nicholasKluge
commited on
Commit
•
e2a3f6a
1
Parent(s):
d6379a0
Upload 6 files
Browse files- completion-tfidf-matrix.pkl +3 -0
- completion-vectorizer.pkl +3 -0
- create-tfidf-matrix.py +46 -0
- prompt-tfidf-matrix.pkl +3 -0
- prompt-vectorizer.pkl +3 -0
completion-tfidf-matrix.pkl
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:16f7cb342da64a6334bb035d162a29579853926af2243c14029fb5043d4fbd81
|
3 |
+
size 116328867
|
completion-vectorizer.pkl
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:add850bde149e5de855d3c0334cd99ef5055289f8d103626250db2b5a1bbd0dc
|
3 |
+
size 4036115
|
create-tfidf-matrix.py
ADDED
@@ -0,0 +1,46 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import joblib
|
2 |
+
import pandas as pd
|
3 |
+
from sklearn.metrics.pairwise import cosine_similarity
|
4 |
+
from sklearn.feature_extraction.text import TfidfVectorizer
|
5 |
+
import argparse
|
6 |
+
|
7 |
+
def main():
|
8 |
+
parser = argparse.ArgumentParser(description='Process some integers.')
|
9 |
+
parser.add_argument('--input', type=str, help="Input file path (file should be in parquet format and have 'prompt' and 'completion' columns)")
|
10 |
+
parser.add_argument('--output', type=str, help='Output file path')
|
11 |
+
args = parser.parse_args()
|
12 |
+
|
13 |
+
df = pd.read_parquet(args.input)
|
14 |
+
|
15 |
+
# fit the vectorizer on the prompt column
|
16 |
+
prompt_tfidf_vectorizer = TfidfVectorizer()
|
17 |
+
prompt_tfidf_vectorizer.fit(df['prompt'])
|
18 |
+
|
19 |
+
# save the vectorizer
|
20 |
+
joblib.dump(prompt_tfidf_vectorizer, args.output + 'prompt-vectorizer.pkl')
|
21 |
+
|
22 |
+
# get the tfidf_matrix
|
23 |
+
prompt_tfidf_matrix = prompt_tfidf_vectorizer.transform(df['prompt'])
|
24 |
+
|
25 |
+
# save the tfidf_matrix
|
26 |
+
joblib.dump(prompt_tfidf_matrix, args.output + 'prompt-tfidf-matrix.pkl')
|
27 |
+
|
28 |
+
# fit the vectorizer on the completion column
|
29 |
+
completion_tfidf_vectorizer = TfidfVectorizer()
|
30 |
+
completion_tfidf_vectorizer.fit(df['completion'])
|
31 |
+
|
32 |
+
# save the vectorizer
|
33 |
+
joblib.dump(completion_tfidf_vectorizer, args.output + 'completion-vectorizer.pkl')
|
34 |
+
|
35 |
+
# get the tfidf_matrix
|
36 |
+
completion_tfidf_matrix = completion_tfidf_vectorizer.transform(df['completion'])
|
37 |
+
|
38 |
+
# save the tfidf_matrix
|
39 |
+
joblib.dump(completion_tfidf_matrix, args.output + 'completion-tfidf-matrix.pkl')
|
40 |
+
|
41 |
+
print("Done!")
|
42 |
+
|
43 |
+
if __name__ == '__main__':
|
44 |
+
main()
|
45 |
+
|
46 |
+
# example usage: python create-tfidf-matrix.py --input fine-tuning-data.parquet --output ./
|
prompt-tfidf-matrix.pkl
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:95d8c1d302b36e5fef3da79e802354972158b247051715c98d55f351b8993fe2
|
3 |
+
size 37977659
|
prompt-vectorizer.pkl
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:739df119b82ef1f2d8dfd4d85bc1ee489d2705b48d1bd701627df9222e15cc8f
|
3 |
+
size 3324940
|