#!/usr/bin/env python import argparse import glob import logging import os import sys import time from collections import defaultdict from pathlib import Path from typing import Dict, List, Tuple import pdb import numpy as np import pytorch_lightning as pl import torch from torch.utils.data import DataLoader from pytorch_lightning.utilities import rank_zero_info from utils.callbacks import Seq2SeqLoggingCallback, get_checkpoint_callback, get_early_stopping_callback from transformers import MBartTokenizer, T5ForConditionalGeneration from transformers.models.bart.modeling_bart import shift_tokens_right from utils.utils_verbalisation_module import ( ROUGE_KEYS, LegacySeq2SeqDataset, Seq2SeqDataset, assert_all_frozen, calculate_bleu, calculate_rouge, flatten_list, freeze_embeds, freeze_params, label_smoothed_nll_loss, lmap, pickle_save, save_json, use_task_specific_params, ) from utils.utils_graph2text import convert_text, eval_meteor, eval_bleu, eval_chrf, eval_meteor_test_webnlg, eval_chrf_test_webnlg # need the parent dir module sys.path.insert(2, str(Path(__file__).resolve().parents[1])) from utils.lightning_base import BaseTransformer, add_generic_args, generic_train # noqa logger = logging.getLogger(__name__) class SummarizationModule(BaseTransformer): mode = "summarization" loss_names = ["loss"] metric_names = ROUGE_KEYS default_val_metric = "rouge2" def __init__(self, hparams, **kwargs): if hparams.sortish_sampler and hparams.gpus > 1: hparams.replace_sampler_ddp = False elif hparams.max_tokens_per_batch is not None: if hparams.gpus > 1: raise NotImplementedError("Dynamic Batch size does not work for multi-gpu training") if hparams.sortish_sampler: raise ValueError("--sortish_sampler and --max_tokens_per_batch may not be used simultaneously") super().__init__(hparams, num_labels=None, mode=self.mode, **kwargs) #use_task_specific_params(self.model, "summarization") self.metrics_save_path = Path('base') / "metrics.json" self.hparams_save_path = Path('base') / "hparams.pkl" pickle_save(self.hparams, self.hparams_save_path) self.step_count = -2 self.metrics = defaultdict(list) self.model_type = self.config.model_type self.vocab_size = self.config.tgt_vocab_size if self.model_type == "fsmt" else self.config.vocab_size if 't5' in hparams.model_name_or_path: self.model.config.prefix = 'translate Graph to English: ' self.dataset_kwargs: dict = dict( data_dir=self.hparams.data_dir, max_source_length=self.hparams.max_source_length, prefix=self.model.config.prefix or "", ) n_observations_per_split = { "train": self.hparams.n_train, "val": self.hparams.n_val, "test_seen": self.hparams.n_test, "test_unseen": self.hparams.n_test, "test_both": self.hparams.n_test, } self.n_obs = {k: v if v >= 0 else None for k, v in n_observations_per_split.items()} self.target_lens = { "train": self.hparams.max_target_length, "val": self.hparams.val_max_target_length, "test_seen": self.hparams.test_max_target_length, "test_unseen": self.hparams.test_max_target_length, "test_both": self.hparams.test_max_target_length, } assert self.target_lens["train"] <= self.target_lens["val"], f"target_lens: {self.target_lens}" assert self.target_lens["train"] <= self.target_lens["test_both"], f"target_lens: {self.target_lens}" if self.hparams.freeze_embeds: freeze_embeds(self.model) if self.hparams.freeze_encoder: freeze_params(self.model.get_encoder()) assert_all_frozen(self.model.get_encoder()) self.num_workers = hparams.num_workers self.decoder_start_token_id = None # default to config if self.model.config.decoder_start_token_id is None and isinstance(self.tokenizer, MBartTokenizer): self.decoder_start_token_id = self.tokenizer.lang_code_to_id[hparams.tgt_lang] self.model.config.decoder_start_token_id = self.decoder_start_token_id self.dataset_class = ( Seq2SeqDataset if hasattr(self.tokenizer, "prepare_seq2seq_batch") else LegacySeq2SeqDataset ) self.already_saved_batch = False self.eval_beams = self.model.config.num_beams if self.hparams.eval_beams is None else self.hparams.eval_beams if self.hparams.eval_max_gen_length is not None: self.eval_max_length = self.hparams.eval_max_gen_length else: self.eval_max_length = self.model.config.max_length self.val_metric = self.default_val_metric if self.hparams.val_metric is None else self.hparams.val_metric def save_readable_batch(self, batch: Dict[str, torch.Tensor]) -> Dict[str, List[str]]: """A debugging utility""" readable_batch = { k: self.tokenizer.batch_decode(v.tolist()) if "mask" not in k else v.shape for k, v in batch.items() } save_json(readable_batch, Path(self.output_dir) / "text_batch.json") tb = {} for k, v in batch.items(): tb[k] = v.tolist() save_json(tb, Path(self.output_dir) / "tok_batch.json") self.already_saved_batch = True return readable_batch def forward(self, input_ids, **kwargs): return self.model(input_ids, **kwargs) def ids_to_clean_text(self, generated_ids: List[int]): gen_text = self.tokenizer.batch_decode( generated_ids, skip_special_tokens=True, clean_up_tokenization_spaces=True ) return lmap(str.strip, gen_text) def _step(self, batch: dict) -> Tuple: pad_token_id = self.tokenizer.pad_token_id src_ids, src_mask = batch["input_ids"], batch["attention_mask"] if isinstance(self.model, T5ForConditionalGeneration): tgt_ids = batch["labels"] decoder_input_ids = self.model._shift_right(tgt_ids) else: #decoder_input_ids = shift_tokens_right(tgt_ids, pad_token_id) y = batch["labels"] decoder_input_ids = y[:, :-1].contiguous() tgt_ids = y[:, 1:].clone() if not self.already_saved_batch: # This would be slightly better if it only happened on rank zero batch["decoder_input_ids"] = decoder_input_ids self.save_readable_batch(batch) outputs = self(src_ids, attention_mask=src_mask, decoder_input_ids=decoder_input_ids, use_cache=False) lm_logits = outputs[0] if self.hparams.label_smoothing == 0: # Same behavior as modeling_bart.py, besides ignoring pad_token_id ce_loss_fct = torch.nn.CrossEntropyLoss(ignore_index=pad_token_id) assert lm_logits.shape[-1] == self.vocab_size loss = ce_loss_fct(lm_logits.view(-1, lm_logits.shape[-1]), tgt_ids.view(-1)) else: lprobs = torch.nn.functional.log_softmax(lm_logits, dim=-1) loss, nll_loss = label_smoothed_nll_loss( lprobs, tgt_ids, self.hparams.label_smoothing, ignore_index=pad_token_id ) return (loss,) @property def pad(self) -> int: return self.tokenizer.pad_token_id def training_step(self, batch, batch_idx) -> Dict: loss_tensors = self._step(batch) logs = {name: loss for name, loss in zip(self.loss_names, loss_tensors)} # tokens per batch logs["tpb"] = batch["input_ids"].ne(self.pad).sum() + batch["labels"].ne(self.pad).sum() logs["bs"] = batch["input_ids"].shape[0] logs["src_pad_tok"] = batch["input_ids"].eq(self.pad).sum() logs["src_pad_frac"] = batch["input_ids"].eq(self.pad).float().mean() # TODO(SS): make a wandb summary metric for this return {"loss": loss_tensors[0], "log": logs} def validation_step(self, batch, batch_idx) -> Dict: return self._generative_step(batch) def validation_epoch_end(self, outputs, prefix="val") -> Dict: self.step_count += 1 val_outputs_folder = "val_outputs" os.system("mkdir -p " + os.path.join(self.hparams.output_dir, val_outputs_folder)) if prefix == "val": output_test_predictions_file = os.path.join(self.hparams.output_dir, val_outputs_folder, "validation_predictions_" + str(self.step_count) + ".txt") output_test_targets_file = os.path.join(self.hparams.output_dir, val_outputs_folder, "validation_targets_" + str(self.step_count) + ".txt") # write predictions and targets for later rouge evaluation. with open(output_test_predictions_file, "w") as p_writer, open(output_test_targets_file, "w") as t_writer: for output_batch in outputs: p_writer.writelines(convert_text(s) + "\n" for s in output_batch["preds"]) t_writer.writelines(convert_text(s) + "\n" for s in output_batch["target"]) p_writer.close() t_writer.close() bleu_info = eval_bleu(self.hparams.data_dir, output_test_predictions_file, 'val') rank_zero_info("%s bleu_info: %s", self.step_count, bleu_info) if bleu_info == -1: bleu_info = float(bleu_info) else: bleu_info = float(bleu_info.split(",")[0].split("BLEU = ")[1]) losses = {k: torch.stack([x[k] for x in outputs]).mean() for k in self.loss_names} loss = losses["loss"] generative_metrics = { k: np.array([x[k] for x in outputs]).mean() for k in self.metric_names + ["gen_time", "gen_len"] } generative_metrics['bleu'] = bleu_info metric_val = ( generative_metrics[self.val_metric] if self.val_metric in generative_metrics else losses[ self.val_metric] ) metric_tensor: torch.FloatTensor = torch.tensor(metric_val).type_as(loss) generative_metrics.update({k: v.item() for k, v in losses.items()}) losses.update(generative_metrics) all_metrics = {f"{prefix}_avg_{k}": x for k, x in losses.items()} all_metrics["step_count"] = self.step_count self.metrics[prefix].append(all_metrics) # callback writes this to self.metrics_save_path preds = flatten_list([x["preds"] for x in outputs]) return { "bleu": bleu_info, "log": all_metrics, "preds": preds, f"{prefix}_loss": loss, f"{prefix}_{self.val_metric}": metric_tensor, } else: data_logs = {} for output in outputs: dataset_idx = output[0]['dataloader_idx'] if dataset_idx == 0: dataset_name = 'test_both' elif dataset_idx == 1: dataset_name = 'test_seen' else: dataset_name = 'test_unseen' if output[0]['bleu'] == -1: bleu_info = float(output[0]['bleu']) else: bleu_info = float(output[0]['bleu'].split(",")[0].split("BLEU = ")[1]) losses = {k: torch.stack([x[k] for x in output]).mean() for k in self.loss_names} loss = losses["loss"] generative_metrics = { k: np.array([x[k] for x in output]).mean() for k in self.metric_names + ["gen_time", "gen_len"] } generative_metrics['bleu'] = bleu_info metric_val = ( generative_metrics[self.val_metric] if self.val_metric in generative_metrics else losses[ self.val_metric] ) metric_tensor: torch.FloatTensor = torch.tensor(metric_val).type_as(loss) generative_metrics.update({k: v.item() for k, v in losses.items()}) losses.update(generative_metrics) all_metrics = {f"{prefix}_avg_{k}": x for k, x in losses.items()} all_metrics["step_count"] = self.step_count self.metrics[prefix].append(all_metrics) # callback writes this to self.metrics_save_path preds = flatten_list([x["preds"] for x in output]) data_logs.update({ "log" + "_" + dataset_name: all_metrics, "preds" + "_" + dataset_name: preds, f"{prefix}_loss" + "_" + dataset_name: loss, f"{prefix}_{self.val_metric}" + "_" + dataset_name: metric_tensor, }) return data_logs ####### def calc_generative_metrics(self, preds, target) -> Dict: return calculate_rouge(preds, target) def _generative_step(self, batch: dict, batch_idx=None, dataloader_idx=None) -> dict: t0 = time.time() # parser.add_argument('--eval_max_gen_length', type=int, default=None, help='never generate more than n tokens') generated_ids = self.model.generate( batch["input_ids"], attention_mask=batch["attention_mask"], use_cache=True, decoder_start_token_id=self.decoder_start_token_id, num_beams=self.eval_beams, max_length=self.eval_max_length, length_penalty=1.0 ) gen_time = (time.time() - t0) / batch["input_ids"].shape[0] preds: List[str] = self.ids_to_clean_text(generated_ids) target: List[str] = self.ids_to_clean_text(batch["labels"]) loss_tensors = self._step(batch) base_metrics = {name: loss for name, loss in zip(self.loss_names, loss_tensors)} rouge: Dict = self.calc_generative_metrics(preds, target) summ_len = np.mean(lmap(len, generated_ids)) base_metrics.update(gen_time=gen_time, gen_len=summ_len, preds=preds, target=target, **rouge) if dataloader_idx is not None: base_metrics.update(batch_idx=batch_idx, dataloader_idx=dataloader_idx) return base_metrics def test_step(self, batch, batch_idx, dataloader_idx): return self._generative_step(batch, batch_idx, dataloader_idx) def test_epoch_end(self, outputs_all_testsets): val_outputs_folder = "val_outputs" os.system("mkdir -p " + os.path.join(self.hparams.output_dir, val_outputs_folder)) for outputs in outputs_all_testsets: dataset_idx = outputs[0]['dataloader_idx'] if dataset_idx == 0: file_name = "test_both_predictions.txt" file_name_tgt = "test_both_targets.txt" dataset_name = 'test_both' elif dataset_idx == 1: file_name = "test_seen_predictions.txt" file_name_tgt = "test_seen_targets.txt" dataset_name = 'test_seen' else: file_name = "test_unseen_predictions.txt" file_name_tgt = "test_unseen_targets.txt" dataset_name = 'test_unseen' file_name += '.debug' file_name_tgt += '.debug' output_test_predictions_file = os.path.join(self.hparams.output_dir, val_outputs_folder, file_name) output_test_targets_file = os.path.join(self.hparams.output_dir, val_outputs_folder, file_name_tgt) # write predictions and targets for later rouge evaluation. with open(output_test_predictions_file, "w") as p_writer, open(output_test_targets_file, "w") as t_writer: for output_batch in outputs: p_writer.writelines(convert_text(s) + "\n" for s in output_batch["preds"]) t_writer.writelines(convert_text(s) + "\n" for s in output_batch["target"]) p_writer.close() t_writer.close() bleu_info = eval_bleu(self.hparams.data_dir, output_test_predictions_file, dataset_name) meteor_info = eval_meteor_test_webnlg(self.hparams.data_dir, output_test_predictions_file, dataset_name) chrf_info = eval_chrf_test_webnlg(self.hparams.data_dir, output_test_predictions_file, dataset_name) rank_zero_info(" %s - bleu_info: %s", dataset_name, bleu_info) rank_zero_info(" %s - meteor_info: %s", dataset_name, meteor_info) rank_zero_info(" %s - chrf_info: %s", dataset_name, chrf_info) outputs[0]['bleu'] = bleu_info return self.validation_epoch_end(outputs_all_testsets, prefix="test") def get_dataset(self, type_path) -> Seq2SeqDataset: n_obs = self.n_obs[type_path] max_target_length = self.target_lens[type_path] dataset = self.dataset_class( self.tokenizer, type_path=type_path, n_obs=n_obs, max_target_length=max_target_length, **self.dataset_kwargs, ) return dataset def get_dataloader(self, type_path: str, batch_size: int, shuffle: bool = False) -> DataLoader: dataset = self.get_dataset(type_path) if self.hparams.sortish_sampler and type_path != "test": sampler = dataset.make_sortish_sampler(batch_size, distributed=self.hparams.gpus > 1) return DataLoader( dataset, batch_size=batch_size, collate_fn=dataset.collate_fn, shuffle=False, num_workers=self.num_workers, sampler=sampler, ) elif self.hparams.max_tokens_per_batch is not None and type_path != "test": batch_sampler = dataset.make_dynamic_sampler( self.hparams.max_tokens_per_batch, distributed=self.hparams.gpus > 1 ) return