A family of 8 models of varied quantized precision and param counts. As of Jul 14th, the models were trained on 1 epoch of the TinyStories dataset.
-
adrahmana/NanoFable-1M-ternary
Text Generation • 1.38M • Updated • 131 -
adrahmana/NanoFable-6M-ternary
Text Generation • 5.87M • Updated • 129 -
adrahmana/NanoFable-16M-ternary
Text Generation • 15.7M • Updated • 141 -
adrahmana/NanoFable-28M-ternary
Text Generation • 27.8M • Updated • 144