nevmenandr commited on
Commit
1024bd2
1 Parent(s): 92dc4d7

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +58 -3
README.md CHANGED
@@ -1,3 +1,58 @@
1
- ---
2
- license: mit
3
- ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: cc0-1.0
3
+ language:
4
+ - de
5
+ tags:
6
+ - natural-language-processing
7
+ - poetry-generation
8
+ - text-generation
9
+ - torch
10
+ - lstm
11
+ ---
12
+
13
+ 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.
14
+
15
+ Model was trained with size 512 and 3 layers, dropout 0.5.
16
+
17
+ ## Usage
18
+
19
+ 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!
20
+
21
+ ```bash
22
+ th sample.lua hexameter_lm_lstm_epoch80.00_1.3702.t7
23
+ ```
24
+
25
+ ## Train data
26
+
27
+ Hexameter lines extracted from [large collection of German verses](https://github.com/tnhaider/DLK) running by Thomas Haider.
28
+
29
+ ## What for?
30
+
31
+ 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.
32
+
33
+ ## Dataset
34
+
35
+ 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.
36
+
37
+ ## Publication
38
+
39
+ There are some texts explaining the goal o these poetic experiments and their place in the history of human culture.
40
+
41
+ * "[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)
42
+
43
+ * 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)
44
+
45
+ ## BibTeX entry and citation info
46
+
47
+ ```
48
+ @article{orekhov2020neural,
49
+ title={Neural reading: Insights from the analysis of poetry generated by artificial neural networks},
50
+ author={Orekhov, Boris and Fischer, Frank},
51
+ journal={Orbis Litterarum},
52
+ volume={75},
53
+ number={5},
54
+ pages={230--246},
55
+ year={2020},
56
+ publisher={Wiley Online Library}
57
+ }
58
+ ```