from transformers import GPT2LMHeadModel, GPT2Tokenizer

Load pre-trained model and tokenizer

model = GPT2LMHeadModel.from_pretrained("gpt2") tokenizer = GPT2Tokenizer.from_pretrained("gpt2")

Encode the input

inputs = tokenizer.encode("The quick brown fox", return_tensors="pt")

Generate text

outputs = model.generate(inputs, max_length=50, do_sample=True) print(tokenizer.decode(outputs[0], skip_special_tokens=True))

Steps

Set up Environment

conda env create -f cre_environment.yml

1. Clean Data

http://localhost:8888/notebooks/66%20CausalMap/Panasonic-IDS/notebooks/EDA.ipynb

• De-duplication
• Fix sentence tokenization

2. Parse/Extract Relations

Go to /home/fiona/79 Mass, run mass_panasonic.py

3. Post-process

cd "D:\66 CausalMap\Panasonic-IDS\src"
conda activate cre
python process_extraction.py

Original workings at: http://localhost:8888/notebooks/79%20Mass/notebooks/Process%20Extraction.ipynb

4. Cluster Nodes

cd "D:\66 CausalMap\Panasonic-IDS\src"
conda activate cre
python convert_to_graph.py
python get_nodes_infos.py
python cluster_nodes.py

5. Display Graph

Test Set

Get Annotations

1. http://localhost:8888/notebooks/66%20CausalMap/Panasonic-IDS/notebooks/Format%20Annotated%20Data.ipynb
2. Fix annotations in CSV
3. http://localhost:8888/notebooks/66%20CausalMap/Panasonic-IDS/notebooks/Format%20Annotated%20Data.ipynb Get Predictions
4. http://localhost:8888/notebooks/66%20CausalMap/Panasonic-IDS/notebooks/Evaluate%20Extraction.ipynb Evaluate Extraction, Store Results
5. http://localhost:8888/notebooks/66%20CausalMap/Panasonic-IDS/notebooks/EDA_nerevaluate.ipynb
6. "D:\66 CausalMap\Paper\paper_tables.xlsx"