|
|
|
""" |
|
Created on Wed Mar 29 16:01:44 2023 |
|
|
|
Source: https://huggingface.co/EleutherAI/gpt-neo-2.7B |
|
|
|
GPT-Neo 2.7B - a transformer model designed using EleutherAI's replication of the |
|
GPT-3 architecture. The model is available on HuggingFace. Although it can be used |
|
for different tasks, the model is best at what it was pretrained for, which is |
|
generating texts from a prompt. |
|
|
|
The task in this script is text generation. |
|
|
|
There are also a 1.3B and 6B versions. |
|
|
|
""" |
|
|
|
import torch |
|
from transformers import AutoTokenizer, GPTNeoForCausalLM |
|
|
|
tokenizer = AutoTokenizer.from_pretrained("EleutherAI/gpt-neo-2.7B") |
|
model = GPTNeoForCausalLM.from_pretrained("EleutherAI/gpt-neo-2.7B") |
|
|
|
inputs = tokenizer("Hello, my dog is cute", return_tensors="pt") |
|
outputs = model(**inputs, labels=inputs["input_ids"]) |
|
|
|
input_ids = inputs["input_ids"] |
|
|
|
gen_tokens = model.generate( |
|
input_ids, |
|
do_sample=True, |
|
temperature=0.9, |
|
max_length=100, |
|
) |
|
|
|
gen_text = tokenizer.batch_decode(gen_tokens)[0] |
|
|
|
print("=========================================================") |
|
print(gen_text) |