4 ms·
But we can probably compress the data very well (e.g. "double delta" is always zero), and then decompress only needed parts.
by deepsun 3y ago
But we can probably compress the data very well (e.g. "double delta" is always zero), and then decompress only needed parts.
- vineyardmike 3y agoThis still requires reading through it all in memory at run time. Maybe we can optimize it with a JIT -what if the package called the ChatGPT api to get the the python code that generates the machine code just for the number queried?