Loading fairseq/data.py +3 −3 Changes for fairseq/data.py: 3 added lines, 3 removed lines. Original line number Diff line number Diff line Loading @@ -21,7 +21,7 @@ from fairseq.indexed_dataset import IndexedDataset, IndexedInMemoryDataset, Inde def has_binary_files(data_dir, splits): for split in splits: if len(glob.glob(os.path.join(data_dir, f'{split}.*-*.*.bin'))) < 2: if len(glob.glob(os.path.join(data_dir, '{}.*-*.*.bin'.format(split)))) < 2: return False return True Loading Loading @@ -107,8 +107,8 @@ def load_raw_text_dataset(path, load_splits, src=None, dst=None): # Load dataset from raw text files for split in load_splits: src_path = os.path.join(path, f'{split}.{src}') dst_path = os.path.join(path, f'{split}.{dst}') src_path = os.path.join(path, '{}.{}'.format(split, src)) dst_path = os.path.join(path, '{}.{}'.format(split, dst)) dataset.splits[split] = LanguagePairDataset( IndexedRawTextDataset(src_path, src_dict), IndexedRawTextDataset(dst_path, dst_dict), Loading fairseq/progress_bar.py +7 −6 Changes for fairseq/progress_bar.py: 7 added lines, 6 removed lines. Original line number Diff line number Diff line Loading @@ -27,9 +27,9 @@ class progress_bar(object): self.epoch = epoch self.prefix = '' if epoch is not None: self.prefix += f'| epoch {epoch:03d}' self.prefix += '| epoch {:03d}'.format(epoch) if prefix is not None: self.prefix += f' | {prefix}' self.prefix += ' | {}'.format(prefix) def __enter__(self): return self Loading Loading @@ -90,7 +90,7 @@ class json_progress_bar(progress_bar): self.log_interval is not None and i % self.log_interval == 0: update = self.epoch + float(i / size) if self.epoch is not None else None stats = self._format_stats(self.stats, epoch=self.epoch, update=update) print("sweep_log: " + json.dumps(stats), flush=True) print('sweep_log: ' + json.dumps(stats), flush=True) def log(self, stats): """Log intermediate stats according to log_interval.""" Loading Loading @@ -152,7 +152,8 @@ class simple_progress_bar(progress_bar): if self.stats is not None and i > 0 and \ self.log_interval is not None and i % self.log_interval == 0: postfix = self._str_commas(self.stats) print(f'{self.prefix}: {i:5d} / {size:d} {postfix}', flush=True) print('{}: {:5d} / {:d} {}'.format(self.prefix, i, size, postfix), flush=True) def log(self, stats): """Log intermediate stats according to log_interval.""" Loading @@ -161,7 +162,7 @@ class simple_progress_bar(progress_bar): def print(self, stats): """Print end-of-epoch stats.""" postfix = self._str_pipes(self._format_stats(stats)) print(f'{self.prefix} | {postfix}', flush=True) print('{} | {}'.format(self.prefix, postfix), flush=True) class tqdm_progress_bar(progress_bar): Loading @@ -181,4 +182,4 @@ class tqdm_progress_bar(progress_bar): def print(self, stats): """Print end-of-epoch stats.""" postfix = self._str_pipes(self._format_stats(stats)) self.tqdm.write(f'{self.tqdm.desc} | {postfix}') self.tqdm.write('{} | {}'.format(self.tqdm.desc, postfix)) Loading
fairseq/data.py +3 −3 Changes for fairseq/data.py: 3 added lines, 3 removed lines. Original line number Diff line number Diff line Loading @@ -21,7 +21,7 @@ from fairseq.indexed_dataset import IndexedDataset, IndexedInMemoryDataset, Inde def has_binary_files(data_dir, splits): for split in splits: if len(glob.glob(os.path.join(data_dir, f'{split}.*-*.*.bin'))) < 2: if len(glob.glob(os.path.join(data_dir, '{}.*-*.*.bin'.format(split)))) < 2: return False return True Loading Loading @@ -107,8 +107,8 @@ def load_raw_text_dataset(path, load_splits, src=None, dst=None): # Load dataset from raw text files for split in load_splits: src_path = os.path.join(path, f'{split}.{src}') dst_path = os.path.join(path, f'{split}.{dst}') src_path = os.path.join(path, '{}.{}'.format(split, src)) dst_path = os.path.join(path, '{}.{}'.format(split, dst)) dataset.splits[split] = LanguagePairDataset( IndexedRawTextDataset(src_path, src_dict), IndexedRawTextDataset(dst_path, dst_dict), Loading
fairseq/progress_bar.py +7 −6 Changes for fairseq/progress_bar.py: 7 added lines, 6 removed lines. Original line number Diff line number Diff line Loading @@ -27,9 +27,9 @@ class progress_bar(object): self.epoch = epoch self.prefix = '' if epoch is not None: self.prefix += f'| epoch {epoch:03d}' self.prefix += '| epoch {:03d}'.format(epoch) if prefix is not None: self.prefix += f' | {prefix}' self.prefix += ' | {}'.format(prefix) def __enter__(self): return self Loading Loading @@ -90,7 +90,7 @@ class json_progress_bar(progress_bar): self.log_interval is not None and i % self.log_interval == 0: update = self.epoch + float(i / size) if self.epoch is not None else None stats = self._format_stats(self.stats, epoch=self.epoch, update=update) print("sweep_log: " + json.dumps(stats), flush=True) print('sweep_log: ' + json.dumps(stats), flush=True) def log(self, stats): """Log intermediate stats according to log_interval.""" Loading Loading @@ -152,7 +152,8 @@ class simple_progress_bar(progress_bar): if self.stats is not None and i > 0 and \ self.log_interval is not None and i % self.log_interval == 0: postfix = self._str_commas(self.stats) print(f'{self.prefix}: {i:5d} / {size:d} {postfix}', flush=True) print('{}: {:5d} / {:d} {}'.format(self.prefix, i, size, postfix), flush=True) def log(self, stats): """Log intermediate stats according to log_interval.""" Loading @@ -161,7 +162,7 @@ class simple_progress_bar(progress_bar): def print(self, stats): """Print end-of-epoch stats.""" postfix = self._str_pipes(self._format_stats(stats)) print(f'{self.prefix} | {postfix}', flush=True) print('{} | {}'.format(self.prefix, postfix), flush=True) class tqdm_progress_bar(progress_bar): Loading @@ -181,4 +182,4 @@ class tqdm_progress_bar(progress_bar): def print(self, stats): """Print end-of-epoch stats.""" postfix = self._str_pipes(self._format_stats(stats)) self.tqdm.write(f'{self.tqdm.desc} | {postfix}') self.tqdm.write('{} | {}'.format(self.tqdm.desc, postfix))