mirror of
https://github.com/onestardao/WFGY.git
synced 2026-04-28 19:50:17 +00:00
24 lines
1 KiB
Python
24 lines
1 KiB
Python
# Evaluation disclaimer (Gradio demo):
|
|
# Scores and labels shown in this app come from specific prompts, models and datasets.
|
|
# They are debugging aids, not scientific proof or product level guarantees.
|
|
|
|
import wfgy_sdk as w, numpy as np, gradio as gr
|
|
from wfgy_sdk.evaluator import compare_logits, pretty_print
|
|
|
|
def run_wfgy(prompt):
|
|
logits = w.call_remote_model(prompt, model_id="gpt2")
|
|
G = np.random.randn(128); G /= np.linalg.norm(G)
|
|
I = G + np.random.normal(scale=0.05, size=128)
|
|
out = w.get_engine().run(input_vec=I, ground_vec=G, logits=logits)
|
|
m = compare_logits(logits, out)
|
|
delta = (1 - m["std_ratio"]) * 100
|
|
return f"variance drop {delta:.0f}% • KL {m['kl_divergence']:.2f}"
|
|
|
|
demo = gr.Interface(fn=run_wfgy,
|
|
inputs=gr.Textbox(),
|
|
outputs="textbox",
|
|
title="WFGY Quick Test",
|
|
description="Type any prompt. GPT-2 baseline, variance/KL will appear.")
|
|
|
|
if __name__ == "__main__":
|
|
demo.launch(server_name="0.0.0.0", server_port=7860)
|