File size: 13,815 Bytes
133928a
f3edc31
133928a
f3edc31
133928a
 
 
f3edc31
 
0acaf35
7970e16
f3edc31
50c16f9
41587e8
f3edc31
 
 
133928a
3d5d967
41587e8
133928a
fd870aa
133928a
 
 
 
 
 
 
 
 
fd870aa
133928a
 
 
 
 
 
 
 
 
 
 
0acaf35
 
32552bc
0acaf35
2aad792
0acaf35
 
 
7970e16
 
6d6b31c
 
 
133928a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
d99fa4e
133928a
 
 
 
 
 
 
 
 
 
 
 
 
 
e27e5ab
f3edc31
 
 
41587e8
f3edc31
eb5a3c6
f3edc31
 
 
c8c9f56
133928a
 
 
 
 
f3edc31
 
 
133928a
f3edc31
 
 
133928a
 
 
f3edc31
 
 
64cb177
0acaf35
 
32552bc
0acaf35
 
 
 
7970e16
 
6d6b31c
 
 
133928a
f3edc31
4932186
 
9671a15
4932186
 
 
 
 
 
 
 
 
f3edc31
4932186
 
f3edc31
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
133928a
 
 
f3edc31
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
133928a
 
 
f3edc31
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
133928a
1451bd6
 
 
7c6f27f
1451bd6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
import logging
from typing import Optional, Tuple, Union

import torch
import torch.nn as nn
from torch.nn import BCEWithLogitsLoss, CrossEntropyLoss, MSELoss
from torch.nn.modules.utils import consume_prefix_in_state_dict_if_present
from transformers import BertPreTrainedModel
from transformers.modeling_outputs import SequenceClassifierOutput
from huggingface_hub import hf_hub_download
from safetensors.torch import load_file

from .bert_layers_mosa import BertModel
from .configuration_bert import BertConfig

logger = logging.getLogger(__name__)


class ClinicalMosaicForEmbeddingGeneration(BertPreTrainedModel):
    config_class = BertConfig

    def __init__(self, config, **kwargs):
        """
        Initializes the BertEmbeddings class.

        Args:
            config (BertConfig): The configuration for the BERT model.
            add_pooling_layer (bool, optional): Whether to add a pooling layer. Defaults to False.
        """
        super().__init__(config)
        self.config = config
        self.bert = BertModel(config, add_pooling_layer=False)
        # this resets the weights
        self.post_init()

    @classmethod
    def from_pretrained(
        cls, pretrained_checkpoint, state_dict=None, config=None, *inputs, **kwargs
    ):
        """Load from pre-trained."""
        # this gets a fresh init model
        model = cls(config, *inputs, **kwargs)

        # Download the model file
        archive_file = hf_hub_download(
            repo_id=pretrained_checkpoint,
            filename="model.safetensors",

        )
        
        # Load the state_dict
        state_dict = load_file(archive_file)

        # add missing bert prefix
        state_dict = {f'bert.{key}': value for key, value in state_dict.items()}
        
        missing_keys, unexpected_keys = model.load_state_dict(state_dict, strict=False)

        if len(missing_keys) > 0:
            logger.warning(
                f"Found these missing keys in the checkpoint: {', '.join(missing_keys)}"
            )

            logger.warning(f"the number of which is equal to {len(missing_keys)}")

        if len(unexpected_keys) > 0:
            logger.warning(
                f"Found these unexpected keys in the checkpoint: {', '.join(unexpected_keys)}",
            )
            logger.warning(f"the number of which is equal to {len(unexpected_keys)}")

        return model

    def forward(
        self,
        input_ids: Optional[torch.Tensor] = None,
        attention_mask: Optional[torch.Tensor] = None,
        token_type_ids: Optional[torch.Tensor] = None,
        position_ids: Optional[torch.Tensor] = None,
        subset_mask: Optional[torch.Tensor] = None,
        output_all_encoded_layers: bool = True,
    ) -> torch.Tensor:

        embedding_output = self.bert.embeddings(input_ids, token_type_ids, position_ids)

        encoder_outputs_all = self.bert.encoder(
            embedding_output,
            attention_mask,
            output_all_encoded_layers=output_all_encoded_layers,
            subset_mask=subset_mask,
        )

