Any, Dict
import torch
from tsm import TSM
from tsn import TSN, TRN, MTRN
verb_class_count, noun_class_count = 125, 352
class_count = (verb_class_count, noun_class_count)
def make_tsn(settings):
return TSN(
class_count,
settings["segment_count"],
settings["modality"],
base_model=settings["arch"],
new_length=settings["flow_length"] if settings["modality"] == "Flow" else 1,
consensus_type=settings["consensus_type"],
dropout=settings["dropout"],
)
def make_trn(settings):
model_type = settings["model_type"]
if model_type == "trn":
cls = TRN
elif model_type == "mtrn":
cls = MTRN
else:
raise ValueError(f"Unknown model_type '{model_type}' for TRN")
return cls(
class_count,
settings["segment_count"],
settings["modality"],
base_model=settings["arch"],
new_length=settings["flow_length"] if settings["modality"] == "Flow" else 1,
img_feature_dim=settings["img_feature_dim"],
dropout=settings["dropout"],
)
def make_tsm(settings):
non_local = settings["model_type"].endswith("-nl")
return TSM(
class_count,
settings["segment_count"],
settings["modality"],
base_model=settings["arch"],
new_length=settings["flow_length"] if settings["modality"] == "Flow" else 1,
consensus_type="avg",
dropout=settings["dropout"],
shift_div=settings["shift_div"],
shift_place=settings["shift_place"],
temporal_pool=settings["temporal_pool"],
non_local=non_local,
)
def make_model(settings: Dict[str, Any]) -> torch.nn.Module:
model_factories = {
"tsn": make_tsn,
"trn": make_trn,
"mtrn": make_trn,
"tsm": make_tsm,
"tsm-nl": make_tsm,
}
return model_factories[settings["model_type"]](settings)
def get_model_settings_from_checkpoint(ckpt: Dict[str, Any]) -> Dict[str, Any]:
settings = {
key: ckpt[key] for key in ["model_type", "segment_count", "modality", "arch"]
}
if ckpt["model_type"] == "tsn":
settings["consensus_type"] = ckpt["consensus_type"]
if ckpt["model_type"] in ["tsm", "tsm-nl"]:
for key in ["shift_place", "shift_div", "temporal_pool", "non_local"]:
settings[key] = ckpt[key]
if ckpt["model_type"] in ["trn", "mtrn"]:
settings["img_feature_dim"] = ckpt["img_feature_dim"]
settings.update(
{key: getattr(ckpt["args"], key) for key in ["flow_length", "dropout"]}
)
return settings
def load_checkpoint(checkpoint_path: Path) -> torch.nn.Module:
ckpt = torch.load(checkpoint_path)
model_settings = get_model_settings_from_checkpoint(ckpt)
model = make_model(model_settings)
model.load_state_dict(ckpt["state_dict"])
return model