Spaces:
Sleeping
Sleeping
File size: 19,522 Bytes
7713b1f |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 |
import torch
import torch.nn
from torch.nn import CrossEntropyLoss
from transformers import BertPreTrainedModel, BertModel, RobertaPreTrainedModel, RobertaModel
from transformers.modeling_outputs import QuestionAnsweringModelOutput
from model.prefix_encoder import PrefixEncoder
from model.deberta import DebertaPreTrainedModel, DebertaModel
class BertForQuestionAnswering(BertPreTrainedModel):
_keys_to_ignore_on_load_unexpected = [r"pooler"]
def __init__(self, config):
super().__init__(config)
self.num_labels = config.num_labels
self.bert = BertModel(config, add_pooling_layer=False)
self.qa_outputs = torch.nn.Linear(config.hidden_size, config.num_labels)
for param in self.bert.parameters():
param.requires_grad = False
self.init_weights()
def forward(
self,
input_ids=None,
attention_mask=None,
token_type_ids=None,
position_ids=None,
head_mask=None,
inputs_embeds=None,
start_positions=None,
end_positions=None,
output_attentions=None,
output_hidden_states=None,
return_dict=None,
):
r"""
start_positions (:obj:`torch.LongTensor` of shape :obj:`(batch_size,)`, `optional`):
Labels for position (index) of the start of the labelled span for computing the token classification loss.
Positions are clamped to the length of the sequence (:obj:`sequence_length`). Position outside of the
sequence are not taken into account for computing the loss.
end_positions (:obj:`torch.LongTensor` of shape :obj:`(batch_size,)`, `optional`):
Labels for position (index) of the end of the labelled span for computing the token classification loss.
Positions are clamped to the length of the sequence (:obj:`sequence_length`). Position outside of the
sequence are not taken into account for computing the loss.
"""
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,
)
sequence_output = outputs[0]
logits = self.qa_outputs(sequence_output)
start_logits, end_logits = logits.split(1, dim=-1)
start_logits = start_logits.squeeze(-1).contiguous()
end_logits = end_logits.squeeze(-1).contiguous()
total_loss = None
if start_positions is not None and end_positions is not None:
# If we are on multi-GPU, split add a dimension
if len(start_positions.size()) > 1:
start_positions = start_positions.squeeze(-1)
if len(end_positions.size()) > 1:
end_positions = end_positions.squeeze(-1)
# sometimes the start/end positions are outside our model inputs, we ignore these terms
ignored_index = start_logits.size(1)
start_positions = start_positions.clamp(0, ignored_index)
end_positions = end_positions.clamp(0, ignored_index)
loss_fct = CrossEntropyLoss(ignore_index=ignored_index)
start_loss = loss_fct(start_logits, start_positions)
end_loss = loss_fct(end_logits, end_positions)
total_loss = (start_loss + end_loss) / 2
if not return_dict:
output = (start_logits, end_logits) + outputs[2:]
return ((total_loss,) + output) if total_loss is not None else output
return QuestionAnsweringModelOutput(
loss=total_loss,
start_logits=start_logits,
end_logits=end_logits,
hidden_states=outputs.hidden_states,
attentions=outputs.attentions,
)
class BertPrefixForQuestionAnswering(BertPreTrainedModel):
def __init__(self, config):
super().__init__(config)
self.num_labels = config.num_labels
self.pre_seq_len = config.pre_seq_len
self.n_layer = config.num_hidden_layers
self.n_head = config.num_attention_heads
self.n_embd = config.hidden_size // config.num_attention_heads
self.bert = BertModel(config, add_pooling_layer=False)
self.qa_outputs = torch.nn.Linear(config.hidden_size, config.num_labels)
self.dropout = torch.nn.Dropout(config.hidden_dropout_prob)
self.prefix_encoder = PrefixEncoder(config)
self.prefix_tokens = torch.arange(self.pre_seq_len).long()
for param in self.bert.parameters():
param.requires_grad = False
self.init_weights()
def get_prompt(self, batch_size):
prefix_tokens = self.prefix_tokens.unsqueeze(0).expand(batch_size, -1).to(self.bert.device)
past_key_values = self.prefix_encoder(prefix_tokens)
bsz, seqlen, _ = past_key_values.shape
past_key_values = past_key_values.view(
bsz,
seqlen,
self.n_layer * 2,
self.n_head,
self.n_embd
)
past_key_values = self.dropout(past_key_values)
past_key_values = past_key_values.permute([2, 0, 3, 1, 4]).split(2)
return past_key_values
def forward(
self,
input_ids=None,
attention_mask=None,
token_type_ids=None,
position_ids=None,
head_mask=None,
inputs_embeds=None,
start_positions=None,
end_positions=None,
output_attentions=None,
output_hidden_states=None,
return_dict=None,
):
r"""
start_positions (:obj:`torch.LongTensor` of shape :obj:`(batch_size,)`, `optional`):
Labels for position (index) of the start of the labelled span for computing the token classification loss.
