Update app.py
Browse files
app.py
CHANGED
@@ -7,7 +7,7 @@ from lavis.models.protein_models.protein_function_opt import Blip2ProteinMistral
|
|
7 |
from lavis.models.base_model import FAPMConfig
|
8 |
import spaces
|
9 |
import gradio as gr
|
10 |
-
from esm_scripts.extract import run_demo
|
11 |
from esm import pretrained, FastaBatchedDataset
|
12 |
from data.evaluate_data.utils import Ontology
|
13 |
import difflib
|
@@ -15,9 +15,29 @@ import re
|
|
15 |
|
16 |
|
17 |
# Load the model
|
18 |
-
model = Blip2ProteinMistral(config=FAPMConfig(), esm_size='3b')
|
19 |
-
model.load_checkpoint("model/checkpoint_mf2.pth")
|
20 |
-
model.to('cuda')
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
21 |
|
22 |
model_esm, alphabet = pretrained.load_model_and_alphabet('esm2_t36_3B_UR50D')
|
23 |
model_esm.to('cuda')
|
@@ -40,7 +60,7 @@ choices = {x.lower(): x for x in choices_mf}
|
|
40 |
|
41 |
|
42 |
@spaces.GPU
|
43 |
-
def generate_caption(protein, prompt):
|
44 |
# Process the image and the prompt
|
45 |
# with open('/home/user/app/example.fasta', 'w') as f:
|
46 |
# f.write('>{}\n'.format("protein_name"))
|
@@ -123,6 +143,7 @@ def generate_caption(protein, prompt):
|
|
123 |
'text_input': ['none'],
|
124 |
'prompt': [prompt]}
|
125 |
|
|
|
126 |
# Generate the output
|
127 |
prediction = model.generate(samples, length_penalty=0., num_beams=15, num_captions=10, temperature=1.,
|
128 |
repetition_penalty=1.0)
|
@@ -151,7 +172,6 @@ description = """Quick demonstration of the FAPM model for protein function pred
|
|
151 |
|
152 |
The model used in this app is available at [Hugging Face Model Hub](https://huggingface.co/wenkai/FAPM) and the source code can be found on [GitHub](https://github.com/xiangwenkai/FAPM/tree/main)."""
|
153 |
|
154 |
-
|
155 |
# iface = gr.Interface(
|
156 |
# fn=generate_caption,
|
157 |
# inputs=[gr.Textbox(type="text", label="Upload sequence"), gr.Textbox(type="text", label="Prompt")],
|
@@ -161,7 +181,6 @@ The model used in this app is available at [Hugging Face Model Hub](https://hugg
|
|
161 |
# # Launch the interface
|
162 |
# iface.launch()
|
163 |
|
164 |
-
|
165 |
css = """
|
166 |
#output {
|
167 |
height: 500px;
|
@@ -175,34 +194,29 @@ with gr.Blocks(css=css) as demo:
|
|
175 |
with gr.Tab(label="Protein caption"):
|
176 |
with gr.Row():
|
177 |
with gr.Column():
|
|
|
178 |
input_protein = gr.Textbox(type="text", label="Upload sequence")
|
179 |
-
|
180 |
-
prompt = gr.Textbox(type="text", label="Taxonomy Prompt")
|
181 |
submit_btn = gr.Button(value="Submit")
|
182 |
with gr.Column():
|
183 |
output_text = gr.Textbox(label="Output Text")
|
184 |
-
# train index
|
185 |
gr.Examples(
|
186 |
examples=[
|
187 |
-
["
|
188 |
-
["
|
189 |
-
["MKTLALFLVLVCVLGLVQSWEWPWNRKPTKFPIPSPNPRDKWCRLNLGPAWGGRC", 'Sophophora'],
|
190 |
-
["MAARGAMLRYLRVNVNPTIQNPRECVLPFSILLRRFSEEVRGSFLDKSEVTDRVLSVVKNFQKVDPSKVTPKANFQNDLGLDSLDSVEVVMALEEEFGFEIPDNEADKIQSIDLAVDFIASHPQAK", 'Arabidopsis'],
|
191 |
["MAAAGGARLLRAASAVLGGPAGRWLHHAGSRAGSSGLLRNRGPGGSAEASRSLSVSARARSSSEDKITVHFINRDGETLTTKGKVGDSLLDVVVENNLDIDGFGACEGTLACSTCHLIFEDHIYEKLDAITDEENDMLDLAYGLTDRSRLGCQICLTKSMDNMTVRVPETVADARQSIDVGKTS", 'Homo'],
|
192 |
