Source code for ofasys.engine.criterion.scst_loss

# Copyright 2022 The OFA-Sys Team. All rights reserved.
# This source code is licensed under the Apache 2.0 license
# found in the LICENSE file in the root directory.

import copy
import math
import string
from collections import OrderedDict
from dataclasses import dataclass, field
from typing import Optional

import torch

from ofasys.configure import register_config
from ofasys.logging import metrics
from ofasys.metric.pyciderevalcap.ciderD.ciderD import CiderD
from ofasys.module import utils
from ofasys.preprocessor.utils import collate_tokens
from ofasys.utils.file_utils import cached_path

from .base import BaseCriterion, CriterionConfig


def scst_loss(lprobs, target, reward, ignore_index=None, reduce=True):
    loss = -lprobs.gather(dim=-1, index=target.unsqueeze(-1)).squeeze()
    loss *= reward.unsqueeze(-1)
    if ignore_index is not None:
        pad_mask = target.eq(ignore_index)
        loss.masked_fill_(pad_mask, 0.0)
        ntokens = (~pad_mask).sum()
    else:
        loss = loss.squeeze(-1)
        ntokens = target.numel()
    if reduce:
        loss = loss.sum()
    return loss, ntokens


[docs]@dataclass class ScstRewardCriterionConfig(CriterionConfig): scst_cider_cached_tokens: str = field( default="coco-train-words.p", metadata={"help": "path to cached cPickle file used to calculate CIDEr scores"}, ) ignore_prefix_size: int = field( default=0, metadata={"help": "Ignore first N tokens"}, ) sentence_avg: bool = field( default=False, metadata={ "help": "normalize gradients by the number of sentences in a batch" " (default is to normalize by number of tokens)" }, ) constraint_range: Optional[str] = field(default=None, metadata={"help": "constraint range"})
[docs]@register_config("ofasys.criterion", "scst_reward_criterion", ScstRewardCriterionConfig) class ScstRewardCriterion(BaseCriterion): """This criterion computes Self-critical Sequence Training (SCST) loss, which is proposed for image captioning. Details for Self-critical Sequence Training for Image Captioning (https://arxiv.org/abs/1612.00563) """ CIDER_REWARD_WEIGHT = 1 def __init__(self, task, cfg: CriterionConfig): super().__init__(task, cfg) local_path = cached_path(cfg.scst_cider_cached_tokens) self.scst_cider_scorer = CiderD(df=local_path) self.sentence_avg = cfg.sentence_avg self.ignore_prefix_size = cfg.ignore_prefix_size self.transtab = str.maketrans({key: None for key in string.punctuation}) self.constraint_start = None self.constraint_end = None if cfg.constraint_range is not None: constraint_start, constraint_end = cfg.constraint_range.split(',') self.constraint_start = int(constraint_start) self.constraint_end = int(constraint_end)
[docs] def forward(self, model, sample, update_num=0, reduce=True): """Compute the loss for the given sample. Returns a tuple with three elements: 1) the loss 2) the sample size, which is used as the denominator for the gradient 3) logging outputs to display while training """ sample_copy = copy.deepcopy(sample) loss, score, ntokens, nsentences = self.compute_loss(model, sample_copy, reduce=reduce) sample_size = nsentences if self.sentence_avg else ntokens logging_output = { "loss": loss.data, "score": score, "ntokens": ntokens, "nsentences": nsentences, "sample_size": sample_size, } return loss * self.weight, sample_size, logging_output
def _calculate_eval_scores(self, gen_res, gt_idx, gt_res): ''' gen_res: generated captions, list of str gt_idx: list of int, of the same length as gen_res gt_res: ground truth captions, list of list of str. gen_res[i] corresponds to gt_res[gt_idx[i]] Each image can have multiple ground truth captions ''' gen_res_size = len(gen_res) res = OrderedDict() for i in range(gen_res_size): res[i] = [self._wrap_sentence(gen_res[i].strip().translate(self.transtab))] gts = OrderedDict() gt_res_ = [ [self._wrap_sentence(gt_res[i][j].strip().translate(self.transtab)) for j in range(len(gt_res[i]))] for i in range(len(gt_res)) ] for i in range(gen_res_size): gts[i] = gt_res_[gt_idx[i]] res_ = [{'image_id': i, 'caption': res[i]} for i in range(len(res))] _, batch_cider_scores = self.scst_cider_scorer.compute_score(gts, res_) scores = self.CIDER_REWARD_WEIGHT * batch_cider_scores return scores @classmethod def _wrap_sentence(self, s): # ensure the sentence ends with <eos> token # in order to keep consisitent with cider_cached_tokens r = s.strip() if r.endswith('.'): r = r[:-1] r += ' <eos>' return r def get_generator_out(self, model, sample): def decode(toks): hypo = toks.int().cpu() hypo_str = self.task.tgt_dict.string(hypo) hypo_str = self.task.bpe.decode(hypo_str).strip() return hypo, hypo_str model.eval() from ofasys.preprocessor import Slot with torch.no_grad(): self.task.scst_generator.model.eval() gen_out = self.task.scst_generator.generate(model, sample) gen_target = [] gen_res = [] gt_res = [] for i in range(len(gen_out)): for j in range(len(gen_out[i])): hypo, hypo_str = decode(gen_out[i][j]["tokens"]) gen_target.append(hypo) gen_res.append(hypo_str) gt_res.append(decode(utils.strip_pad(sample["target"][i], self.padding_idx))[1].split('&&')) return gen_target, gen_res, gt_res def get_reward_and_scores(self, gen_res, gt_res, device): batch_size = len(gt_res) gen_res_size = len(gen_res) seq_per_img = gen_res_size // batch_size gt_idx = [i // seq_per_img for i in range(gen_res_size)] scores = self._calculate_eval_scores(gen_res, gt_idx, gt_res) sc_ = scores.reshape(batch_size, seq_per_img) baseline = (sc_.sum(1, keepdims=True) - sc_) / (sc_.shape[1] - 1) # sample - baseline reward = scores.reshape(batch_size, seq_per_img) reward = reward - baseline reward = reward.reshape(gen_res_size) reward = torch.as_tensor(reward, device=device, dtype=torch.float64) return reward, scores def get_net_output(self, model, sample, gen_target): def merge(sample_list, eos=self.task.tgt_dict.eos(), move_eos_to_beginning=False): return collate_tokens( sample_list, pad_idx=self.padding_idx, eos_idx=eos, left_pad=False, move_eos_to_beginning=move_eos_to_beginning, ) batch_size = len(sample["target"]) gen_target_size = len(gen_target) seq_per_img = gen_target_size // batch_size model.train() for slot in sample['net_input']['slots']: if slot.is_src: slot.value = torch.repeat_interleave(slot.value, seq_per_img, dim=0) else: assert slot.modality.value == 1 slot.value = torch.as_tensor( merge(gen_target, eos=self.task.tgt_dict.bos(), move_eos_to_beginning=True), device=sample["target"].device, dtype=torch.int64, ) gen_target_tokens = torch.as_tensor(merge(gen_target), device=sample["target"].device, dtype=torch.int64) net_output = model(**sample["net_input"]) return net_output, gen_target_tokens def get_lprobs_and_target(self, model, net_output, gen_target): if self.constraint_start is not None and self.constraint_end is not None: net_output[0][:, :, 4 : self.constraint_start] = -math.inf net_output[0][:, :, self.constraint_end :] = -math.inf lprobs = model.get_normalized_probs(net_output, log_probs=True) if self.ignore_prefix_size > 0: if getattr(lprobs, "batch_first", False): lprobs = lprobs[:, self.ignore_prefix_size :, :].contiguous() gen_target = gen_target[:, self.ignore_prefix_size :].contiguous() else: lprobs = lprobs[self.ignore_prefix_size :, :, :].contiguous() gen_target = gen_target[self.ignore_prefix_size :, :].contiguous() return lprobs, gen_target def compute_loss(self, model, sample, reduce=True): gen_target, gen_res, gt_res = self.get_generator_out(model, sample) reward, scores = self.get_reward_and_scores(gen_res, gt_res, device=sample["target"].device) net_output, gen_target_tokens = self.get_net_output(model, sample, gen_target) gen_lprobs, gen_target_tokens = self.get_lprobs_and_target(model, net_output, gen_target_tokens) loss, ntokens = scst_loss(gen_lprobs, gen_target_tokens, reward, ignore_index=self.padding_idx, reduce=reduce) nsentences = gen_target_tokens.size(0) return loss, scores.sum(), ntokens, nsentences
[docs] @classmethod def reduce_metrics(cls, logging_outputs, prefix_name=None) -> None: """Aggregate logging outputs from data parallel training.""" task_name = prefix_name + '/' if prefix_name else '' loss_sum = sum(log.get("loss", 0) for log in logging_outputs) score_sum = sum(log.get("score", 0) for log in logging_outputs) ntokens = sum(log.get("ntokens", 0) for log in logging_outputs) nsentences = sum(log.get("nsentences", 0) for log in logging_outputs) sample_size = sum(log.get("sample_size", 0) for log in logging_outputs) metrics.log_scalar("loss", loss_sum / sample_size, sample_size, priority=0, round=3) metrics.log_scalar(f"{task_name}loss", loss_sum / sample_size, sample_size, round=3) metrics.log_scalar(f"{task_name}score", score_sum / nsentences, nsentences, round=3) metrics.log_scalar(f"{task_name}ntokens", ntokens, 1, round=3) metrics.log_scalar(f"{task_name}nsentences", nsentences, 1, round=3) metrics.log_scalar(f"{task_name}sample_size", sample_size, 1, round=3)
[docs] @staticmethod def logging_outputs_can_be_summed() -> bool: """ Whether the logging outputs returned by `forward` can be summed across workers prior to calling `reduce_metrics`. Setting this to True will improve distributed training speed. """ return True