Zeb
commited on
Commit
•
6ee996c
1
Parent(s):
af95aac
Add option to load in original files
Browse files
BabyLM.py
CHANGED
@@ -26,14 +26,24 @@ filenames = [
|
|
26 |
class BabyLM(datasets.GeneratorBasedBuilder):
|
27 |
|
28 |
BUILDER_CONFIGS = [
|
|
|
|
|
|
|
|
|
|
|
29 |
datasets.BuilderConfig(
|
30 |
name="strict_small",
|
31 |
-
description="Small version of the dataset with 10M words",
|
|
|
|
|
|
|
|
|
|
|
32 |
version="1.0.0",
|
33 |
),
|
34 |
datasets.BuilderConfig(
|
35 |
name="strict",
|
36 |
-
description="Full version of the dataset with 100M words",
|
37 |
version="1.0.0",
|
38 |
),
|
39 |
datasets.BuilderConfig(
|
@@ -71,12 +81,14 @@ class BabyLM(datasets.GeneratorBasedBuilder):
|
|
71 |
Returns data for different splits
|
72 |
"""
|
73 |
|
74 |
-
if self.config.name
|
75 |
train_data_dir = "10M"
|
76 |
else:
|
77 |
train_data_dir = "100M"
|
78 |
if 'gold' in self.config.name:
|
79 |
folder = 'tagged_gold'
|
|
|
|
|
80 |
else:
|
81 |
folder = 'tagged'
|
82 |
|
@@ -120,20 +132,28 @@ class BabyLM(datasets.GeneratorBasedBuilder):
|
|
120 |
|
121 |
global_idx = 0
|
122 |
|
123 |
-
|
124 |
-
|
125 |
-
|
126 |
-
|
127 |
-
|
128 |
-
|
129 |
-
|
130 |
-
|
131 |
-
|
132 |
-
|
133 |
-
|
134 |
-
|
135 |
-
|
136 |
-
|
137 |
-
|
138 |
-
|
139 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
26 |
class BabyLM(datasets.GeneratorBasedBuilder):
|
27 |
|
28 |
BUILDER_CONFIGS = [
|
29 |
+
datasets.BuilderConfig(
|
30 |
+
name="original_strict_small",
|
31 |
+
description="Small version of the dataset with 10M words, not cleaned",
|
32 |
+
version="1.0.0",
|
33 |
+
),
|
34 |
datasets.BuilderConfig(
|
35 |
name="strict_small",
|
36 |
+
description="Small version of the dataset with 10M words, cleaned",
|
37 |
+
version="1.0.0",
|
38 |
+
),
|
39 |
+
datasets.BuilderConfig(
|
40 |
+
name="original_strict",
|
41 |
+
description="Full version of the dataset with 10M words, not cleaned",
|
42 |
version="1.0.0",
|
43 |
),
|
44 |
datasets.BuilderConfig(
|
45 |
name="strict",
|
46 |
+
description="Full version of the dataset with 100M words, cleaned",
|
47 |
version="1.0.0",
|
48 |
),
|
49 |
datasets.BuilderConfig(
|
|
|
81 |
Returns data for different splits
|
82 |
"""
|
83 |
|
84 |
+
if "strict_small" in self.config.name:
|
85 |
train_data_dir = "10M"
|
86 |
else:
|
87 |
train_data_dir = "100M"
|
88 |
if 'gold' in self.config.name:
|
89 |
folder = 'tagged_gold'
|
90 |
+
elif 'original' in self.config.name:
|
91 |
+
folder = 'original'
|
92 |
else:
|
93 |
folder = 'tagged'
|
94 |
|
|
|
132 |
|
133 |
global_idx = 0
|
134 |
|
135 |
+
if 'original' in self.config.name:
|
136 |
+
for filepath in filepaths:
|
137 |
+
with open(filepath, encoding="utf-8") as f:
|
138 |
+
for line in f:
|
139 |
+
yield global_idx, {"text": line.strip(), "filename": '', 'tagged_text': line.strip()}
|
140 |
+
global_idx += 1
|
141 |
+
|
142 |
+
else:
|
143 |
+
for filepath in filepaths:
|
144 |
+
with open(filepath, encoding="utf-8") as f:
|
145 |
+
is_tags = False
|
146 |
+
text = ""
|
147 |
+
filename = ""
|
148 |
+
# Every other row contains POS tags. First row is the filename (we can't use filepath since the file path changes upon caching)
|
149 |
+
for row in f:
|
150 |
+
if filename == "":
|
151 |
+
filename = row.strip()
|
152 |
+
continue
|
153 |
+
if is_tags:
|
154 |
+
yield global_idx, {"text": text.strip(), "tagged_text": row.strip(), "filename": filename}
|
155 |
+
global_idx += 1
|
156 |
+
is_tags = False
|
157 |
+
else:
|
158 |
+
text = row
|
159 |
+
is_tags = True
|