Spaces:
Runtime error
Runtime error
stevenkolawole
commited on
Commit
•
d352f8e
1
Parent(s):
d9b4b87
add file uploader & download functions
Browse files- app.py +70 -26
- requirements.txt +1 -1
app.py
CHANGED
@@ -10,8 +10,7 @@ import tensorflow as tf
|
|
10 |
|
11 |
def main():
|
12 |
st.title("Interactive demo: T5 Multitasking Demo")
|
13 |
-
st.
|
14 |
-
text summarization, document similarity, and grammatical correctness of sentences.**")
|
15 |
saved_model_path = load_model_cache()
|
16 |
|
17 |
# Model is loaded in st.session_state to remain stateless across reloading
|
@@ -33,43 +32,88 @@ def load_model_cache():
|
|
33 |
snapshot_download(repo_id="stevekola/T5", cache_dir=CACHE_DIR)
|
34 |
saved_model_path = os.path.join(CACHE_DIR, os.listdir(CACHE_DIR)[0])
|
35 |
return saved_model_path
|
36 |
-
|
37 |
|
38 |
def dashboard(model):
|
39 |
-
"""
|
40 |
params:
|
41 |
model stateless model to run inference from
|
42 |
"""
|
43 |
-
st.sidebar.write("**Select the Task Type over here**")
|
44 |
task_type = st.sidebar.radio("Task Type",
|
45 |
[
|
46 |
-
|
47 |
-
|
48 |
-
|
49 |
-
|
50 |
-
|
51 |
-
|
52 |
])
|
53 |
-
|
54 |
-
|
55 |
-
|
56 |
-
|
57 |
-
|
58 |
-
|
59 |
-
|
60 |
-
|
61 |
-
|
62 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
63 |
else:
|
64 |
-
|
65 |
-
|
|
|
|
|
|
|
|
|
|
|
66 |
|
67 |
st.write("**Output Text**")
|
68 |
-
with st.spinner("
|
69 |
output_text = predict(task_type, sentence, model)
|
70 |
st.write(output_text)
|
71 |
-
|
72 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
73 |
|
74 |
def predict(task_type, sentence, model):
|
75 |
"""Function to parse the user inputs, run the parsed text through the
|
10 |
|
11 |
def main():
|
12 |
st.title("Interactive demo: T5 Multitasking Demo")
|
13 |
+
st.sidebar.image("https://i.gzn.jp/img/2020/02/25/google-ai-t5/01.png")
|
|
|
14 |
saved_model_path = load_model_cache()
|
15 |
|
16 |
# Model is loaded in st.session_state to remain stateless across reloading
|
32 |
snapshot_download(repo_id="stevekola/T5", cache_dir=CACHE_DIR)
|
33 |
saved_model_path = os.path.join(CACHE_DIR, os.listdir(CACHE_DIR)[0])
|
34 |
return saved_model_path
|
35 |
+
|
36 |
|
37 |
def dashboard(model):
|
38 |
+
"""Function to display the inputs and results
|
39 |
params:
|
40 |
model stateless model to run inference from
|
41 |
"""
|
|
|
42 |
task_type = st.sidebar.radio("Task Type",
|
43 |
[
|
44 |
+
"Translate English to French",
|
45 |
+
"Translate English to German",
|
46 |
+
"Translate English to Romanian",
|
47 |
+
"Grammatical Correctness of Sentence",
|
48 |
+
"Text Summarization",
|
49 |
+
"Document Similarity Score"
|
50 |
])
|
51 |
+
|
52 |
+
default_sentence = "I am Steven and I live in Lagos, Nigeria."
|
53 |
+
text_summarization_sentence = "I don't care about those doing the comparison, but comparing \
|
54 |
+
the Ghanaian Jollof Rice to Nigerian Jollof Rice is an insult to Nigerians."
|
55 |
+
doc_similarity_sentence1 = "I reside in the commercial capital city of Nigeria, which is Lagos."
|
56 |
+
doc_similarity_sentence2 = "I live in Lagos."
