Add application file

app.py

@@ -0,0 +1,71 @@
+from transformers import (
+    AlbertTokenizerFast, DistilBertTokenizerFast, RobertaTokenizerFast,
+    AlbertForSequenceClassification, DistilBertForSequenceClassification, RobertaForSequenceClassification
+)
+import torch.nn.functional as F
+import torch
+import gradio as gr
+
+from gradio.components import Textbox
+
+albert_tokenizer = AlbertTokenizerFast.from_pretrained('albert-base-v2')
+bert_tokenizer = DistilBertTokenizerFast.from_pretrained('distilbert-base-uncased')
+roberta_tokenizer = RobertaTokenizerFast.from_pretrained('distilroberta-base')
+
+albert_model = AlbertForSequenceClassification.from_pretrained("./model_albert")
+bert_model = DistilBertForSequenceClassification.from_pretrained("./model_bert")
+roberta_model = RobertaForSequenceClassification.from_pretrained("./model_roberta")
+
+def majority_voting(votes):
+    vote_result = max(set(votes), key=votes.count)
+    return vote_result
+
+def predict_news(news_title, news_text):
+    combined_text = f"{news_title} - {news_text}"
+    
+    albert_input = albert_tokenizer(combined_text, return_tensors='pt', truncation=True, padding=True)
+    bert_input = bert_tokenizer(combined_text, return_tensors='pt', truncation=True, padding=True)
+    roberta_input = roberta_tokenizer(combined_text, return_tensors='pt', truncation=True, padding=True)
+    
+    with torch.no_grad():
+        albert_logits = albert_model(**albert_input).logits
+        bert_logits = bert_model(**bert_input).logits
+        roberta_logits = roberta_model(**roberta_input).logits
+
+    albert_probs = F.softmax(albert_logits, dim=-1)
+    bert_probs = F.softmax(bert_logits, dim=-1)
+    roberta_probs = F.softmax(roberta_logits, dim=-1)
+
+    albert_pred_class = albert_probs.argmax(-1).item()
+    bert_pred_class = bert_probs.argmax(-1).item()
+    roberta_pred_class = roberta_probs.argmax(-1).item()
+
+    albert_pred_prob = albert_probs[0, albert_pred_class].item()
+    bert_pred_prob = bert_probs[0, bert_pred_class].item()
+    roberta_pred_prob = roberta_probs[0, roberta_pred_class].item()
+
+    average_confidence = (albert_pred_prob + bert_pred_prob + roberta_pred_prob) / 3
+
+    final_pred_class = majority_voting([albert_pred_class, bert_pred_class, roberta_pred_class])
+
+    prediction = 'REAL' if final_pred_class == 1 else 'FAKE'
+    return prediction, f"{average_confidence:.2f}", f"{albert_pred_prob:.2f}", f"{bert_pred_prob:.2f}", f"{roberta_pred_prob:.2f}"
+
+iface = gr.Interface(
+    fn=predict_news,
+    inputs=[
+        Textbox(lines=2, label="News Title", placeholder="Enter News Title Here..."),
+        Textbox(lines=7, label="News Text", placeholder="Enter News Text Here...")
+    ],
+    outputs=[
+        Textbox(label="Prediction"),
+        Textbox(label="Average Confidence"),
+        Textbox(label="ALBERT Confidence"),
+        Textbox(label="BERT Confidence"),
+        Textbox(label="RoBERTa Confidence")
+    ],
+    title="Fake News Classification with Ensemble Learning Transformer Models",
+    description="UAS Deep Learning - 2501985836 - Arya Adyatma"
+)
+
+iface.launch()

model_albert/config.json

@@ -0,0 +1,34 @@
+{
+  "_name_or_path": "albert-base-v2",
+  "architectures": [
+    "AlbertForSequenceClassification"
+  ],
+  "attention_probs_dropout_prob": 0,
+  "bos_token_id": 2,
+  "classifier_dropout_prob": 0.1,
+  "down_scale_factor": 1,
+  "embedding_size": 128,
+  "eos_token_id": 3,
+  "gap_size": 0,
+  "hidden_act": "gelu_new",
+  "hidden_dropout_prob": 0,
+  "hidden_size": 768,
+  "initializer_range": 0.02,
+  "inner_group_num": 1,
+  "intermediate_size": 3072,
+  "layer_norm_eps": 1e-12,
+  "max_position_embeddings": 512,
+  "model_type": "albert",
+  "net_structure_type": 0,
+  "num_attention_heads": 12,
+  "num_hidden_groups": 1,
+  "num_hidden_layers": 12,
+  "num_memory_blocks": 0,
+  "pad_token_id": 0,
+  "position_embedding_type": "absolute",
+  "problem_type": "single_label_classification",
+  "torch_dtype": "float32",
+  "transformers_version": "4.36.0",
+  "type_vocab_size": 2,
+  "vocab_size": 30000
+}

model_albert/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d3801acd748ab99ea2ff11f1097fd594e3f672de21ebc2fb12a82c6ff7c88051
+size 46743912

model_albert/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f7255591348fe6bb850b47d3d540f13261b2ef94508a8f005cadee19b780f4cd
+size 4347

model_bert/config.json

@@ -0,0 +1,25 @@
+{
+  "_name_or_path": "distilbert-base-uncased",
+  "activation": "gelu",
+  "architectures": [
+    "DistilBertForSequenceClassification"
+  ],
+  "attention_dropout": 0.1,
+  "dim": 768,
+  "dropout": 0.1,
+  "hidden_dim": 3072,
+  "initializer_range": 0.02,
+  "max_position_embeddings": 512,
+  "model_type": "distilbert",
+  "n_heads": 12,
+  "n_layers": 6,
+  "pad_token_id": 0,
+  "problem_type": "single_label_classification",
+  "qa_dropout": 0.1,
+  "seq_classif_dropout": 0.2,
+  "sinusoidal_pos_embds": false,
+  "tie_weights_": true,
+  "torch_dtype": "float32",
+  "transformers_version": "4.36.0",
+  "vocab_size": 30522
+}

model_bert/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6dd5ae00c4acfaada74cca42d9181dc55e9cc907c84f6eb806bd24103e191944
+size 267832560

model_bert/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6b01cfc38acaaaa60208e5937e4dd3ac6ab311aef216d5c12b1474f7a7fe513e
+size 4155

model_roberta/config.json

@@ -0,0 +1,28 @@
+{
+  "_name_or_path": "distilroberta-base",
+  "architectures": [
+    "RobertaForSequenceClassification"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "bos_token_id": 0,
+  "classifier_dropout": null,
+  "eos_token_id": 2,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 768,
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "layer_norm_eps": 1e-05,
+  "max_position_embeddings": 514,
+  "model_type": "roberta",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 6,
+  "pad_token_id": 1,
+  "position_embedding_type": "absolute",
+  "problem_type": "single_label_classification",
+  "torch_dtype": "float32",
+  "transformers_version": "4.36.0",
+  "type_vocab_size": 1,
+  "use_cache": true,
+  "vocab_size": 50265
+}

model_roberta/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:38b9bdb64964633cb7a1fd069c326a35e118a8c798175c99a8b417c65ed1ae8e
+size 328492280

model_roberta/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ef2cea918d8f268fa604d0638b5947092b595b80cc8417f1e63a28ab520199c7
+size 4155

requirements.txt

@@ -0,0 +1,3 @@
+gradio==3.50
+transformers
+torch