20241204 ○

too sleepy :cry:

Is it possible to mix up embedding models?

No, it isn’t. Different embedding models generate embeddings that are not compatible with each other. Each embedding model has its onwn way of encoding semantics, and a vector from one model might have no meaning relationship with a vector from another model. Also, embeddings from different models may have varying dimensions.


159.0 lb

Yogurt 20 g Fruits 0 g Doughnut 5 g Salad 10 g Japanese Matcha Soy 10 g Protein chips 20 g Fujitas 35 g Protein shake 50

total protein -> 150 g (>= 140)


MUST:


index 20241203 20241205