|
57 |
+
help_msg = "You could either type in the sentences to run inferences on or use the upload button to \
|
58 |
+
upload text files containing those sentences. The input sentence box, by default, displays sample \
|
59 |
+
texts or the texts in the files that you've uploaded. Feel free to erase them and type in new sentences."
|
60 |
+
|
61 |
+
if task_type.startswith("Document Similarity"): # document similarity requires two documents
|
62 |
+
uploaded_file = upload_files(help_msg, text="Upload 2 documents for similarity check", accept_multiple_files=True)
|
63 |
+
if uploaded_file:
|
64 |
+
sentence1 = st.text_area("Enter first document/sentence", uploaded_file[0], help=help_msg)
|
65 |
+
sentence2 = st.text_area("Enter second document/sentence", uploaded_file[1], help=help_msg)
|
66 |
+
else:
|
67 |
+
sentence1 = st.text_area("Enter first document/sentence", doc_similarity_sentence1)
|
68 |
+
sentence2 = st.text_area("Enter second document/sentence", doc_similarity_sentence2)
|
69 |
+
sentence = sentence1 + "---" + sentence2 # to be processed like other tasks' single sentences
|
70 |
else:
|
71 |
+
uploaded_file = upload_files(help_msg)
|
72 |
+
if uploaded_file:
|
73 |
+
sentence = st.text_area("Enter sentence", uploaded_file, help=help_msg)
|
74 |
+
elif task_type.startswith("Text Summarization"): # text summarization's default input should be longer
|
75 |
+
sentence = st.text_area("Enter sentence", text_summarization_sentence, help=help_msg)
|
76 |
+
else:
|
77 |
+
sentence = st.text_area("Enter sentence", default_sentence, help=help_msg)
|
78 |
|
79 |
st.write("**Output Text**")
|
80 |
+
with st.spinner("Waiting for prediction..."): # spinner while model is running inferences
|
81 |
output_text = predict(task_type, sentence, model)
|
82 |
st.write(output_text)
|
83 |
+
try: # to workaround the environment's Streamlit version
|
84 |
+
st.download_button("Download output text", output_text)
|
85 |
+
except AttributeError:
|
86 |
+
st.text("File download not enabled for this Streamlit version \U0001F612")
|
87 |
+
|
88 |
+
|
89 |
+
def upload_files(help_msg, text="Upload a text file here", accept_multiple_files=False):
|
90 |
+
"""Function to upload text files and return as string text
|
91 |
+
params:
|
92 |
+
text Display label for the upload button
|
93 |
+
accept_multiple_files params for the file_uploader function to accept more than a file
|
94 |
+
returns:
|
95 |
+
a string or a list of strings (in case of multiple files being uploaded)
|
96 |
+
"""
|
97 |
+
|
98 |
+
def upload():
|
99 |
+
uploaded_files = st.file_uploader(label="Upload text files only",
|
100 |
+
type="txt", help=help_msg,
|
101 |
+
accept_multiple_files=accept_multiple_files)
|
102 |
+
if st.button("Process"):
|
103 |
+
if not uploaded_files:
|
104 |
+
st.write("**No file uploaded!**")
|
105 |
+
return None
|
106 |
+
st.write("**Upload successful!**")
|
107 |
+
if type(uploaded_files) == list:
|
108 |
+
return [f.read().decode("utf-8") for f in uploaded_files]
|
109 |
+
return uploaded_files.read().decode("utf-8")
|
110 |
+
|
111 |
+
try: # to workaround the environment's Streamlit version
|
112 |
+
with st.expander(text):
|
113 |
+
return upload()
|
114 |
+
except AttributeError:
|
115 |
+
return upload()
|
116 |
+
|
117 |
|
118 |
def predict(task_type, sentence, model):
|
119 |
"""Function to parse the user inputs, run the parsed text through the
|
requirements.txt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
t5
|
2 |
huggingface_hub
|
3 |
-
streamlit
|
1 |
t5
|
2 |
huggingface_hub
|
3 |
+
streamlit==1.0.0
|