DataLoader( dataset, batch_sampler=batch_sampler, collate_fn=dataset.collate_fn, # shuffle=False, num_workers=self.num_workers, # batch_size=None, ) else: return DataLoader( dataset, batch_size=batch_size, collate_fn=dataset.collate_fn, shuffle=shuffle, num_workers=self.num_workers, sampler=None, ) def train_dataloader(self) -> DataLoader: dataloader = self.get_dataloader("train", batch_size=self.hparams.train_batch_size, shuffle=True) return dataloader def val_dataloader(self) -> DataLoader: return self.get_dataloader("val", batch_size=self.hparams.eval_batch_size) def test_dataloader(self) -> List[DataLoader]: test_dataloader = self.get_dataloader("test_both", batch_size=self.hparams.eval_batch_size) test_seen_dataloader = self.get_dataloader("test_seen", batch_size=self.hparams.eval_batch_size) test_unseen_dataloader = self.get_dataloader("test_unseen", batch_size=self.hparams.eval_batch_size) return [test_dataloader, test_seen_dataloader, test_unseen_dataloader] @staticmethod def add_model_specific_args(parser, root_dir): BaseTransformer.add_model_specific_args(parser, root_dir) add_generic_args(parser, root_dir) parser.add_argument( "--max_source_length", default=1024, type=int, help="The maximum total input sequence length after tokenization. Sequences longer " "than this will be truncated, sequences shorter will be padded.", ) parser.add_argument( "--max_target_length", default=56, type=int, help="The maximum total input sequence length after tokenization. Sequences longer " "than this will be truncated, sequences shorter will be padded.", ) parser.add_argument( "--val_max_target_length", default=142, # these defaults are optimized for CNNDM. For xsum, see README.md. type=int, help="The maximum total input sequence length after tokenization. Sequences longer " "than this will be truncated, sequences shorter will be padded.", ) parser.add_argument( "--test_max_target_length", default=142, type=int, help="The maximum total input sequence length after tokenization. Sequences longer " "than this will be truncated, sequences shorter will be padded.", ) parser.add_argument("--freeze_encoder", action="store_true") parser.add_argument("--freeze_embeds", action="store_true") parser.add_argument("--sortish_sampler", action="store_true", default=False) parser.add_argument("--max_tokens_per_batch", type=int, default=None) parser.add_argument("--logger_name", type=str, choices=["default", "wandb", "wandb_shared"], default="default") parser.add_argument("--n_train", type=int, default=-1, required=False, help="# examples. -1 means use all.") parser.add_argument("--n_val", type=int, default=-1, required=False, help="# examples. -1 means use all.") parser.add_argument("--n_test", type=int, default=-1, required=False, help="# examples. -1 means use all.") parser.add_argument( "--task", type=str, default="summarization", required=False, help="# examples. -1 means use all." ) parser.add_argument("--label_smoothing", type=float, default=0.0, required=False) parser.add_argument("--src_lang", type=str, default="", required=False) parser.add_argument("--tgt_lang", type=str, default="", required=False) parser.add_argument("--eval_beams", type=int, default=None, required=False) parser.add_argument("--checkpoint", type=str, default=None, required=False) parser.add_argument( "--val_metric", type=str, default=None, required=False, choices=["bleu", "rouge2", "loss", None] ) parser.add_argument("--eval_max_gen_length", type=int, default=None, help="never generate more than n tokens") parser.add_argument("--save_top_k", type=int, default=1, required=False, help="How many checkpoints to save") parser.add_argument( "--early_stopping_patience", type=int, default=-1, required=False, help="-1 means never early stop. early_stopping_patience is measured in validation checks, not epochs. So val_check_interval will effect it.", ) return parser class TranslationModule(SummarizationModule): mode = "translation" loss_names = ["loss"] metric_names = ["bleu"] default_val_metric = "bleu" def __init__(self, hparams, **kwargs): super().