['MASAELSREENVYMAKLAEQAERYEEMVEFMEKVAKTVDSEELTVEERNLLSVAYKNVIGARRASWRIISSIEQKEEGRGNEDRVTLIKDYRGKIETELTKICDGILKLLETHLVPSSTAPESKVFYLKMKGDYYRYLAEFKTGAERKDAAENTMVAYKAAQDIALAELAPTHPIRLGLALNFSVFYYEILNSPDRACSLAKQAFDEAISELDTLSEESYKDSTLIMQLLRDNLTLWTSDISEDPAEEIREAPKRDSSEGQ', 'Zea'],
|
193 |
['MIKAAVTKESLYRMNTLMEAFQGFLGLDLGEFTFKVKPGVFLLTDVKSYLIGDKYDDAFNALIDFVLRNDRDAVEGTETDVSIRLGLSPSDMVVKRQDKTFTFTHGDLEFEVHWINL', 'Bacteriophage'],
|
194 |
-
['MQMYKLTAGTTGYHTLLTRTQAEHMLSLWGDKYSIDDCTPSNPIYSPSRYTKLELVYMAANATA', 'Bacteriophage'],
|
195 |
-
['MSITAMDAKLQRILEESTCFGIGHDPNVKECKMCDVREQCKAKTQGMNVPTPTRKKPEDVAPAKEKPTTKKTTAKKSTAKEEKKETAPKAKETKAKPKSKPKKAKAPENPNLPNFKEMSFEELVELAKERNVEWKDYNSPNITRMRLIMALKASY', 'Bacteriophage'],
|
196 |
['MNDLMIQLLDQFEMGLRERAIKVMATINDEKHRFPMELNKKQCSLMLLGTTDTTTFDMRFNSKKDFPRIKGAREKYPRDAVIEWYHQNWMRTEVKQ', 'Bacteriophage'],
|
197 |
],
|
198 |
-
inputs=[input_protein, prompt],
|
199 |
outputs=[output_text],
|
200 |
fn=generate_caption,
|
201 |
cache_examples=True,
|
202 |
label='Try examples'
|
203 |
)
|
204 |
-
|
205 |
-
submit_btn.click(generate_caption, [input_protein, prompt], [output_text])
|
206 |
|
207 |
demo.launch(debug=True)
|
208 |
|
|
|
7 |
from lavis.models.base_model import FAPMConfig
|
8 |
import spaces
|
9 |
import gradio as gr
|
10 |
+
# from esm_scripts.extract import run_demo
|
11 |
from esm import pretrained, FastaBatchedDataset
|
12 |
from data.evaluate_data.utils import Ontology
|
13 |
import difflib
|
|
|
15 |
|
16 |
|
17 |
# Load the model
|
18 |
+
# model = Blip2ProteinMistral(config=FAPMConfig(), esm_size='3b')
|
19 |
+
# model.load_checkpoint("model/checkpoint_mf2.pth")
|
20 |
+
# model.to('cuda')
|
21 |
+
|
22 |
+
def get_model(type='Molecule Function'):
|
23 |
+
model = Blip2ProteinMistral(config=FAPMConfig(), esm_size='3b')
|
24 |
+
if type == 'Molecule Function':
|
25 |
+
model.load_checkpoint("model/checkpoint_mf2.pth")
|
26 |
+
model.to('cuda')
|
27 |
+
elif type == 'Biological Process':
|
28 |
+
model.load_checkpoint("model/checkpoint_bp1.pth")
|
29 |
+
model.to('cuda')
|
30 |
+
elif type == 'Cellar Component':
|
31 |
+
model.load_checkpoint("model/checkpoint_cc2.pth")
|
32 |
+
model.to('cuda')
|
33 |
+
|
34 |
+
|
35 |
+
models = {
|
36 |
+
'Molecule Function': get_model('Molecule Function'),
|
37 |
+
'Biological Process': get_model('Biological Process'),
|
38 |
+
'Cellar Component': get_model('Cellar Component'),
|
39 |
+
}
|
40 |
+
|
41 |
|
42 |
model_esm, alphabet = pretrained.load_model_and_alphabet('esm2_t36_3B_UR50D')
|
43 |
model_esm.to('cuda')
|
|
|
60 |
|
61 |
|
62 |
@spaces.GPU
|
63 |
+
def generate_caption(protein, prompt, model_id):
|
64 |
# Process the image and the prompt
|
65 |
# with open('/home/user/app/example.fasta', 'w') as f:
|
66 |
# f.write('>{}\n'.format("protein_name"))
|
|
|
143 |
'text_input': ['none'],
|
144 |
'prompt': [prompt]}
|
145 |
|
146 |
+
model = models[model_id]
|
147 |
# Generate the output
|
148 |
prediction = model.generate(samples, length_penalty=0., num_beams=15, num_captions=10, temperature=1.,
|
149 |
repetition_penalty=1.0)
|
|
|
172 |
|
173 |
The model used in this app is available at [Hugging Face Model Hub](https://huggingface.co/wenkai/FAPM) and the source code can be found on [GitHub](https://github.com/xiangwenkai/FAPM/tree/main)."""