Positions are clamped to the length of the sequence (:obj:`sequence_length`). Position outside of the
sequence are not taken into account for computing the loss.
end_positions (:obj:`torch.LongTensor` of shape :obj:`(batch_size,)`, `optional`):
Labels for position (index) of the end of the labelled span for computing the token classification loss.
Positions are clamped to the length of the sequence (:obj:`sequence_length`). Position outside of the
sequence are not taken into account for computing the loss.
"""
return_dict = return_dict if return_dict is not None else self.config.use_return_dict
batch_size = input_ids.shape[0]
past_key_values = self.get_prompt(batch_size=batch_size)
prefix_attention_mask = torch.ones(batch_size, self.pre_seq_len).to(self.bert.device)
attention_mask = torch.cat((prefix_attention_mask, attention_mask), dim=1)
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,
past_key_values=past_key_values,
)
sequence_output = outputs[0]
logits = self.qa_outputs(sequence_output)
start_logits, end_logits = logits.split(1, dim=-1)
start_logits = start_logits.squeeze(-1).contiguous()
end_logits = end_logits.squeeze(-1).contiguous()
total_loss = None
if start_positions is not None and end_positions is not None:
# If we are on multi-GPU, split add a dimension
if len(start_positions.size()) > 1:
start_positions = start_positions.squeeze(-1)
if len(end_positions.size()) > 1:
end_positions = end_positions.squeeze(-1)
# sometimes the start/end positions are outside our model inputs, we ignore these terms
ignored_index = start_logits.size(1)
start_positions = start_positions.clamp(0, ignored_index)
end_positions = end_positions.clamp(0, ignored_index)
loss_fct = CrossEntropyLoss(ignore_index=ignored_index)
start_loss = loss_fct(start_logits, start_positions)
end_loss = loss_fct(end_logits, end_positions)
total_loss = (start_loss + end_loss) / 2
if not return_dict:
output = (start_logits, end_logits) + outputs[2:]
return ((total_loss,) + output) if total_loss is not None else output
return QuestionAnsweringModelOutput(
loss=total_loss,
start_logits=start_logits,
end_logits=end_logits,
hidden_states=outputs.hidden_states,
attentions=outputs.attentions,
)
class RobertaPrefixModelForQuestionAnswering(RobertaPreTrainedModel):
def __init__(self, config):
super().__init__(config)
self.num_labels = config.num_labels
self.pre_seq_len = config.pre_seq_len
self.n_layer = config.num_hidden_layers
self.n_head = config.num_attention_heads
self.n_embd = config.hidden_size // config.num_attention_heads
self.roberta = RobertaModel(config, add_pooling_layer=False)
self.qa_outputs = torch.nn.Linear(config.hidden_size, config.num_labels)
self.init_weights()
self.dropout = torch.nn.Dropout(config.hidden_dropout_prob)
self.prefix_encoder = PrefixEncoder(config)
self.prefix_tokens = torch.arange(self.pre_seq_len).long()
for param in self.roberta.parameters():
param.requires_grad = False
def get_prompt(self, batch_size):
prefix_tokens = self.prefix_tokens.unsqueeze(0).expand(batch_size, -1).to(self.roberta.device)
past_key_values = self.prefix_encoder(prefix_tokens)
bsz, seqlen, _ = past_key_values.shape
past_key_values = past_key_values.view(
bsz,
seqlen,
self.n_layer * 2,
self.n_head,
self.n_embd
)
past_key_values = self.dropout(past_key_values)
past_key_values = past_key_values.permute([2, 0, 3, 1, 4]).split(2)
return past_key_values
def forward(
self,
input_ids=None,
attention_mask=None,
token_type_ids=None,
position_ids=None,
head_mask=None,
inputs_embeds=None,
start_positions=None,
end_positions=None,
output_attentions=None,
output_hidden_states=None,
return_dict=None,
):
r"""
start_positions (:obj:`torch.LongTensor` of shape :obj:`(batch_size,)`, `optional`):
Labels for position (index) of the start of the labelled span for computing the token classification loss.