__init__(hparams, **kwargs) self.dataset_kwargs["src_lang"] = hparams.src_lang self.dataset_kwargs["tgt_lang"] = hparams.tgt_lang def calc_generative_metrics(self, preds, target) -> dict: return calculate_bleu(preds, target) class Graph2TextModule(SummarizationModule): mode = "graph2text" loss_names = ["loss"] metric_names = ["sacrebleu"] default_val_metric = "bleu" def __init__(self, hparams, **kwargs): if type(hparams) == dict: hparams = argparse.Namespace(**hparams) print(f'Graph2Text hparams are: {hparams}') super().__init__(hparams, **kwargs) self.hparams.update(vars(hparams)) rank_zero_info("parameters %s", hparams) def calc_generative_metrics(self, preds, target) -> dict: return calculate_bleu(preds, target) def main(args, model=None) -> SummarizationModule: Path(args.output_dir).mkdir(exist_ok=True) if len(os.listdir(args.output_dir)) > 3 and args.do_train: raise ValueError("Output directory ({}) already exists and is not empty.".format(args.output_dir)) if model is None: if "summarization" in args.task: model: SummarizationModule = SummarizationModule(args) elif "translation" in args.task: model: SummarizationModule = TranslationModule(args) else: model: SummarizationModule = Graph2TextModule(args) dataset = Path(args.data_dir).name if ( args.logger_name == "default" or args.fast_dev_run or str(args.output_dir).startswith("/tmp") or str(args.output_dir).startswith("/var") ): logger = True # don't pollute wandb logs unnecessarily elif args.logger_name == "wandb": from pytorch_lightning.loggers import WandbLogger project = os.environ.get("WANDB_PROJECT", dataset) logger = WandbLogger(name=model.output_dir.name, project=project) elif args.logger_name == "wandb_shared": from pytorch_lightning.loggers import WandbLogger logger = WandbLogger(name=model.output_dir.name, project=f"hf_{dataset}") if args.early_stopping_patience >= 0: es_callback = get_early_stopping_callback(model.val_metric, args.early_stopping_patience) else: es_callback = False lower_is_better = args.val_metric == "loss" trainer: pl.Trainer = generic_train( model, args, logging_callback=Seq2SeqLoggingCallback(), checkpoint_callback=get_checkpoint_callback( args.output_dir, model.val_metric, args.save_top_k, lower_is_better ), early_stopping_callback=es_callback, logger=logger, ) pickle_save(model.hparams, model.output_dir / "hparams.pkl") if not args.do_predict: return model model.hparams.test_checkpoint = "" if not args.checkpoint: checkpoints = list(sorted(glob.glob(os.path.join(args.output_dir, "*.ckpt"), recursive=True))) else: checkpoints = [args.checkpoint] if checkpoints: model.hparams.test_checkpoint = checkpoints[-1] trainer.resume_from_checkpoint = checkpoints[-1] if args.do_predict and not args.do_train: checkpoint = checkpoints[-1] print(checkpoint) #trainer.test(ckpt_path=checkpoints[-1]) trainer.test(model, ckpt_path=checkpoint) return model trainer.logger.log_hyperparams(model.hparams) # test() without a model tests using the best checkpoint automatically trainer.test() return model if __name__ == "__main__": parser = argparse.ArgumentParser() parser = pl.Trainer.add_argparse_args(parser) parser = SummarizationModule.add_model_specific_args(parser, os.getcwd()) args = parser.parse_args() main(args)