app.py

from pathlib import Path

import streamlit as st
import streamlit.components.v1 as components
import torch
import torch.nn.functional as F
from transformers import AutoModelForCausalLM, AutoTokenizer, BatchEncoding

root_dir = Path(__file__).resolve().parent
_highlighted_text_fn = components.declare_component(
    "highlighted_text", path=root_dir / "highlighted_text" / "build"
)

def highlighted_text(tokens, scores, key=None):
    return _highlighted_text_fn(tokens=tokens, scores=scores, key=key, default=0)

def get_windows_batched(examples: BatchEncoding, window_len: int, stride: int = 1, pad_id: int = 0) -> BatchEncoding:
    return BatchEncoding({
        k: [
            t[i][j : j + window_len] + [
                pad_id if k == "input_ids" else 0
            ] * (j + window_len - len(t[i]))
            for i in range(len(examples["input_ids"]))
            for j in range(0, len(examples["input_ids"][i]) - 1, stride)
        ]
        for k, t in examples.items()
    })

BAD_CHAR = chr(0xfffd)

def ids_to_readable_tokens(tokenizer, ids, strip_whitespace=False):
    cur_ids = []
    result = []
    for idx in ids:
        cur_ids.append(idx)
        decoded = tokenizer.decode(cur_ids)
        if BAD_CHAR not in decoded:
            if strip_whitespace:
                decoded = decoded.strip()
            result.append(decoded)
            del cur_ids[:]
        else:
            result.append("")
    return result

model_name = st.selectbox("Model", ["distilgpt2"])

tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name)

window_len = st.select_slider("Window size", options=[8, 16, 32, 64, 128, 256, 512, 1024], value=512)
text = st.text_area("Input text", "The complex houses married and single soldiers and their families.")

inputs = tokenizer([text])
[input_ids] = inputs["input_ids"]
window_len = min(window_len, len(input_ids))
tokens = ids_to_readable_tokens(tokenizer, input_ids)

inputs_sliding = get_windows_batched(
    inputs,
    window_len=window_len,
    pad_id=tokenizer.eos_token_id
)
with torch.inference_mode():
    logits = model(**inputs_sliding.convert_to_tensors("pt")).logits.to(torch.float16)
    logits = F.pad(logits, (0, 0, 0, window_len - 1, 0, 0), value=torch.nan)
    logits = logits.view(-1, logits.shape[-1])[:(window_len - 1) * (len(input_ids) + window_len - 2)]
    logits = logits.view(window_len - 1, len(input_ids) + window_len - 2, logits.shape[-1])

scores = logits.to(torch.float32).softmax(dim=-1)
scores = scores[:, torch.arange(len(input_ids[1:])), input_ids[1:]]
scores = scores.diff(dim=0).transpose(0, 1)
scores = scores.nan_to_num()
scores /= scores.abs().max(dim=1, keepdim=True).values + 1e-9
scores = scores.to(torch.float16)
print(scores)
st.markdown("---")
highlighted_text(tokens=tokens, scores=scores.tolist())