        # batch_size, hidden_dim
        return encoder_outputs_all

class ClinicalMosaicForSequenceClassification(BertPreTrainedModel):
    """Bert Model transformer with a sequence classification/regression head.
    This head is just a linear layer on top of the pooled output.
    """
    config_class = BertConfig

    def __init__(self, config, **kwargs):
        super().__init__(config)
        self.num_labels = config.num_labels
        self.config = config
        self.bert = BertModel(config, add_pooling_layer=True)
        classifier_dropout = (
            config.classifier_dropout
            if config.classifier_dropout is not None
            else config.hidden_dropout_prob
        )
        self.dropout = nn.Dropout(classifier_dropout)
        self.classifier = nn.Linear(config.hidden_size, config.num_labels)

        # this resets the weights
        self.post_init()

    @classmethod
    def from_pretrained(
        cls, pretrained_checkpoint, state_dict=None, config=None, *inputs, **kwargs
    ):
        """Load from pre-trained."""
        # this gets a fresh init model
        model = cls(config, *inputs, **kwargs)
    
        # Download the model file
        archive_file = hf_hub_download(
            repo_id=pretrained_checkpoint,
            filename="model.safetensors",
        )
        
        # Load the state_dict
        state_dict = load_file(archive_file)

        # add missing bert prefix
        state_dict = {f'bert.{key}': value for key, value in state_dict.items()}
        
        missing_keys, unexpected_keys = model.load_state_dict(state_dict, strict=False)

        # Calculate classifier parameters
        num_classifier_params = config.hidden_size * config.num_labels + config.num_labels
        classifier_keys = {"classifier.weight", "classifier.bias", "bert.pooler.dense.weight", "bert.pooler.dense.bias"}
        
        # Check if only the classification layer is missing
        if set(missing_keys) == classifier_keys:
            print(
                f"Checkpoint does not contain the classification layer "
                f"({config.hidden_size}x{config.num_labels} + {config.num_labels} = {num_classifier_params} params). "
                "It will be randomly initialized."
            )
        elif len(missing_keys) > 0:
            logger.warning(
                f"Checkpoint is missing {len(missing_keys)} parameters, including possibly critical ones: "
                f"{', '.join(missing_keys)}"
            )

        if len(unexpected_keys) > 0:
            logger.warning(
                f"Found these unexpected keys in the checkpoint: {', '.join(unexpected_keys)}",
            )
            logger.warning(f"the number of which is equal to {len(unexpected_keys)}")

        return model

    def forward(
        self,
        input_ids: Optional[torch.Tensor] = None,
        attention_mask: Optional[torch.Tensor] = None,
        token_type_ids: Optional[torch.Tensor] = None,
        position_ids: Optional[torch.Tensor] = None,
        head_mask: Optional[torch.Tensor] = None,
        inputs_embeds: Optional[torch.Tensor] = None,
        labels: Optional[torch.Tensor] = None,
        output_attentions: Optional[bool] = None,
        output_hidden_states: Optional[bool] = None,
        return_dict: Optional[bool] = None,
    ) -> Union[Tuple[torch.Tensor], SequenceClassifierOutput]:

        return_dict = (
            return_dict if return_dict is not None else self.config.use_return_dict
        )

        outputs = self.bert(
            input_ids,
            attention_mask=attention_mask,
            token_type_ids=token_type_ids,
            position_ids=position_ids,
            head_mask=head_mask,
            inputs_embeds=inputs_embeds,
            output_attentions=output_attentions,
            output_hidden_states=output_hidden_states,
            return_dict=return_dict,
        )

        pooled_output = outputs[1]

        pooled_output = self.dropout(pooled_output)
        logits = self.classifier(pooled_output)

        loss = None
        if labels is not None:
            if self.config.problem_type is None:
                if self.num_labels == 1:
                    self.config.problem_type = "regression"
                elif self.num_labels > 1 and (
                    labels.dtype == torch.long or labels.dtype == torch.int
                ):
                    self.config.problem_type = "single_label_classification"
                else:
                    self.config.problem_type = "multi_label_classification"

            if self.config.problem_type == "regression":
                loss_fct = MSELoss()
                if self.num_labels == 1:
                    loss = loss_fct(logits.squeeze(), labels.squeeze())
                else:
                    loss = loss_fct(logits, labels)
            elif self.config.problem_type == "single_label_classification":
                loss_fct = CrossEntropyLoss()
                loss = loss_fct(logits.view(-1, self.num_labels), labels.view(-1))
            elif self.config.problem_type == "multi_label_classification":
                loss_fct = BCEWithLogitsLoss()
                loss = loss_fct(logits, labels)
        if not return_dict:
            output = (logits,) + outputs[2:]
            return ((loss,) + output) if loss is not None else output

        return SequenceClassifierOutput(
            loss=loss,
            logits=logits,
            hidden_states=None,
            attentions=None,
        )


class ClinicalMosaicForForMaskedLM(BertPreTrainedModel):
    config_class = BertConfig

    def __init__(self, config):
        super().__init__(config)

        if config.is_decoder:
            warnings.warn(
                'If you want to use `BertForMaskedLM` make sure `config.is_decoder=False` for '
                'bi-directional self-attention.')

