Spaces:
No application file
No application file
| # -*- coding: utf-8 -*- | |
| """final_gradio.ipynb | |
| Automatically generated by Colaboratory. | |
| Original file is located at | |
| https://colab.research.google.com/drive/1Laxh069wv-cX4NqcOnNB2AyEDW0gmvbL | |
| """ | |
| from google.colab import drive | |
| drive.mount('/content/drive') | |
| !pip install gradio | |
| !pip install transformers==3.0.2 | |
| import pandas as pd | |
| import numpy as np | |
| from sklearn.model_selection import train_test_split | |
| import torch | |
| import seaborn as sns | |
| import transformers | |
| import json | |
| from tqdm import tqdm | |
| from torch.utils.data import Dataset, DataLoader | |
| from transformers import RobertaModel, RobertaTokenizer | |
| import logging | |
| logging.basicConfig(level=logging.ERROR) | |
| from torch import cuda | |
| device = 'cuda' if cuda.is_available() else 'cpu' | |
| class RobertaClass(torch.nn.Module): | |
| def __init__(self): | |
| super(RobertaClass, self).__init__() | |
| self.l1 = RobertaModel.from_pretrained("roberta-base") | |
| self.pre_classifier = torch.nn.Linear(768, 768) | |
| self.dropout = torch.nn.Dropout(0.3) | |
| self.classifier = torch.nn.Linear(768, 5) | |
| def forward(self, input_ids, attention_mask, token_type_ids): | |
| output_1 = self.l1(input_ids=input_ids, attention_mask=attention_mask, token_type_ids=token_type_ids) | |
| hidden_state = output_1[0] | |
| pooler = hidden_state[:, 0] | |
| pooler = self.pre_classifier(pooler) | |
| pooler = torch.nn.ReLU()(pooler) | |
| pooler = self.dropout(pooler) | |
| output = self.classifier(pooler) | |
| return output | |
| model = RobertaClass() | |
| model.to(device) | |
| #url = f"https://github.com/udacity/deep-learning-v2-pytorch/blob/master/convolutional-neural-networks/mnist-mlp/model.pt" | |
| #!wget --no-cache --backups=1 {url} | |
| #model.load_state_dict(torch.load("/content/model.pt")) | |
| model.load_state_dict(torch.load("/content/drive/MyDrive/avicenna_model.pt")) | |
| model.eval() | |
| class SyllogismData(Dataset): | |
| def __init__(self, dataframe, tokenizer, max_len): | |
| self.tokenizer = tokenizer | |
| self.data = dataframe | |
| self.text = dataframe.Premises | |
| self.targets = self.data.label | |
| self.max_len = max_len | |
| def __len__(self): | |
| return len(self.text) | |
| def __getitem__(self, index): | |
| text = str(self.text[index]) | |
| text = " ".join(text.split()) | |
| inputs = self.tokenizer.encode_plus( | |
| text, | |
| None, | |
| add_special_tokens=True, | |
| max_length=self.max_len, | |
| pad_to_max_length=True, | |
| return_token_type_ids=True | |
| ) | |
| ids = inputs['input_ids'] | |
| mask = inputs['attention_mask'] | |
| token_type_ids = inputs["token_type_ids"] | |
| return { | |
| 'ids': torch.tensor(ids, dtype=torch.long), | |
| 'mask': torch.tensor(mask, dtype=torch.long), | |
| 'token_type_ids': torch.tensor(token_type_ids, dtype=torch.long), | |
| 'targets': torch.tensor(self.targets[index], dtype=torch.float) | |
| } | |
| MAX_LEN = 256 | |
| TRAIN_BATCH_SIZE = 8 | |
| VALID_BATCH_SIZE = 4 | |
| LEARNING_RATE = 1e-05 | |
| tokenizer = RobertaTokenizer.from_pretrained('roberta-base', truncation=True, do_lower_case=True) | |
| def avicenna(input): | |
| p1=input | |
| data = [[p1, 10]] | |
| testnew_df = pd.DataFrame(data, columns=['Premises', 'label']) | |
| data | |
| test_size = 1 | |
| test_data=testnew_df | |
| testing_set = SyllogismData(test_data, tokenizer, MAX_LEN) | |
| test_params = {'batch_size': VALID_BATCH_SIZE, | |
| 'shuffle': False, | |
| 'num_workers': 0 | |
| } | |
| testing_loader = DataLoader(testing_set, **test_params) | |
| model.eval() | |
| n_correct = 0; n_wrong = 0; total = 0; tr_loss=0; nb_tr_steps=0; nb_tr_examples=0 | |
| with torch.no_grad(): | |
| for _, data in tqdm(enumerate(testing_loader, 0)): | |
| ids = data['ids'].to(device, dtype = torch.long) | |
| mask = data['mask'].to(device, dtype = torch.long) | |
| token_type_ids = data['token_type_ids'].to(device, dtype=torch.long) | |
