app.py

import gradio as gr
from transformers import pipeline
import numpy as np
import pandas as pd
import re
import torch

number_re = re.compile(r"\.[0-9]*\.")

STATE_DICT = {}
DATA = pd.DataFrame()


def scatter_plot_fn(group_name):
    global DATA
    df = DATA[DATA.group_name == group_name]
    return gr.LinePlot.update(
        value=df,
        x="rank",
        y="val",
        color="layer",
        tooltip=["val", "rank", "layer"],
        caption="",
    )


def find_choices(state_dict):
    global DATA
    layered_tensors = [
        k for k, v in state_dict.items() if number_re.findall(k) and len(v.shape) == 2
    ]
    choices = set()
    data = []
    for name in layered_tensors:
        group_name = number_re.sub(".{N}.", name)
        choices.add(group_name)
        layer = int(number_re.search(name).group()[1:-1])

        svdvals = torch.linalg.svdvals(state_dict[name])
        svdvals /= svdvals.sum()
        for rank, val in enumerate(svdvals.tolist()[:20]):
            data.append((name, layer, group_name, rank, val))
    data = np.array(data)
    DATA = pd.DataFrame(data, columns=["name", "layer", "group_name", "rank", "val"])
    DATA["val"] = DATA["val"].astype("float")
    DATA["layer"] = DATA["layer"].astype("category")
    DATA["rank"] = DATA["rank"].astype("int32")
    return choices


def weights_fn(model_id):
    global STATE_DICT
    try:
        pipe = pipeline(model=model_id)
        STATE_DICT = pipe.model.state_dict()
    except Exception as e:
        print(e)
        STATE_DICT = {}
    choices = find_choices(STATE_DICT)
    return gr.Dropdown.update(choices=choices)


with gr.Blocks() as scatter_plot:
    with gr.Row():
        with gr.Column():
            model_id = gr.Textbox(value="gpt")
            weights = gr.Dropdown(choices=["qkv", "c_fc"])
        with gr.Column():
            plot = gr.LinePlot(show_label=False).style(container=True)
    model_id.change(weights_fn, inputs=model_id, outputs=weights)
    weights.change(fn=scatter_plot_fn, inputs=weights, outputs=plot)

if __name__ == "__main__":
    scatter_plot.launch()