# Copyright 2022 Cerebras Systems.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import torch
from modelzoo.transformers.pytorch.bert.bert_finetune_models import (
BertForSummarization,
BertForSummarizationLoss,
)
from modelzoo.transformers.pytorch.bert.utils import check_unused_model_params
[docs]class BertSummarizationModel(torch.nn.Module):
[docs] def __init__(self, params):
super().__init__()
model_params = params["model"].copy()
dropout_rate = model_params.pop("dropout_rate")
embedding_dropout_rate = model_params.pop(
"embedding_dropout_rate", dropout_rate
)
num_labels = 2
loss_weight = model_params.pop("loss_weight")
use_cls_bias = model_params.pop("use_cls_bias")
model_kwargs = {
"vocab_size": model_params.pop("vocab_size"),
"hidden_size": model_params.pop("hidden_size"),
"num_hidden_layers": model_params.pop("num_hidden_layers"),
"num_heads": model_params.pop("num_heads"),
"filter_size": model_params.pop("filter_size"),
"nonlinearity": model_params.pop("encoder_nonlinearity"),
"pooler_nonlinearity": model_params.pop(
"pooler_nonlinearity", None
),
"embedding_dropout_rate": embedding_dropout_rate,
"dropout_rate": dropout_rate,
"attention_dropout_rate": model_params.pop(
"attention_dropout_rate"
),
"max_position_embeddings": model_params.pop(
"max_position_embeddings"
),
"layer_norm_epsilon": float(model_params.pop("layer_norm_epsilon")),
}
self.model = BertForSummarization(
num_labels=num_labels,
loss_weight=loss_weight,
use_cls_bias=use_cls_bias,
**model_kwargs,
)
self.loss_fn = BertForSummarizationLoss(num_labels, loss_weight,)
self.compute_eval_metrics = model_params.pop(
"compute_eval_metrics", False
)
self.vocab_file = model_params.pop("vocab_file")
check_unused_model_params(model_params)
if self.compute_eval_metrics:
raise NotImplementedError(
"RougeScoreMetric not yet supported in weight streaming"
)
self.rouge1_score = RougeScoreMetric(
max_n=1, vocab_file=self.vocab_file, name="eval/rouge1"
)
self.rouge2_score = RougeScoreMetric(
max_n=2, vocab_file=self.vocab_file, name="eval/rouge2"
)
def forward(self, data):
logits = self.model(
input_ids=data["input_ids"],
attention_mask=data["attention_mask"],
token_type_ids=data["token_type_ids"],
cls_tokens_positions=data["cls_indices"],
)
loss = self.loss_fn(
logits, data["labels"], data["cls_weights"].clone().to(logits.dtype)
)
if not self.model.training and self.compute_eval_metrics:
labels = data["labels"].clone()
predictions = logits.argmax(-1).int()
input_ids = data["input_ids"].clone()
cls_indices = data["cls_indices"].clone()
cls_weights = data["cls_weights"].clone()
self.rouge1_score(
labels, predictions, cls_indices, cls_weights, input_ids,
)
self.rouge2_score(
labels, predictions, cls_indices, cls_weights, input_ids,
)
return loss