File size: 1,864 Bytes
5028de7
 
 
 
 
 
 
 
 
d42693d
106dfc2
5028de7
 
d42693d
5028de7
 
64981e3
5028de7
 
64981e3
7d4764c
64981e3
5028de7
 
 
fe4c8cf
 
5028de7
64981e3
090d30d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
64981e3
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
---
license: apache-2.0
datasets:
- mrm8488/CHISTES_spanish_jokes
language:
- es
pipeline_tag: text-generation
---

# Adapter for BERTIN-GPT-J-6B fine-tuned on Jokes for jokes generation


## Adapter Description
This adapter was created by using the [PEFT](https://github.com/huggingface/peft) library and allows the base model **BERTIN-GPT-J-6B** to be fine-tuned on the dataset **mrm8488/CHISTES_spanish_jokes** for **Spanish jokes generation** by using the method **LoRA**.

## Model Description
[BERTIN-GPT-J-6B](https://huggingface.co/bertin-project/bertin-gpt-j-6B) is a Spanish finetuned version of GPT-J 6B, a transformer model trained using Ben Wang's Mesh Transformer JAX. "GPT-J" refers to the class of model, while "6B" represents the number of trainable parameters.

## Training data
Dataset from [Workshop for NLP introduction with Spanish jokes](https://github.com/liopic/chistes-nlp)

[More Information needed](https://github.com/huggingface/datasets/blob/main/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)

### Training procedure

TBA

## How to use
```py
import torch
from peft import PeftModel, PeftConfig
from transformers import AutoModelForCausalLM, AutoTokenizer

peft_model_id = "mrm8488/bertin-gpt-j-6B-es-finetuned-chistes_spanish_jokes-500"
config = PeftConfig.from_pretrained(peft_model_id)
model = AutoModelForCausalLM.from_pretrained(config.base_model_name_or_path, return_dict=True, load_in_8bit=True, device_map='auto')
tokenizer = AutoTokenizer.from_pretrained(config.base_model_name_or_path)

# Load the Lora model
model = PeftModel.from_pretrained(model, peft_model_id)

# Inference
batch = tokenizer("Esto son dos amigos", return_tensors='pt')

with torch.cuda.amp.autocast():
  output_tokens = model.generate(**batch, max_new_tokens=50)

print('\n\n', tokenizer.decode(output_tokens[0], skip_special_tokens=True))
```