Rename old datasets interface

#231
rwth-i6 · Oct 24, 2022 · 000ee8d · 000ee8d
1 parent 8f652b9
commit 000ee8d
Show file tree

Hide file tree

Showing 20 changed files with 7 additions and 5 deletions.
diff --git a/datasets_old_2022_10/README.md b/datasets_old_2022_10/README.md
@@ -0,0 +1,2 @@
+This was an initial try to define a common interface for datasets.
+We decided to go with a new interface, see [#231](https://github.com/rwth-i6/returnn_common/issues/231).
diff --git a/datasets/__init__.py → datasets_old_2022_10/__init__.py b/datasets/__init__.py → datasets_old_2022_10/__init__.py
diff --git a/datasets/asr/__init__.py → datasets_old_2022_10/asr/__init__.py b/datasets/asr/__init__.py → datasets_old_2022_10/asr/__init__.py
diff --git a/datasets/asr/features.py → datasets_old_2022_10/asr/features.py b/datasets/asr/features.py → datasets_old_2022_10/asr/features.py
diff --git a/datasets/asr/librispeech/__init__.py → ...s_old_2022_10/asr/librispeech/__init__.py b/datasets/asr/librispeech/__init__.py → ...s_old_2022_10/asr/librispeech/__init__.py
diff --git a/datasets/asr/librispeech/lm.py → datasets_old_2022_10/asr/librispeech/lm.py b/datasets/asr/librispeech/lm.py → datasets_old_2022_10/asr/librispeech/lm.py
diff --git a/.../librispeech/norm_stats/stats.40.mean.txt → .../librispeech/norm_stats/stats.40.mean.txt b/.../librispeech/norm_stats/stats.40.mean.txt → .../librispeech/norm_stats/stats.40.mean.txt
diff --git a/...brispeech/norm_stats/stats.40.std_dev.txt → ...brispeech/norm_stats/stats.40.std_dev.txt b/...brispeech/norm_stats/stats.40.std_dev.txt → ...brispeech/norm_stats/stats.40.std_dev.txt
diff --git a/...s/asr/librispeech/norm_stats/stats.40.txt → ...0/asr/librispeech/norm_stats/stats.40.txt b/...s/asr/librispeech/norm_stats/stats.40.txt → ...0/asr/librispeech/norm_stats/stats.40.txt
diff --git a/datasets/asr/librispeech/oggzip.py → ...ets_old_2022_10/asr/librispeech/oggzip.py b/datasets/asr/librispeech/oggzip.py → ...ets_old_2022_10/asr/librispeech/oggzip.py
@@ -72,7 +72,7 @@ def get_train_dataset(self) -> Dict[str, Any]:
 
   def get_eval_datasets(self) -> Dict[str, Dict[str, Any]]:
     """
-    Get eval datasets
+    Get eval datasets_old_2022_10
     """
     return {
       "dev": self.get_dataset("dev", train=False, subset=3000),

diff --git a/datasets/asr/librispeech/vocabs.py → ...ets_old_2022_10/asr/librispeech/vocabs.py b/datasets/asr/librispeech/vocabs.py → ...ets_old_2022_10/asr/librispeech/vocabs.py
diff --git a/...asr/librispeech/vocabs/trans.bpe10k.codes → ...asr/librispeech/vocabs/trans.bpe10k.codes b/...asr/librispeech/vocabs/trans.bpe10k.codes → ...asr/librispeech/vocabs/trans.bpe10k.codes
diff --git a/...asr/librispeech/vocabs/trans.bpe10k.vocab → ...asr/librispeech/vocabs/trans.bpe10k.vocab b/...asr/librispeech/vocabs/trans.bpe10k.vocab → ...asr/librispeech/vocabs/trans.bpe10k.vocab
diff --git a/.../asr/librispeech/vocabs/trans.bpe1k.codes → .../asr/librispeech/vocabs/trans.bpe1k.codes b/.../asr/librispeech/vocabs/trans.bpe1k.codes → .../asr/librispeech/vocabs/trans.bpe1k.codes
diff --git a/.../asr/librispeech/vocabs/trans.bpe1k.vocab → .../asr/librispeech/vocabs/trans.bpe1k.vocab b/.../asr/librispeech/vocabs/trans.bpe1k.vocab → .../asr/librispeech/vocabs/trans.bpe1k.vocab
diff --git a/datasets/asr/timit/__init__.py → datasets_old_2022_10/asr/timit/__init__.py b/datasets/asr/timit/__init__.py → datasets_old_2022_10/asr/timit/__init__.py
diff --git a/datasets/asr/timit/nltk.py → datasets_old_2022_10/asr/timit/nltk.py b/datasets/asr/timit/nltk.py → datasets_old_2022_10/asr/timit/nltk.py
@@ -54,7 +54,7 @@ def get_train_dataset(self) -> Dict[str, Any]:
 
   def get_eval_datasets(self) -> Dict[str, Dict[str, Any]]:
     """
-    Get eval datasets
+    Get eval datasets_old_2022_10
     """
     return {
       "dev": self.get_dataset("dev"),
@@ -72,7 +72,7 @@ def get_main_dataset(self) -> Dict[str]:
 
   def get_dataset(self, key, subset=None):
     """
-    Get datasets
+    Get datasets_old_2022_10
     """
     assert key in {"train", "dev"}
     assert not subset

diff --git a/datasets/interface.py → datasets_old_2022_10/interface.py b/datasets/interface.py → datasets_old_2022_10/interface.py
diff --git a/nn_raw/lm.py b/nn_raw/lm.py
@@ -3,7 +3,7 @@
 Language model functions.
 """
 
-from ..datasets.interface import VocabConfig
+from ..datasets_old_2022_10.interface import VocabConfig
 from typing import Dict, Any
 
 

diff --git a/nn_raw/transducer/transducer_fullsum.py b/nn_raw/transducer/transducer_fullsum.py
@@ -64,7 +64,7 @@
 from .recomb_recog import targetb_recomb_recog
 from .loss import rnnt_loss, rnnt_loss_out_type
 from ..collect_out_str import make_out_str_func
-from ...datasets.interface import TargetConfig
+from ...datasets_old_2022_10.interface import TargetConfig
 
 
 class Context:
Original file line number	Diff line number	Diff line change
		@@ -0,0 +1,2 @@
		This was an initial try to define a common interface for datasets.
		We decided to go with a new interface, see [#231](https://github.com/rwth-i6/returnn_common/issues/231).