BitDelta: Your Fine-Tune May Only Be Worth One Bit

<a href="https:&#x2F;&#x2F;github.com&#x2F;FasterDecoding&#x2F;BitDelta">https:&#x2F;&#x2F;github.com&#x2F;FasterDecoding&#x2F;BitDelta</a>

&quot;By enabling the use of a single high-precision base model accompanied by multiple 1-bit deltas, BitDelta dramatically reduces GPU memory requirements by more than 10x, which can also be translated to enhanced generation latency in multi-tenant settings.&quot;