The model I will introduce today is Dleunji’s Kant model.
The ‘Kant Model’ is a sentence generation model that writes an essay about Kant when you enter an introduction.
Our teammate, Dleunji, said that during her philosophy major, she had to read Kant’s work all the time and she found that this philosopher’s thoughts were so complex and profound that she had a really hard time with the assignments, so she made a model to solve this problem.
If you would like to check the project right away, please refer to the following link.
What do you think when you hear the word ‘philosophy’? In my case, I think that philosophy is profound and difficult. Then, since philosophy is so difficult and profound, can it be excluded from life? I don’t think so. We believe that philosophy can improve our way of thinking about important everyday problems.
However, it is not easy for people with little to no knowledge of philosophy to think philosophically. It would be nice to ask for guidance to the philosophers of the past directly, but needless to say, they have already passed away. So how about asking artificial intelligence, which can learn all the philosopher’s thoughts and doctrine. If I ask to AI, will it give a similar answer to the philosopher’s thoughts? Let’s take a look at the model trained on this idea based on Kant!
The critique of pure reason is one of the most influential books in the history of philosophy. The original text of the criticism of pure reason is available from Project Gutenburg.
When I looked at the data for data preprocessing, there were unnecessary headings, spaces, and symbols, so I proceeded to remove them.
- Using the DEMO provided by Ainize
Let’s use the DEMO provided by Ainize.
Enter the text you want in the heading and Kant will write the rest in accordance with the philosopher’s thoughts . The DEMO can be found on the link
- Using APIs provided by Ainize
Next, we will use the Kant model using the Kant API provided by Ainize. You can check the information on the API in the link .
- Train and use the model yourself
Now, we will use the model by fine-tuning GPT-2 through Teachable NLP .
In order to fine-tune the GPT-2 model, first upload the pre-processed data in Teachable-NLP and set the model size to small and epoch to 5. Now, when the training process is over, click Test your model.
You can become Kant in the TabTab screen that opens when you click it.
I will write in the word judgment.
judgement of the principles of morality; for the judgement, to which lies the foundation of human reason, relates to itself, and must consequently be identical with morality, and is the origin of a kind of à priori knowledge. The attempt to make an à priori representation of a thinking being in general, and to construct a judgement in the relation of the determining cause of this thinking being, will here give us only a partial knowledge of the subject. For, to make a cognition of it (judgements being merely the judgements of persuasion), we must have an experience, and therefore also a principle of reason, which explains the empirical conditions under which we are affected by effects. If an…
I think this model is really well trained. I don’t even know what it’s talking about.
This model was trained with Kant’s pure rational criticism, but furthermore, if you train the model through other books written by Kant, it will have the same effect and perhaps new Kant’s ideas may emerge. This is the end of the article. Thanks for reading!