gpt2-tokenizer / index.html
julien-c's picture
julien-c HF Staff
Working prototype
7349b7e verified
raw
history blame
1.17 kB
<!DOCTYPE html>
<html>
<head>
<script src="https://cdn.jsdelivr.net/pyodide/v0.19.1/full/pyodide.js"></script>
</head>
<body>
Pyodide test page <br>
Open your browser console to see Pyodide output
<script type="text/javascript">
const URL_VOCAB = "https://huggingface.co/gpt2/resolve/main/vocab.json";
const URL_MERGES = "https://huggingface.co/gpt2/resolve/main/merges.txt";
(async function main() {
const vocab = await (await fetch(URL_VOCAB)).text();
const merges = await (await fetch(URL_MERGES)).text();
const py_code = await (await fetch("./encoder.py")).text();
const c = console;
const pyodide = await loadPyodide({
indexURL : "https://cdn.jsdelivr.net/pyodide/v0.19.1/full/"
});
await pyodide.loadPackagesFromImports(py_code);
pyodide.runPython(py_code);
pyodide.globals.set("vocab", vocab);
pyodide.globals.set("merges", merges);
pyodide.runPython(`encoder = get_encoder_from_strings(vocab, merges)`);
const out = pyodide.runPython(`encoder.encode(${JSON.stringify("Hello my name is")})`);
c.log(Array.from(out));
})();
</script>
</body>
</html>