Positions are clamped to the length of the sequence (:obj:`sequence_length`). Position outside of the
sequence are not taken into account for computing the loss.
end_positions (:obj:`torch.LongTensor` of shape :obj:`(batch_size,)`, `optional`):
Labels for position (index) of the end of the labelled span for computing the token classification loss.
Positions are clamped to the length of the sequence (:obj:`sequence_length`). Position outside of the
sequence are not taken into account for computing the loss.
"""
return_dict = return_dict if return_dict is not None else self.config.use_return_dict
batch_size = input_ids.shape[0]
past_key_values = self.get_prompt(batch_size=batch_size)
prefix_attention_mask = torch.ones(batch_size, self.pre_seq_len).to(self.roberta.device)
attention_mask = torch.cat((prefix_attention_mask, attention_mask), dim=1)
outputs = self.roberta(
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,
past_key_values=past_key_values,
)
sequence_output = outputs[0]
logits = self.qa_outputs(sequence_output)
start_logits, end_logits = logits.split(1, dim=-1)
start_logits = start_logits.squeeze(-1).contiguous()
end_logits = end_logits.squeeze(-1).contiguous()
total_loss = None
if start_positions is not None and end_positions is not None:
# If we are on multi-GPU, split add a dimension
if len(start_positions.size()) > 1:
start_positions = start_positions.squeeze(-1)
if len(end_positions.size()) > 1:
end_positions = end_positions.squeeze(-1)
# sometimes the start/end positions are outside our model inputs, we ignore these terms
ignored_index = start_logits.size(1)
start_positions = start_positions.clamp(0, ignored_index)
end_positions = end_positions.clamp(0, ignored_index)
loss_fct = CrossEntropyLoss(ignore_index=ignored_index)
start_loss = loss_fct(start_logits, start_positions)
end_loss = loss_fct(end_logits, end_positions)
total_loss = (start_loss + end_loss) / 2
if not return_dict:
output = (start_logits, end_logits) + outputs[2:]
return ((total_loss,) + output) if total_loss is not None else output
return QuestionAnsweringModelOutput(
loss=total_loss,
start_logits=start_logits,
end_logits=end_logits,
hidden_states=outputs.hidden_states,
attentions=outputs.attentions,
)
class DebertaPrefixModelForQuestionAnswering(DebertaPreTrainedModel):
def __init__(self, config):
super().__init__(config)
self.num_labels = config.num_labels
self.deberta = DebertaModel(config)
self.dropout = torch.nn.Dropout(config.hidden_dropout_prob)
self.qa_outputs = torch.nn.Linear(config.hidden_size, config.num_labels)
self.init_weights()
for param in self.deberta.parameters():
param.requires_grad = False
self.pre_seq_len = config.pre_seq_len
self.n_layer = config.num_hidden_layers
self.n_head = config.num_attention_heads
self.n_embd = config.hidden_size // config.num_attention_heads
# Use a two layered MLP to encode the prefix
self.prefix_tokens = torch.arange(self.pre_seq_len).long()
self.prefix_encoder = PrefixEncoder(config)
deberta_param = 0
for name, param in self.deberta.named_parameters():
deberta_param += param.numel()
all_param = 0
for name, param in self.named_parameters():
all_param += param.numel()
total_param = all_param - deberta_param
print('total param is {}'.format(total_param)) # 9860105
def get_prompt(self, batch_size):
