Upload 6 files

app.py

@@ -0,0 +1,31 @@
+import gradio as gr
+from tape import ProteinBertModel, ProteinBertConfig, TAPETokenizer  # type: ignore
+from tape.models import modeling_bert
+import numpy as np
+import torch
+
+tokenizer = TAPETokenizer(vocab='iupac')
+config=modeling_bert.ProteinBertConfig(num_hidden_layers=5,num_attention_heads=8,hidden_size=400)
+
+bert_model = torch.load('models/bert.pt')
+class_model=torch.load('models/class.pt')
+
+
+
+def greet(name):
+    token_ids = torch.tensor([tokenizer.encode(name)])
+    token_ids = token_ids
+    bert_output = bert_model(token_ids)
+    class_output=class_model(bert_output[1])
+    cluster = torch.argmax(class_output, dim=1) + 1
+    cluster=cluster.item()
+
+    return "cluster "+str(cluster)
+demo = gr.Interface(
+    fn=greet,
+    # 自定义输入框
+    # 具体设置方法查看官方文档
+    inputs=gr.Textbox(lines=3, placeholder="Name Here...",label="my input"),
+    outputs="text",
+)
+demo.launch(share=True)

downstream_model.py

@@ -0,0 +1,61 @@
+import torch
+import torch.nn as nn
+from tape.models import modeling_bert
+
+
+config=modeling_bert.ProteinBertConfig(hidden_size=400)
+class ClassificationModel1(torch.nn.Module):
+    def __init__(self):
+        super(ClassificationModel1, self).__init__()
+        self.dense_layer1 = nn.Linear(1024, 2048)
+        self.dense_layer2 = nn.Linear(2048, 512)
+        self.output_layer = nn.Linear(512, 45)
+        self.dropout1 = nn.Dropout(0.2)
+        self.dropout2 = nn.Dropout(0.3)
+
+    def forward(self, protein_sequence):
+        hidden_layer_output1 = torch.relu(self.dense_layer1(protein_sequence))
+        #hidden_layer_output1 = self.dropout1(hidden_layer_output1)
+
+        hidden_layer_output2 = torch.relu(self.dense_layer2(hidden_layer_output1))
+        #hidden_layer_output2 = self.dropout2(hidden_layer_output2)
+
+        output = self.output_layer(hidden_layer_output2)
+
+        return output
+
+
+
+
+class ClassificationModel2(nn.Module):
+    def __init__(self):
+        super(ClassificationModel2,self).__init__()
+        self.attention_layer=nn.Linear(400,1)
+        self.hidden_layer=nn.Linear(400,1024)
+        self.output_layer=nn.Linear(1024,45)
+        self.relu=nn.ReLU()
+        self.dropout=nn.Dropout(0.1)
+        self.layernorm1=nn.LayerNorm(400)
+        self.layernorm2=nn.LayerNorm(1024)
+        self.attention_selfoutput= modeling_bert.ProteinBertSelfOutput(config)
+
+    def forward(self,sequence):
+        sequence=self.layernorm1(sequence)
+        attention_values=self.attention_layer(sequence)
+        attention_values=self.dropout(attention_values)
+        attention_weights=torch.softmax(attention_values,dim=1)
+
+
+
+        weighted_embeddings=sequence*attention_weights
+        attention_embeddings=torch.sum(weighted_embeddings,dim=1)
+
+        attention_embeddings=self.attention_selfoutput(attention_embeddings,attention_embeddings)
+        attention_embeddings=self.dropout(attention_embeddings)
+        attention_embeddings = self.attention_selfoutput(attention_embeddings, attention_embeddings)
+
+        hidden_output=self.dropout(self.relu(self.hidden_layer(attention_embeddings)))
+        hidden_output=self.layernorm2(hidden_output)
+        output=self.output_layer(hidden_output)
+
+        return output

flagged/log.csv

@@ -0,0 +1,2 @@
+my input,output,flag,username,timestamp
+,,,,2023-12-18 20:31:08.484538

models/bert.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d8202c06d805df2ca08b9a460e71cb651eb7deb594f37c13c15e685d77e02b5d
+size 80272735

models/class.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:54ffee6f21ff8d0c743fbf8cad638ed609030e79a6359ec115005c98f0cef9c7
+size 2491739