#this is version two with flagging features
import gradio as gr
from transformers import pipeline, AutoTokenizer, AutoModelForCausalLM
# initialize the environment
model_name = 'anugrahap/gpt2-indo-textgen'
HF_TOKEN = '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'
hf_writer = gr.HuggingFaceDatasetSaver(HF_TOKEN, "gpt2-output")
# define the tokenization method
tokenizer = AutoTokenizer.from_pretrained(model_name,
model_max_length=1e30,
padding_side='right',
return_tensors='pt')
# add the EOS token as PAD token to avoid warnings
model = AutoModelForCausalLM.from_pretrained(model_name, pad_token_id=tokenizer.eos_token_id)
generator = pipeline('text-generation', model=model, tokenizer=tokenizer)
# create the decoder parameter to generate the text
def single_generation(text,min_length,max_length,temperature,top_k,top_p,num_beams,repetition_penalty,do_sample):
# create local variable for error parameter
error_rep=gr.Error(f"ERROR: repetition penalty cannot be lower than one! Given rep penalty = {repetition_penalty}")
error_temp=gr.Error(f"ERROR: temperature cannot be zero or lower! Given temperature = {temperature}")
error_minmax=gr.Error(f"ERROR: min length must be lower than or equal to max length! Given min length = {min_length}")
error_numbeams_type=gr.Error(f"ERROR: number of beams must be an integer not {type(num_beams)}")
error_topk_type=gr.Error(f"ERROR: top k must be an integer not {type(top_k)}")
error_minmax_type=gr.Error(f"ERROR: min length and max length must be an integer not {type(min_length)} and {type(max_length)}")
error_empty_temprep=gr.Error("ERROR: temperature and repetition penalty cannot be empty!")
error_empty_text=gr.Error("ERROR: Input Text cannot be empty!")
error_unknown=gr.Error("Unknown Error.")
if text != '':
if type(min_length) == int and type(max_length) == int:
if type(top_k) == int:
if type(num_beams) == int:
if min_length <= max_length:
if temperature > 0:
if repetition_penalty >= 1:
if temperature and repetition_penalty is not None:
result = generator(text,
min_length=min_length,
max_length=max_length,
temperature=temperature,
top_k=top_k,
top_p=top_p,
num_beams=num_beams,
repetition_penalty=repetition_penalty,
do_sample=do_sample,
no_repeat_ngram_size=2,
num_return_sequences=1)
return result[0]["generated_text"]
elif temperature or repetition_penalty is None:
raise error_empty_temprep
elif repetition_penalty < 1:
raise error_rep
elif temperature <= 0:
raise error_temp
elif min_length > max_length:
raise error_minmax
elif type(num_beams) != int:
raise error_numbeams_type
elif type(top_k) != int:
raise error_topk_type
elif type(min_length) != int or type(max_length) != int:
raise error_minmax_type
elif text == '':
raise error_empty_text
else:
raise error_unknown
# create the variable needed for the gradio app
forinput=[gr.Textbox(lines=5, label="Input Text"),
gr.Slider(label="Min Length", minimum=10, maximum=50, value=10, step=5),
gr.Slider(label="Max Length", minimum=10, maximum=100, value=30, step=10),
gr.Number(label="Temperature Sampling", value=1.5),
gr.Slider(label="Top K Sampling", minimum=0, maximum=100, value=30, step=5),
gr.Slider(label="Top P Sampling", minimum=0.01, maximum=1, value=0.93),
gr.Slider(label="Number of Beams", minimum=1, maximum=10, value=5, step=1),
gr.Number(label="Repetition Penalty", value=2.0),
gr.Dropdown(label="Do Sample?", choices=[True,False], value=True, multiselect=False)]
foroutput=gr.Textbox(lines=5, max_lines=50, label="Generated Text with Greedy/Beam Search Decoding")
examples = [
["Indonesia adalah negara kepulauan", 10, 30, 1.0, 25, 0.92, 5, 2.0, True],
["Indonesia adalah negara kepulauan", 10, 30, 1.0, 25, 0.92, 5, 1.0, False],
["Skripsi merupakan tugas akhir mahasiswa", 20, 40, 1.0, 50, 0.92, 1, 2.0, True],
["Skripsi merupakan tugas akhir mahasiswa", 20, 40, 1.0, 50, 0.92, 1, 1.0, False],
["Pemandangan di pantai kuta Bali sangatlah indah.", 30, 50, 0.5, 40, 0.98, 10, 1.0, True],
["Pemandangan di pantai kuta Bali sangatlah indah.", 10, 30, 1.5, 30, 0.93, 5, 2.0, True]]
title = """
This project is a part of thesis requirement of Anugrah Akbar Praramadhan
" article = """
Link to the Trained Model |
Link to the Project Repository |
Link to the Autosaved Generated Output |
Original Paper
Trained on Indo4B Benchmark Dataset of Indonesian language Wikipedia with a Causal Language Modeling (CLM) objective
Copyright Anugrah Akbar Praramadhan 2023
""" # using gradio interfaces app = gr.Interface( fn=single_generation, inputs=forinput, outputs=foroutput, examples=examples, title=title, description=description, article=article, allow_flagging='manual', flagging_options=['Well Performed', 'Inappropriate Word Selection', 'Wordy', 'Strange Word', 'Others'], flagging_callback=hf_writer) if __name__=='__main__': app.launch()