File size: 3,193 Bytes
97037a8
 
 
 
 
 
 
 
 
 
d23d2f6
 
 
 
 
 
 
3ab8618
 
 
 
 
 
97037a8
 
 
 
 
 
 
cfaffad
97037a8
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
cfaffad
97037a8
cfaffad
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
20eaaec
 
 
 
 
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
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
---
language:
- en
tags:
- text generation
- pytorch
- causal-lm
license: apache-2.0
datasets:
- custom
widget:
- text: "style: Pilsner\nbatch_size: 20\nefficiency: 75\nboil_size:"
  example_title: "Pilsener"
- text: "style: IPA\nbatch_size: 20\nefficiency: 75\nboil_size:"
  example_title: "IPA"
- text: "style: Scottish Ale\nbatch_size: 20\nefficiency: 75\nboil_size:"
  example_title: "Scottish Ale"
inference:
  parameters:
    do_sample: true
    top_k: 10
    top_p: 0.99
    max_length: 500
---

# GPT-Neo 125M finetuned with beer recipes

## Model Description

GPT-Neo 125M is a transformer model based on EleutherAI's replication of the GPT-3 architecture https://huggingface.co/EleutherAI/gpt-neo-125M.
It generates recipes for brewing beer in a YAML-like format which can be easily used for different purposes.

## Training data

This model was trained on a custom dataset of ~ 76,800 beer recipes from the internet. It includes recipes for the following 
styles of beer:

* Strong American Ale 
* Pale American Ale
* India Pale Ale (IPA)
* Standard American Beer
* Stout
* English Pale Ale
* IPA
* American Porter and Stout
* Sour Ale
* Irish Beer
* Strong British Ale
* Belgian and French Ale
* German Wheat and Rye Beer
* Czech Lager
* Spice/Herb/Vegetable Beer
* Specialty Beer
* American Ale
* Pilsner
* Belgian Ale
* Strong Belgian Ale
* Bock
* Brown British Beer
* German Wheat Beer
* Fruit Beer
* Amber Malty European Lager
* Pale Malty European Lager
* British Bitter
* Amber and Brown American Beer
* Light Hybrid Beer
* Pale Commonwealth Beer
* American Wild Ale
* European Amber Lager
* Belgian Strong Ale
* International Lager
* Amber Bitter European Lager
* Light Lager
* Scottish and Irish Ale
* European Sour Ale
* Trappist Ale
* Strong European Beer
* Porter
* Historical Beer
* Pale Bitter European Beer
* Amber Hybrid Beer
* Smoke Flavored/Wood-Aged Beer
* Spiced Beer
* Dark European Lager
* Alternative Fermentables Beer
* Mead
* Strong Ale
* Dark British Beer
* Scottish Ale
* Smoked Beer
* English Brown Ale
* Dark Lager
* Cider or Perry
* Wood Beer

### How to use

You can use this model directly with a pipeline for text generation. This example generates a different recipe each time it's run:

```py
>>> from transformers import pipeline
>>> generator = pipeline('text-generation', model='b3ck1/gpt-neo-125M-finetuned-beer-recipes')
>>> generator("style: Pilsner\nbatch_size: 20\nefficiency: 75\nboil_size:", do_sample=True, min_length=50, max_length=500)
>>> print(output[0]['generated_text'])

style: Pilsner
batch_size: 20
efficiency: 70
boil_size: 24
boil_time: 60
fermentables:
- name: Pale Ale
  type: Grain
  amount: 6.5
hops:
- name: Saaz
  alpha: 3.5
  use: Boil
  time: 60
  amount: 0.06
- name: Saaz
  alpha: 3.5
  use: Boil
  time: 30
  amount: 0.06
- name: Saaz
  alpha: 3.5
  use: Boil
  time: 10
  amount: 0.06
- name: Saaz
  alpha: 3.5
  use: Boil
  time: 0
  amount: 0.06
yeasts:
- name: Safale - American Ale Yeast US-05
  amount: 0.11
  min_temperature: 12
  max_temperature: 25
primary_temp: null
mash_steps:
- step_temp: 65
  step_time: 60
miscs: []
```

### See this model in action

This model was used to build https://beerai.net.