albertvillanova HF staff commited on
Commit
043c6cc
1 Parent(s): 2d7620b

Fix RuntimeError: Sharding is ambiguous for this dataset (#4)

Browse files

- Pass tuple in gen_kwargs to avoid ambiguous sharding error (81710454f9f6a612e8e466cb73051f67970918b5)

Files changed (1) hide show
  1. ami.py +4 -4
ami.py CHANGED
@@ -396,7 +396,7 @@ class AMI(datasets.GeneratorBasedBuilder):
396
  gen_kwargs={
397
  "annotation_path": annotation_path,
398
  "samples_paths": train_path,
399
- "ids": _TRAIN_SAMPLE_IDS,
400
  "verification_ids": train_ids,
401
  },
402
  ),
@@ -405,7 +405,7 @@ class AMI(datasets.GeneratorBasedBuilder):
405
  gen_kwargs={
406
  "annotation_path": annotation_path,
407
  "samples_paths": validation_path,
408
- "ids": _VALIDATION_SAMPLE_IDS,
409
  "verification_ids": validation_ids,
410
  },
411
  ),
@@ -414,7 +414,7 @@ class AMI(datasets.GeneratorBasedBuilder):
414
  gen_kwargs={
415
  "annotation_path": annotation_path,
416
  "samples_paths": eval_path,
417
- "ids": _EVAL_SAMPLE_IDS,
418
  "verification_ids": eval_ids,
419
  },
420
  ),
@@ -501,7 +501,7 @@ class AMI(datasets.GeneratorBasedBuilder):
501
  sample_paths = samples_paths[num_audios * i : num_audios * (i + 1)]
502
  sample_verification_id = set(verification_ids[num_audios * i : num_audios * (i + 1)])
503
 
504
- # make sure that multi microphone samples are correctly atttributed to labels
505
  if len(sample_verification_id) > 1 or next(iter(sample_verification_id)) != _id:
506
  raise ValueError(
507
  f"Incorrect dataset generation. The files {sample_paths} of id {_id} have incorrect verification_ids {sample_verification_id}."
 
396
  gen_kwargs={
397
  "annotation_path": annotation_path,
398
  "samples_paths": train_path,
399
+ "ids": tuple(_TRAIN_SAMPLE_IDS),
400
  "verification_ids": train_ids,
401
  },
402
  ),
 
405
  gen_kwargs={
406
  "annotation_path": annotation_path,
407
  "samples_paths": validation_path,
408
+ "ids": tuple(_VALIDATION_SAMPLE_IDS),
409
  "verification_ids": validation_ids,
410
  },
411
  ),
 
414
  gen_kwargs={
415
  "annotation_path": annotation_path,
416
  "samples_paths": eval_path,
417
+ "ids": tuple(_EVAL_SAMPLE_IDS),
418
  "verification_ids": eval_ids,
419
  },
420
  ),
 
501
  sample_paths = samples_paths[num_audios * i : num_audios * (i + 1)]
502
  sample_verification_id = set(verification_ids[num_audios * i : num_audios * (i + 1)])
503
 
504
+ # make sure that multi microphone samples are correctly attributed to labels
505
  if len(sample_verification_id) > 1 or next(iter(sample_verification_id)) != _id:
506
  raise ValueError(
507
  f"Incorrect dataset generation. The files {sample_paths} of id {_id} have incorrect verification_ids {sample_verification_id}."