        self.bert = BertModel(config, add_pooling_layer=False)
        self.cls = BertOnlyMLMHead(config,
                                   self.bert.embeddings.word_embeddings.weight)

        # Initialize weights and apply final processing
        self.post_init()

    def get_output_embeddings(self):
        return self.cls.predictions.decoder

    def set_output_embeddings(self, new_embeddings):
        self.cls.predictions.decoder = new_embeddings

    def forward(
        self,
        input_ids: Optional[torch.Tensor] = None,
        attention_mask: Optional[torch.Tensor] = None,
        token_type_ids: Optional[torch.Tensor] = None,
        position_ids: Optional[torch.Tensor] = None,
        head_mask: Optional[torch.Tensor] = None,
        inputs_embeds: Optional[torch.Tensor] = None,
        encoder_hidden_states: Optional[torch.Tensor] = None,
        encoder_attention_mask: Optional[torch.Tensor] = None,
        labels: Optional[torch.Tensor] = None,
        output_attentions: Optional[bool] = None,
        output_hidden_states: Optional[bool] = None,
        return_dict: Optional[bool] = None,
    ) -> Union[Tuple[torch.Tensor], MaskedLMOutput]:
        # labels should be a `torch.LongTensor` of shape
        # `(batch_size, sequence_length)`. These are used for computing the
        #  masked language modeling loss.
        #
        # Indices should be in `[-100, 0, ..., config.vocab_size]` (see
        # `input_ids` docstring) Tokens with indices set to `-100` are ignored
        # (masked), the loss is only computed for the tokens with labels in `[0,
        # ..., config.vocab_size]`
        #
        # Prediction scores are only computed for masked tokens and the (bs,
        # seqlen) dimensions are flattened
        if (input_ids is not None) == (inputs_embeds is not None):
            raise ValueError('Must specify either input_ids or input_embeds!')

        if labels is None:
            masked_tokens_mask = None
        else:
            masked_tokens_mask = labels > 0

        return_dict = return_dict if return_dict is not None else self.config.use_return_dict

        outputs = self.bert(
            input_ids,
            attention_mask=attention_mask,
            token_type_ids=token_type_ids,
            position_ids=position_ids,
            head_mask=head_mask,
            inputs_embeds=inputs_embeds,
            encoder_hidden_states=encoder_hidden_states,
            encoder_attention_mask=encoder_attention_mask,
            output_attentions=output_attentions,
            output_hidden_states=output_hidden_states,
            return_dict=return_dict,
            masked_tokens_mask=masked_tokens_mask,
        )
        
        sequence_output = outputs[0]
        prediction_scores = self.cls(sequence_output)

        loss = None
        if labels is not None:
            # Compute loss
            loss_fct = nn.CrossEntropyLoss()
            masked_token_idx = torch.nonzero(labels.flatten() > 0,
                                             as_tuple=False).flatten()
            loss = loss_fct(prediction_scores,
                            labels.flatten()[masked_token_idx])

            assert input_ids is not None, 'Coding error; please open an issue'
            batch, seqlen = input_ids.shape[:2]
            prediction_scores = rearrange(index_put_first_axis(
                prediction_scores, masked_token_idx, batch * seqlen),
                                          '(b s) d -> b s d',
                                          b=batch)

        if not return_dict:
            output = (prediction_scores,) + outputs[2:]
            return ((loss,) + output) if loss is not None else output

        return MaskedLMOutput(
            loss=loss,
            logits=prediction_scores,
            hidden_states=outputs[0],
            attentions=None,
        )

    def prepare_inputs_for_generation(self, input_ids: torch.Tensor,
                                      attention_mask: torch.Tensor,
                                      **model_kwargs):
        input_shape = input_ids.shape
        effective_batch_size = input_shape[0]

        #  add a dummy token
        if self.config.pad_token_id is None:
            raise ValueError('The PAD token should be defined for generation')

        attention_mask = torch.cat([
            attention_mask,
            attention_mask.new_zeros((attention_mask.shape[0], 1))
        ],
                                   dim=-1)
        dummy_token = torch.full((effective_batch_size, 1),
                                 self.config.pad_token_id,
                                 dtype=torch.long,
                                 device=input_ids.device)
        input_ids = torch.cat([input_ids, dummy_token], dim=1)

        return {'input_ids': input_ids, 'attention_mask': attention_mask}