Using 2.0.0 train model, but update transformers and sbert, can't load model by: SentenceTransformer(model_path)

[rollmark]

OSError: Error no file named pytorch_model.bin, tf_model.h5, model.ckpt.index or flax_model.msgpack found in directory ............

I need move pytorch_model.bin file from 2_Dense directory to out director, then it work, but a lots of warning like:
ome weights of the model checkpoint at /model_path/ were not used when initializing DistilBertModel: ['linear.bias', 'linear.weight']

• This IS expected if you are initializing DistilBertModel from the checkpoint of a model trained on another task or with another architecture (e.g. initializing a BertForSequenceClassification model from a BertForPreTraining model).
• This IS NOT expected if you are initializing DistilBertModel from the checkpoint of a model that you expect to be exactly identical (initializing a BertForSequenceClassification model from a BertForSequenceClassification model).
  Some weights of DistilBertModel were not initialized from the model checkpoint at /Users/mark.qin/derbysoft-ai-mapping/model_hubs/train_models/training_MultipleNegativesRankingLoss-latest/ and are newly initialized: ['transformer.layer.2.ffn.lin2.weight', 'transformer.layer.2.attention.out_lin.weight', 'transformer.layer.4.output_layer_norm.weight', 'transformer.layer.3.attention.k_lin.weight', 'transformer.layer.3.attention.v_lin.weight', 'transformer.layer.4.attention.out_lin.weight', 'transformer.layer.2.sa_layer_norm.bias', 'transformer.layer.4.sa_layer_norm.bias', 'embeddings.position_embeddings.weight', 'transformer.layer.1.attention.v_lin.bias', 'transformer.layer.1.ffn.lin1.bias', 'embeddings.LayerNorm.bias', 'transformer.layer.1.attention.q_lin.weight', 'transformer.layer.0.attention.q_lin.weight', 'transformer.layer.2.ffn.lin2.bias', 'transformer.layer.4.attention.out_lin.bias', 'transformer.layer.3.sa_layer_norm.weight', 'transformer.layer.5.output_layer_norm.bias', 'transformer.layer.3.sa_layer_norm.bias', 'embeddings.LayerNorm.weight', 'transformer.layer.5.attention.v_lin.weight', 'transformer.layer.5.attention.q_lin.weight', 'embeddings.word_embeddings.weight', 'transformer.layer.5.attention.out_lin.weight', 'transformer.layer.0.attention.k_lin.weight', 'transformer.layer.1.output_layer_norm.weight', 'transformer.layer.1.attention.out_lin.bias', 'transformer.layer.2.output_layer_norm.weight', 'transformer.layer.3.attention.out_lin.weight', 'transformer.layer.4.attention.q_lin.weight', 'transformer.layer.3.attention.k_lin.bias', 'transformer.layer.2.attention.q_lin.weight', 'transformer.layer.3.output_layer_norm.bias', 'transformer.layer.0.sa_layer_norm.weight', 'transformer.layer.2.attention.v_lin.bias', 'transformer.layer.5.attention.k_lin.weight', 'transformer.layer.5.attention.v_lin.bias', 'transformer.layer.3.ffn.lin2.weight', 'transformer.layer.5.attention.q_lin.bias', 'transformer.layer.3.attention.q_lin.weight', 'transformer.layer.5.ffn.lin1.bias', 'transformer.layer.0.attention.v_lin.bias', 'transformer.layer.2.ffn.lin1.bias', 'transformer.layer.4.ffn.lin2.weight', 'transformer.layer.2.output_layer_norm.bias', 'transformer.layer.4.attention.v_lin.bias', 'transformer.layer.2.attention.q_lin.bias', 'transformer.layer.4.ffn.lin2.bias', 'transformer.layer.4.attention.k_lin.weight', 'transformer.layer.1.attention.k_lin.bias', 'transformer.layer.0.ffn.lin1.bias', 'transformer.layer.3.output_layer_norm.weight', 'transformer.layer.3.ffn.lin1.weight', 'transformer.layer.0.attention.out_lin.bias', 'transformer.layer.1.attention.out_lin.weight', 'transformer.layer.4.attention.q_lin.bias', 'transformer.layer.2.attention.k_lin.bias', 'transformer.layer.5.ffn.lin1.weight', 'transformer.layer.5.sa_layer_norm.weight', 'transformer.layer.5.attention.k_lin.bias', 'transformer.layer.5.ffn.lin2.bias', 'transformer.layer.3.attention.v_lin.bias', 'transformer.layer.0.output_layer_norm.weight', 'transformer.layer.1.sa_layer_norm.bias', 'transformer.layer.1.attention.q_lin.bias', 'transformer.layer.0.attention.k_lin.bias', 'transformer.layer.1.ffn.lin2.bias', 'transformer.layer.2.attention.v_lin.weight', 'transformer.layer.1.attention.k_lin.weight', 'transformer.layer.5.attention.out_lin.bias', 'transformer.layer.0.attention.v_lin.weight', 'transformer.layer.5.ffn.lin2.weight', 'transformer.layer.1.attention.v_lin.weight', 'transformer.layer.2.attention.k_lin.weight', 'transformer.layer.2.ffn.lin1.weight', 'transformer.layer.0.ffn.lin2.bias', 'transformer.layer.1.sa_layer_norm.weight', 'transformer.layer.2.attention.out_lin.bias', 'transformer.layer.0.sa_layer_norm.bias', 'transformer.layer.1.output_layer_norm.bias', 'transformer.layer.3.ffn.lin2.bias', 'transformer.layer.4.attention.v_lin.weight', 'transformer.layer.4.sa_layer_norm.weight', 'transformer.layer.0.output_layer_norm.bias', 'transformer.layer.0.attention.q_lin.bias', 'transformer.layer.5.output_layer_norm.weight', 'transformer.layer.3.ffn.lin1.bias', 'transformer.layer.2.sa_layer_norm.weight', 'transformer.layer.4.attention.k_lin.bias', 'transformer.layer.0.ffn.lin2.weight', 'transformer.layer.0.ffn.lin1.weight', 'transformer.layer.4.ffn.lin1.weight', 'transformer.layer.1.ffn.lin2.weight', 'transformer.layer.3.attention.out_lin.bias', 'transformer.layer.3.attention.q_lin.bias', 'transformer.layer.4.ffn.lin1.bias', 'transformer.layer.4.output_layer_norm.bias', 'transformer.layer.5.sa_layer_norm.bias', 'transformer.layer.1.ffn.lin1.weight', 'transformer.layer.0.attention.out_lin.weight']
  You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.

my question: how to fix it that i don't need retrain my model