import json
import os
import shutil
import requests
import gradio as gr
from huggingface_hub import Repository
from text_generation import Client
from share_btn import community_icon_html, loading_icon_html, share_js, share_btn_css
HF_TOKEN = os.environ.get("HF_TOKEN", None)
API_URL = "https://api-inference.huggingface.co/models/"
model_id_1, model_id_2 = "Phind/Phind-CodeLlama-34B-v2", "WizardLM/WizardCoder-Python-34B-V1.0"
FIM_PREFIX = "<PRE> "
FIM_MIDDLE = " <MID>"
FIM_SUFFIX = " <SUF>"
FIM_INDICATOR = "<FILL_ME>"
EOS_STRING = "</s>"
EOT_STRING = "<EOT>"
theme = gr.themes.Monochrome(
primary_hue="indigo",
secondary_hue="blue",
neutral_hue="slate",
radius_size=gr.themes.sizes.radius_sm,
font=[
gr.themes.GoogleFont("Open Sans"),
"ui-sans-serif",
"system-ui",
"sans-serif",
],
)
def generate(
model_id, prompt, temperature=0.9, max_new_tokens=256, top_p=0.95, repetition_penalty=1.0,
):
client = Client(
f"{API_URL}{model_id}",
headers={"Authorization": f"Bearer {HF_TOKEN}"},
)
temperature = float(temperature)
if temperature < 1e-2:
temperature = 1e-2
top_p = float(top_p)
fim_mode = False
generate_kwargs = dict(
temperature=temperature,
max_new_tokens=max_new_tokens,
top_p=top_p,
repetition_penalty=repetition_penalty,
do_sample=True,
seed=42,
)
if FIM_INDICATOR in prompt:
fim_mode = True
try:
prefix, suffix = prompt.split(FIM_INDICATOR)
except:
raise ValueError(f"Only one {FIM_INDICATOR} allowed in prompt!")
prompt = f"{FIM_PREFIX}{prefix}{FIM_SUFFIX}{suffix}{FIM_MIDDLE}"
stream = client.generate_stream(prompt, **generate_kwargs)
if fim_mode:
output = prefix
else:
output = prompt
previous_token = ""
for response in stream:
if any([end_token in response.token.text for end_token in [EOS_STRING, EOT_STRING]]):
if fim_mode:
output += suffix
yield output
return output
print("output", output)
else:
return output
else:
output += response.token.text
previous_token = response.token.text
yield output
return output
def generate_both(prompt, temperature=0.9, max_new_tokens=256, top_p=0.95, repetition_penalty=1.0):
generator_1, generator_2 = generate(model_id_1, prompt, temperature, max_new_tokens, top_p, repetition_penalty), generate(model_id_2, prompt, temperature, max_new_tokens, top_p, repetition_penalty)
output_1, output_2 = "", ""
output_1_end, output_2_end = False, False
while True:
try:
output_1 = next(generator_1)
except StopIteration:
output_1_end = True
try:
output_2 = next(generator_2)
except StopIteration:
output_2_end = True
if output_1_end and output_2_end:
yield output_1, output_2
return output_1, output_2
yield output_1, output_2
examples = [
"X_train, y_train, X_test, y_test = train_test_split(X, y, test_size=0.1)\n\n# Train a logistic regression model, predict the labels on the test set and compute the accuracy score",
"// Returns every other value in the array as a new array.\nfunction everyOther(arr) {",
"Poor English: She no went to the market. Corrected English:",
"def alternating(list1, list2):\n results = []\n for i in range(min(len(list1), len(list2))):\n results.append(list1[i])\n results.append(list2[i])\n if len(list1) > len(list2):\n <FILL_ME>\n else:\n results.extend(list2[i+1:])\n return results",
"def remove_non_ascii(s: str) -> str:\n \"\"\" <FILL_ME>\nprint(remove_non_ascii('afkdj$$('))",
]
def process_example(args):
for x in generate_both(args):
pass
return x
css = ".generating {visibility: hidden}"
monospace_css = """
#q-input textarea {
font-family: monospace, 'Consolas', Courier, monospace;
}
"""
css += share_btn_css + monospace_css + ".gradio-container {color: black}"
description = f"""
<div style="text-align: center;">
<h1> Phind VS WizardCoder Playground</h1>
</div>
<div style="text-align: left;">
<p>Compare python code generations from <a href="https://hf.co/{model_id_1}">{model_id_1}</a> (73.8% pass@1 on HumanEval) & <a href="https://hf.co/{model_id_2}">{model_id_2}</a> (73.2% pass@1 on HumanEval), which makes them surpass GPT4 (2023/03/15) on the same benchmark</p>
<p>Moreover, you can try those models on VSCode using HF Autocomplete extenson. Read more <a href="https://github.com/huggingface/huggingface-vscode#phind-and-wizardcoder">here</a>.</p>
<p>This space is cloned from <a href="https://hf.co/spaces/codellama/codellama-playground">codellama/codellama-playground</a></p>
</div>
"""
with gr.Blocks(theme=theme, analytics_enabled=False, css=css) as demo:
with gr.Column():
gr.Markdown(description)
with gr.Row():
with gr.Column():
instruction = gr.Textbox(
placeholder="Enter your code here",
lines=5,
label="Input",
elem_id="q-input",
)
submit = gr.Button("Generate", variant="primary")
with gr.Row():
output_1 = gr.Code(elem_id="q-output", lines=30, label=f"{model_id_1} Output", language="python")
output_2 = gr.Code(elem_id="q-output", lines=30, label=f"{model_id_2} Output", language="python")
with gr.Row():
with gr.Column():
with gr.Accordion("Advanced settings", open=False):
with gr.Row():
column_1, column_2 = gr.Column(), gr.Column()
with column_1:
temperature = gr.Slider(
label="Temperature",
value=0.1,
minimum=0.0,
maximum=1.0,
step=0.05,
interactive=True,
info="Higher values produce more diverse outputs",
)
max_new_tokens = gr.Slider(
label="Max new tokens",
value=256,
minimum=0,
maximum=8192,
step=64,
interactive=True,
info="The maximum numbers of new tokens",
)
with column_2:
top_p = gr.Slider(
label="Top-p (nucleus sampling)",
value=0.90,
minimum=0.0,
maximum=1,
step=0.05,
interactive=True,
info="Higher values sample more low-probability tokens",
)
repetition_penalty = gr.Slider(
label="Repetition penalty",
value=1.05,
minimum=1.0,
maximum=2.0,
step=0.05,
interactive=True,
info="Penalize repeated tokens",
)
gr.Examples(
examples=examples,
inputs=[instruction],
cache_examples=False,
fn=process_example,
outputs=[output_1],
)
submit.click(
generate_both,
inputs=[instruction, temperature, max_new_tokens, top_p, repetition_penalty],
outputs=[output_1, output_2],
)
demo.queue(concurrency_count=16).launch(debug=True)