gregtatum/static-embeddings

1from model2vec import StaticModel2from tokenizers import Tokenizer3import torch4 5model = StaticModel.from_pretrained("minishlab/potion-multilingual-128M")6embeddings = torch.from_numpy(model.embedding)7 8print("Embedding shape:", embeddings.shape)9bytes = embeddings.shape[0] * embeddings.shape[1] * 410 11print("MiB:", bytes / 1024 / 1024)12 13tokenizer: Tokenizer = model.tokenizer14print(tokenizer.to_str())15