Spaces:
Runtime error
Runtime error
File size: 2,822 Bytes
8f67185 2203509 8f67185 2203509 8f67185 d18e04b 8f67185 1f57ae8 d18e04b 8f67185 1f57ae8 |
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 |
import transformers
import streamlit as st
from transformers import AutoTokenizer, AutoModelWithLMHead
tokenizer = AutoTokenizer.from_pretrained("mrm8488/spanish-gpt2")
model = AutoModelWithLMHead.from_pretrained("mrm8488/spanish-gpt2")
def infer(input_ids, max_length, temperature, top_k, top_p, num_return_sequences):
output_sequences = model.generate(
input_ids=input_ids,
max_length=max_length,
temperature=temperature,
top_k=top_k,
top_p=top_p,
do_sample=True,
num_return_sequences=num_return_sequences,
)
return output_sequences
default_value = "Vea cómo una red neuronal moderna completa automáticamente su texto 🤗 Este sitio, creado por el equipo de Hugging Face, le permite escribir un documento completo directamente desde su navegador, y puede activar el Transformer (Spanish GPT-2) en cualquier lugar usando la tecla Tab. Es como tener una máquina inteligente que completa tus pensamientos 😀 Comienza escribiendo un fragmento personalizado."
#prompts
st.title("Write with Spanish GPT-2 🦄")
st.write("Demo del modelo Spanish GPT-2 creado por Manuel Romero y su equipo en la Flax/Jax Commnunity Event orgranizado por Hugging Face y Google")
sent = st.text_area("Text", default_value, height = 275)
max_length = st.sidebar.slider("Max Length", min_value = 10, max_value=30)
temperature = st.sidebar.slider("Temperature", value = 1.0, min_value = 0.0, max_value=1.0, step=0.05)
top_k = st.sidebar.slider("Top-k", min_value = 0, max_value=5, value = 0)
top_p = st.sidebar.slider("Top-p", min_value = 0.0, max_value=1.0, step = 0.05, value = 0.9)
num_return_sequences = st.sidebar.number_input('Number of Return Sequences', min_value=1, max_value=5, value=1, step=1)
encoded_prompt = tokenizer.encode(sent, add_special_tokens=False, return_tensors="pt")
if encoded_prompt.size()[-1] == 0:
input_ids = None
else:
input_ids = encoded_prompt
output_sequences = infer(input_ids, max_length, temperature, top_k, top_p, num_return_sequences)
for generated_sequence_idx, generated_sequence in enumerate(output_sequences):
print(f"=== GENERATED SEQUENCE {generated_sequence_idx + 1} ===")
generated_sequences = generated_sequence.tolist()
# Decode text
text = tokenizer.decode(generated_sequence, clean_up_tokenization_spaces=True)
# Remove all text after the stop token
#text = text[: text.find(args.stop_token) if args.stop_token else None]
# Add the prompt at the beginning of the sequence. Remove the excess text that was used for pre-processing
total_sequence = (
sent + text[len(tokenizer.decode(encoded_prompt[0], clean_up_tokenization_spaces=True)) :]
)
generated_sequences.append(total_sequence)
print(total_sequence)
st.write(generated_sequences[-1]) |