mirror of
https://github.com/hiyouga/LLaMA-Factory.git
synced 2025-08-03 04:02:49 +08:00
parent
b12d4beb8a
commit
8ecf606230
@ -120,8 +120,8 @@ def load_single_dataset(
|
|||||||
logger.info("Sampled {} examples from dataset {}.".format(dataset_attr.num_samples, dataset_attr))
|
logger.info("Sampled {} examples from dataset {}.".format(dataset_attr.num_samples, dataset_attr))
|
||||||
|
|
||||||
if data_args.max_samples is not None: # truncate dataset
|
if data_args.max_samples is not None: # truncate dataset
|
||||||
indexes = np.random.permutation(len(dataset))[: data_args.max_samples]
|
max_samples = min(data_args.max_samples, len(dataset))
|
||||||
dataset = dataset.select(indexes)
|
dataset = dataset.select(range(max_samples))
|
||||||
|
|
||||||
return align_dataset(dataset, dataset_attr, data_args)
|
return align_dataset(dataset, dataset_attr, data_args)
|
||||||
|
|
||||||
|
Loading…
x
Reference in New Issue
Block a user