File size: 5,486 Bytes
0b32ad6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
import logging
from collections import defaultdict

from s3prl.base.container import Container
from s3prl.base.workspace import Workspace
from s3prl.corpus.iemocap import iemocap_for_superb
from s3prl.dataset.utterance_classification_pipe import UtteranceClassificationPipe
from s3prl.nn import MeanPoolingLinear
from s3prl.sampler import FixedBatchSizeBatchSampler, MaxTimestampBatchSampler
from s3prl.task.utterance_classification_task import UtteranceClassificationTask
from s3prl.util.configuration import default_cfg, field

from .base import SuperbProblem

logger = logging.getLogger(__name__)


class SuperbER(SuperbProblem):
    """
    Superb Emotion Classification problem
    """

    @default_cfg(
        **SuperbProblem.setup.default_except(
            corpus=dict(
                CLS=iemocap_for_superb,
                dataset_root="???",
                test_fold=field(
                    "???",
                    "The session in IEMOCAP used for testing.\n"
                    "The other sessions will be used for training and validation.",
                ),
            ),
            train_datapipe=dict(
                CLS=UtteranceClassificationPipe,
                train_category_encoder=True,
            ),
            train_sampler=dict(
                CLS=FixedBatchSizeBatchSampler,
                batch_size=4,
                shuffle=True,
            ),
            valid_datapipe=dict(
                CLS=UtteranceClassificationPipe,
            ),
            valid_sampler=dict(
                CLS=FixedBatchSizeBatchSampler,
                batch_size=4,
            ),
            test_datapipe=dict(
                CLS=UtteranceClassificationPipe,
            ),
            test_sampler=dict(
                CLS=FixedBatchSizeBatchSampler,
                batch_size=4,
            ),
            downstream=dict(
                CLS=MeanPoolingLinear,
                hidden_size=256,
            ),
            task=dict(
                CLS=UtteranceClassificationTask,
            ),
        )
    )
    @classmethod
    def setup(cls, **cfg):
        """
        This setups the ER problem, containing train/valid/test datasets & samplers and a task object
        """
        super().setup(**cfg)

    @default_cfg(
        **SuperbProblem.train.default_except(
            optimizer=dict(
                CLS="torch.optim.Adam",
                lr=1.0e-4,
            ),
            trainer=dict(
                total_steps=30000,
                log_step=500,
                eval_step=1000,
                save_step=1000,
                gradient_clipping=1.0,
                gradient_accumulate_steps=8,
                valid_metric="accuracy",
                valid_higher_better=True,
            ),
        )
    )
    @classmethod
    def train(cls, **cfg):
        """
        Train the setup problem with the train/valid datasets & samplers and the task object
        """
        super().train(**cfg)

    @default_cfg(**SuperbProblem.inference.default_cfg)
    @classmethod
    def inference(cls, **cfg):
        super().inference(**cfg)

    @default_cfg(
        **SuperbProblem.run.default_except(
            stages=["setup", "train", "inference"],
            start_stage="setup",
            final_stage="inference",
            setup=setup.default_cfg.deselect("workspace", "resume"),
            train=train.default_cfg.deselect("workspace", "resume"),
            inference=inference.default_cfg.deselect("workspace", "resume"),
        )
    )
    @classmethod
    def run(cls, **cfg):
        super().run(**cfg)

    @default_cfg(
        num_fold=field(5, "The number of folds to run cross validation", int),
        **run.default_except(
            workspace=field(
                "???",
                "The root workspace for all folds.\n"
                "Each fold will use a 'fold_{id}' sub-workspace under this root workspace",
            ),
            setup=dict(
                corpus=dict(
                    test_fold=field(
                        "TBD", "This will be auto-set by 'run_cross_validation'"
                    )
                )
            ),
        ),
    )
    @classmethod
    def cross_validation(cls, **cfg):
        """
        Except 'num_fold', all other fields are for 'run' for every fold. That is, all folds shared the same
        config (training hypers, dataset root, etc) except 'workspace' and 'test_fold' are different
        """
        cfg = Container(cfg)
        workspaces = [
            str(Workspace(cfg.workspace) / f"fold_{fold_id}")
            for fold_id in range(cfg.num_fold)
        ]
        for fold_id, workspace in enumerate(workspaces):
            fold_cfg = cfg.clone().deselect("num_fold")

            fold_cfg.workspace = workspace
            fold_cfg.setup.corpus.test_fold = fold_id
            cls.run(
                **fold_cfg,
            )
        metrics = defaultdict(list)
        for fold_id, workspace in enumerate(workspaces):
            workspace = Workspace(workspace)
            metric = workspace["test_metrics"]
            for key, value in metric.items():
                metrics[key].append(value)

        avg_result = dict()
        for key, values in metrics.items():
            avg_score = sum(values) / len(values)
            avg_result[key] = avg_score
            logger.info(f"Average {key}: {avg_score}")

        Workspace(cfg.workspace).put(avg_result, "avg_test_metrics", "yaml")