fix for squad

This commit is contained in:
VictorSanh
2019-08-22 00:25:42 -04:00
parent e00b4ff1de
commit b006a7a12f

View File

@@ -272,7 +272,7 @@ def evaluate(args, model, tokenizer, prefix=""):
def load_and_cache_examples(args, tokenizer, evaluate=False, output_examples=False): def load_and_cache_examples(args, tokenizer, evaluate=False, output_examples=False):
if args.local_rank not in [-1, 0]: if args.local_rank not in [-1, 0] and not evaluate:
torch.distributed.barrier() # Make sure only the first process in distributed training process the dataset, and the others will use the cache torch.distributed.barrier() # Make sure only the first process in distributed training process the dataset, and the others will use the cache
# Load data features from cache or dataset file # Load data features from cache or dataset file
@@ -299,7 +299,7 @@ def load_and_cache_examples(args, tokenizer, evaluate=False, output_examples=Fal
logger.info("Saving features into cached file %s", cached_features_file) logger.info("Saving features into cached file %s", cached_features_file)
torch.save(features, cached_features_file) torch.save(features, cached_features_file)
if args.local_rank == 0: if args.local_rank == 0 and not evaluate:
torch.distributed.barrier() # Make sure only the first process in distributed training process the dataset, and the others will use the cache torch.distributed.barrier() # Make sure only the first process in distributed training process the dataset, and the others will use the cache
# Convert to Tensors and build dataset # Convert to Tensors and build dataset