File size: 2,454 Bytes
1024bd2 7a58881 1024bd2 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 |
---
license: cc0-1.0
language:
- de
tags:
- natural-language-processing
- poetry-generation
- text-generation
- torch
- lstm
---
This generative text model was trained using [Andrej Karpathy's code](https://github.com/karpathy/char-rnn) on texts of German [Hexameter](https://en.wikipedia.org/w/index.php?title=Hexameter&oldid=1198526559). Models of this type well represent individual style.
Model was trained with size 512 and 3 layers, dropout 0.5.
## Usage
The procedure for installing the required software is described [by Karpathy](https://github.com/karpathy/char-rnn), torch is required, the code is written in lua. Be careful, versions of libraries written many years ago are used!
```bash
th sample.lua hexameter_lm_lstm_epoch80.00_1.3702.t7
```
## Train data
Hexameter lines extracted from [large collection of German verses](https://github.com/tnhaider/DLK) running by Thomas Haider. Included in this repository as `input.txt`.
## What for?
In an era of winning Transformers, ancient RNN models seem archaic. But I see that they still work better than modern architectures with such important categories from the humanities point of view as poetic style.
## Dataset
There is a [repository](https://github.com/nevmenandr/german-generated-poetic-texts) that publishes German poetic texts generated by RNN LSTM models (inclding this one) with different temperature.
## Publication
There are some texts explaining the goal o these poetic experiments and their place in the history of human culture.
* "[Der digitale Superdichter](http://nevmenandr.net/personalia/holderlin.pdf). Vor 250 Jahren wurde Friedrich Hölderlin geboren. Heute kann Computertechnik neue Gedichte im Hölderlin-Sound generieren. Ein Werkstattbericht" *Die Literarische Welt*, 14 March 2020, p. 29. (included in this repository as pdf)
* Orekhov, Boris, and Frank Fischer. "Neural reading: Insights from the analysis of poetry generated by artificial neural networks." *Orbis Litterarum* 75.5 (2020): 230-246. [DOI: 10.1111/oli.12274](https://onlinelibrary.wiley.com/doi/abs/10.1111/oli.12274)
## BibTeX entry and citation info
```
@article{orekhov2020neural,
title={Neural reading: Insights from the analysis of poetry generated by artificial neural networks},
author={Orekhov, Boris and Fischer, Frank},
journal={Orbis Litterarum},
volume={75},
number={5},
pages={230--246},
year={2020},
publisher={Wiley Online Library}
}
```
|