Text Generation
Transformers
PyTorch
Safetensors
English
mistral
conversational
Inference Endpoints
text-generation-inference

Phi-3 mini is a bhit to dumb for dolphin. -> Phi-3 small ???

#3
by sebastienbo - opened

Phi-3 mini was a bit to small making it dumb, but phi-3 mini is the exact right ammount of data to make it suitabhle for everyb day use without requiring a GPU

Phi-3 medium is too big, -> it requireds a GPU

Could you make a dolphin with phi-3 small ?

image.png

As you can see the medium does'n't add much except for size and complexity, whereas the mini is dumber then gpt 3.5

Sign up or log in to comment