| targets = data['targets'].to(device, dtype = torch.long) | |
| outputs = model(ids, mask, token_type_ids).squeeze() | |
| big_val, big_idx = torch.max(outputs.data, dim=0) | |
| return big_idx.item() | |
| #valid(model, testing_loader) | |
| #print(big_idx) | |
| import gradio as gr | |
| iface = gr.Interface( | |
| fn = avicenna, | |
| title="Syllogistic NLI", | |
| description="Select pair of sentences and see if the Avicenna-trained model can gauge the relation correctly", | |
| #false | |
| #inputs = gr.inputs.Dropdown(["Buildings should not be constructed on organic soils. All buildings should be constructed on stiff underlying soil with enough strength.","The launching of satellites while still contributing to national prestige is a significant economic activity. For each type of economic activity, there is some threshold for the launch cost per payload mass beyond which the economic activity will not be sustainable."], label="Select an example from Avicenna test set"), | |
| #true | |
| inputs = gr.inputs.Dropdown(["All humans are mortal. Socrates is a human.","Avicenna wrote the famous book the Canon of Medicine. The Canon of Medicine has influenced modern medicine","Police found signs of active bleeding before death around the corpse. A large volume of blood released from a body may indicate that the individual has died of exsanguination.","During the first trimester of pregnancy, the body undergoes hormonal fluctuations. Hormonal changes regularly are followed by extreme tiredness.","Influenza spreads the virus to the lungs. Garlic is useful to cure all infections.","With a single currency, there will no longer be a cost involved in changing currencies. It was no longer cost-effective for the government to convert metals into coins.","Pain and tension around your head and neck are known as tension headaches. Tension headaches are dull pain, tightness, or pressure around your head and neck.","Heavy rain can cause flooding. Many different health conditions can cause heavy breathing.","Eating foods that are in the Mediterranean diet helps with healthy weight loss and metabolism. Ana knows how to eat healthy and lose weight with the Mediterranean diet."], label="Select an example"), | |
| #f and t | |
| #inputs = gr.inputs.Dropdown(["All humans are mortal. Socrates is a human.","Avicenna wrote the famous book the Canon of Medicine. The Canon of Medicine has influenced modern medicine","Police found signs of active bleeding before death around the corpse. A large volume of blood released from a body may indicate that the individual has died of exsanguination.","During the first trimester of pregnancy, the body undergoes hormonal fluctuations. Hormonal changes regularly are followed by extreme tiredness.","Influenza spreads the virus to the lungs. Garlic is useful to cure all infections.","With a single currency, there will no longer be a cost involved in changing currencies. It was no longer cost-effective for the government to convert metals into coins.","Pain and tension around your head and neck are known as tension headaches. Tension headaches are dull pain, tightness, or pressure around your head and neck.","Heavy rain can cause flooding. Many different health conditions can cause heavy breathing.","Eating foods that are in the Mediterranean diet helps with healthy weight loss and metabolism. Ana knows how to eat healthy and lose weight with the Mediterranean diet.","Buildings should not be constructed on organic soils. All buildings should be constructed on stiff underlying soil with enough strength.","The launching of satellites while still contributing to national prestige is a significant economic activity. For each type of economic activity, there is some threshold for the launch cost per payload mass beyond which the economic activity will not be sustainable."], label="Select an example"), | |
| outputs = gr.outputs.Textbox(label="Syllogistic relation") | |
| ) | |
| iface.launch(debug=True,enable_queue=True)) |