|
174 |
|
|
|
175 |
# iface = gr.Interface(
|
176 |
# fn=generate_caption,
|
177 |
# inputs=[gr.Textbox(type="text", label="Upload sequence"), gr.Textbox(type="text", label="Prompt")],
|
|
|
181 |
# # Launch the interface
|
182 |
# iface.launch()
|
183 |
|
|
|
184 |
css = """
|
185 |
#output {
|
186 |
height: 500px;
|
|
|
194 |
with gr.Tab(label="Protein caption"):
|
195 |
with gr.Row():
|
196 |
with gr.Column():
|
197 |
+
model_selector = gr.Dropdown(choices=list(models.keys()), label="Model", value='Molecule Function')
|
198 |
input_protein = gr.Textbox(type="text", label="Upload sequence")
|
199 |
+
prompt = gr.Textbox(type="text", label="Taxonomy Prompt (Optional)")
|
|
|
200 |
submit_btn = gr.Button(value="Submit")
|
201 |
with gr.Column():
|
202 |
output_text = gr.Textbox(label="Output Text")
|
203 |
+
# O14813 train index 127, 266, 738, 1060 test index 4
|
204 |
gr.Examples(
|
205 |
examples=[
|
206 |
+
["MDYSYLNSYDSCVAAMEASAYGDFGACSQPGGFQYSPLRPAFPAAGPPCPALGSSNCALGALRDHQPAPYSAVPYKFFPEPSGLHEKRKQRRIRTTFTSAQLKELERVFAETHYPDIYTREELALKIDLTEARVQVWFQNRRAKFRKQERAASAKGAAGAAGAKKGEARCSSEDDDSKESTCSPTPDSTASLPPPPAPGLASPRLSPSPLPVALGSGPGPGPGPQPLKGALWAGVAGGGGGGPGAGAAELLKAWQPAESGPGPFSGVLSSFHRKPGPALKTNLF", ''],
|
207 |
+
["MKTLALFLVLVCVLGLVQSWEWPWNRKPTKFPIPSPNPRDKWCRLNLGPAWGGRC", ''],
|
|
|
|
|
208 |
["MAAAGGARLLRAASAVLGGPAGRWLHHAGSRAGSSGLLRNRGPGGSAEASRSLSVSARARSSSEDKITVHFINRDGETLTTKGKVGDSLLDVVVENNLDIDGFGACEGTLACSTCHLIFEDHIYEKLDAITDEENDMLDLAYGLTDRSRLGCQICLTKSMDNMTVRVPETVADARQSIDVGKTS", 'Homo'],
|
209 |
['MASAELSREENVYMAKLAEQAERYEEMVEFMEKVAKTVDSEELTVEERNLLSVAYKNVIGARRASWRIISSIEQKEEGRGNEDRVTLIKDYRGKIETELTKICDGILKLLETHLVPSSTAPESKVFYLKMKGDYYRYLAEFKTGAERKDAAENTMVAYKAAQDIALAELAPTHPIRLGLALNFSVFYYEILNSPDRACSLAKQAFDEAISELDTLSEESYKDSTLIMQLLRDNLTLWTSDISEDPAEEIREAPKRDSSEGQ', 'Zea'],
|
210 |
['MIKAAVTKESLYRMNTLMEAFQGFLGLDLGEFTFKVKPGVFLLTDVKSYLIGDKYDDAFNALIDFVLRNDRDAVEGTETDVSIRLGLSPSDMVVKRQDKTFTFTHGDLEFEVHWINL', 'Bacteriophage'],
|
|
|
|
|
211 |
['MNDLMIQLLDQFEMGLRERAIKVMATINDEKHRFPMELNKKQCSLMLLGTTDTTTFDMRFNSKKDFPRIKGAREKYPRDAVIEWYHQNWMRTEVKQ', 'Bacteriophage'],
|
212 |
],
|
213 |
+
inputs=[input_protein, prompt, model_selector],
|
214 |
outputs=[output_text],
|
215 |
fn=generate_caption,
|
216 |
cache_examples=True,
|
217 |
label='Try examples'
|
218 |
)
|
219 |
+
submit_btn.click(generate_caption, [input_protein, prompt, model_selector], [output_text])
|
|
|
220 |
|
221 |
demo.launch(debug=True)
|
222 |
|