Tekkonetes
commited on
Commit
·
e088236
1
Parent(s):
b26eb37
yas
Browse files- README.md +0 -9
- corpus.txt +0 -0
- predict.py +24 -0
- pytorch_model.bin +0 -3
- requirements.txt +2 -0
- train.py +26 -0
README.md
DELETED
@@ -1,9 +0,0 @@
|
|
1 |
-
---
|
2 |
-
language: py
|
3 |
-
library_name: PyTorch
|
4 |
-
pipeline_tag: text-generation
|
5 |
-
tags:
|
6 |
-
- pytorch
|
7 |
-
- text2text
|
8 |
-
- tekkonetes
|
9 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
corpus.txt
ADDED
File without changes
|
predict.py
ADDED
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# Import libraries
|
2 |
+
import nltk
|
3 |
+
import numpy as np
|
4 |
+
import pickle
|
5 |
+
nltk.download('punkt')
|
6 |
+
# Define predict_word function
|
7 |
+
def predict_word(model, last_word):
|
8 |
+
if last_word in model:
|
9 |
+
return np.random.choice(model[last_word])
|
10 |
+
else:
|
11 |
+
return None
|
12 |
+
# Load the model
|
13 |
+
with open("model.pkl", "rb") as f:
|
14 |
+
model = pickle.load(f)
|
15 |
+
|
16 |
+
# Run the prediction for 10 words
|
17 |
+
input_words = input('Input words: ')
|
18 |
+
for i in range(10):
|
19 |
+
input_words_list = nltk.word_tokenize(input_words)
|
20 |
+
last_word = input_words_list[-1]
|
21 |
+
predicted_word = predict_word(model, last_word)
|
22 |
+
input_words = f"{input_words}" + " " + f'{predicted_word}'
|
23 |
+
|
24 |
+
print(input_words)
|
pytorch_model.bin
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:a4d1d24b0310816195ff64dfb22d22ea033844dc332f0c2bbc53a0af421483c3
|
3 |
-
size 15748927
|
|
|
|
|
|
|
|
requirements.txt
ADDED
@@ -0,0 +1,2 @@
|
|
|
|
|
|
|
1 |
+
nltk==3.8.1
|
2 |
+
numpy==1.24.2
|
train.py
ADDED
@@ -0,0 +1,26 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# Download nltk and numpy
|
2 |
+
import os
|
3 |
+
os.system('pip install nltk numpy')
|
4 |
+
import nltk
|
5 |
+
import numpy as np
|
6 |
+
|
7 |
+
nltk.download('punkt')
|
8 |
+
def train_model(corpus):
|
9 |
+
tokens = nltk.word_tokenize(corpus)
|
10 |
+
model = {}
|
11 |
+
for i in range(len(tokens) - 1):
|
12 |
+
if tokens[i] in model:
|
13 |
+
model[tokens[i]].append(tokens[i + 1])
|
14 |
+
else:
|
15 |
+
model[tokens[i]] = [tokens[i + 1]]
|
16 |
+
return model
|
17 |
+
|
18 |
+
import pickle
|
19 |
+
|
20 |
+
# Train the model on a given corpus
|
21 |
+
corpus = open('corpus.txt').read()
|
22 |
+
model = train_model(corpus)
|
23 |
+
|
24 |
+
# Save the model to a file
|
25 |
+
with open("model.pkl", "wb") as f:
|
26 |
+
pickle.dump(model, f)
|