prefix_tokens = self.prefix_tokens.unsqueeze(0).expand(batch_size, -1).to(self.deberta.device)
past_key_values = self.prefix_encoder(prefix_tokens)
# bsz, seqlen, _ = past_key_values.shape
past_key_values = past_key_values.view(
batch_size,
self.pre_seq_len,
self.n_layer * 2,
self.n_head,
self.n_embd
)
past_key_values = self.dropout(past_key_values)
past_key_values = past_key_values.permute([2, 0, 3, 1, 4]).split(2)
return past_key_values
def forward(
self,
input_ids=None,
attention_mask=None,
token_type_ids=None,
position_ids=None,
# head_mask=None,
inputs_embeds=None,
start_positions=None,
end_positions=None,
output_attentions=None,
output_hidden_states=None,
return_dict=None,
):
r"""
start_positions (:obj:`torch.LongTensor` of shape :obj:`(batch_size,)`, `optional`):
Labels for position (index) of the start of the labelled span for computing the token classification loss.
Positions are clamped to the length of the sequence (:obj:`sequence_length`). Position outside of the
sequence are not taken into account for computing the loss.
end_positions (:obj:`torch.LongTensor` of shape :obj:`(batch_size,)`, `optional`):
Labels for position (index) of the end of the labelled span for computing the token classification loss.
Positions are clamped to the length of the sequence (:obj:`sequence_length`). Position outside of the
sequence are not taken into account for computing the loss.
"""
return_dict = return_dict if return_dict is not None else self.config.use_return_dict
batch_size = input_ids.shape[0]
past_key_values = self.get_prompt(batch_size=batch_size)
prefix_attention_mask = torch.ones(batch_size, self.pre_seq_len).to(self.deberta.device)
attention_mask = torch.cat((prefix_attention_mask, attention_mask), dim=1)
outputs = self.deberta(
input_ids,
attention_mask=attention_mask,
token_type_ids=token_type_ids,
position_ids=position_ids,
inputs_embeds=inputs_embeds,
output_attentions=output_attentions,
output_hidden_states=output_hidden_states,
return_dict=return_dict,
past_key_values=past_key_values,
)
sequence_output = outputs[0]
logits = self.qa_outputs(sequence_output)
start_logits, end_logits = logits.split(1, dim=-1)
start_logits = start_logits.squeeze(-1).contiguous()
end_logits = end_logits.squeeze(-1).contiguous()
total_loss = None
if start_positions is not None and end_positions is not None:
# If we are on multi-GPU, split add a dimension
if len(start_positions.size()) > 1:
start_positions = start_positions.squeeze(-1)
if len(end_positions.size()) > 1:
end_positions = end_positions.squeeze(-1)
# sometimes the start/end positions are outside our model inputs, we ignore these terms
ignored_index = start_logits.size(1)
start_positions = start_positions.clamp(0, ignored_index)
end_positions = end_positions.clamp(0, ignored_index)
loss_fct = CrossEntropyLoss(ignore_index=ignored_index)
start_loss = loss_fct(start_logits, start_positions)
end_loss = loss_fct(end_logits, end_positions)
total_loss = (start_loss + end_loss) / 2
if not return_dict:
output = (start_logits, end_logits) + outputs[2:]
return ((total_loss,) + output) if total_loss is not None else output
return QuestionAnsweringModelOutput(
loss=total_loss,
start_logits=start_logits,
end_logits=end_logits,
hidden_states=outputs.hidden_states,
attentions=outputs.attentions,
) |