Spaces:
Sleeping
Sleeping
ndhieunguyen
commited on
Commit
·
ad32d4f
1
Parent(s):
23a7a4b
feat: use gpu space
Browse files
app.py
CHANGED
@@ -6,6 +6,7 @@ from src.improved_diffusion import gaussian_diffusion as gd
|
|
6 |
from src.improved_diffusion.respace import SpacedDiffusion
|
7 |
from src.improved_diffusion.transformer_model import TransformerNetModel
|
8 |
import streamlit as st
|
|
|
9 |
import os
|
10 |
|
11 |
|
@@ -57,28 +58,18 @@ def get_diffusion():
|
|
57 |
training_mode="e2e",
|
58 |
)
|
59 |
|
60 |
-
|
61 |
-
|
62 |
-
|
63 |
-
tokenizer = get_tokenizer()
|
64 |
-
encoder = get_encoder(device)
|
65 |
-
model = get_model(device)
|
66 |
-
diffusion = get_diffusion()
|
67 |
-
|
68 |
-
st.title("Lang2mol-Diff")
|
69 |
-
text_input = st.text_area("Enter molecule description")
|
70 |
-
button = st.button("Submit")
|
71 |
-
if button:
|
72 |
with st.spinner("Please wait..."):
|
73 |
output = tokenizer(
|
74 |
-
|
75 |
-
|
76 |
-
|
77 |
-
|
78 |
-
|
79 |
-
|
80 |
-
|
81 |
-
|
82 |
caption_state = encoder(
|
83 |
input_ids=output["input_ids"].to(device),
|
84 |
attention_mask=output["attention_mask"].to(device),
|
@@ -103,5 +94,18 @@ if button:
|
|
103 |
result = sf.decoder(
|
104 |
outputs[0].replace("<pad>", "").replace("</s>", "").replace("\t", "")
|
105 |
).replace("\t", "")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
106 |
|
107 |
-
|
|
|
|
|
|
|
|
|
|
|
|
6 |
from src.improved_diffusion.respace import SpacedDiffusion
|
7 |
from src.improved_diffusion.transformer_model import TransformerNetModel
|
8 |
import streamlit as st
|
9 |
+
import spaces
|
10 |
import os
|
11 |
|
12 |
|
|
|
58 |
training_mode="e2e",
|
59 |
)
|
60 |
|
61 |
+
@spaces.GPU
|
62 |
+
def generate(text_input):
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
63 |
with st.spinner("Please wait..."):
|
64 |
output = tokenizer(
|
65 |
+
text_input,
|
66 |
+
max_length=256,
|
67 |
+
truncation=True,
|
68 |
+
padding="max_length",
|
69 |
+
add_special_tokens=True,
|
70 |
+
return_tensors="pt",
|
71 |
+
return_attention_mask=True,
|
72 |
+
)
|
73 |
caption_state = encoder(
|
74 |
input_ids=output["input_ids"].to(device),
|
75 |
attention_mask=output["attention_mask"].to(device),
|
|
|
94 |
result = sf.decoder(
|
95 |
outputs[0].replace("<pad>", "").replace("</s>", "").replace("\t", "")
|
96 |
).replace("\t", "")
|
97 |
+
return result
|
98 |
+
|
99 |
+
device = torch.device("cuda") if torch.cuda.is_available() else torch.device("cpu")
|
100 |
+
|
101 |
+
tokenizer = get_tokenizer()
|
102 |
+
encoder = get_encoder(device)
|
103 |
+
model = get_model(device)
|
104 |
+
diffusion = get_diffusion()
|
105 |
|
106 |
+
st.title("Lang2mol-Diff")
|
107 |
+
text_input = st.text_area("Enter molecule description")
|
108 |
+
button = st.button("Submit")
|
109 |
+
if button:
|
110 |
+
result = generate(text_input)
|
111 |
+
st.write(result)
|