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.
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:
You can use this model directly with a pipeline for text generation. This example generates a different recipe each time it's run:
>>> 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: []
This model was used to build https://beerai.net.
pytorch_model.bin 551 MBtraining_args.bin 3.1 kB