Encountering _IncompatibleKeys while training LoRA for flux

I encountered the following output while training LoRA for Flux, should I ignore it?
`INFO Loaded T5xxl: _IncompatibleKeys(missing_keys=['encoder.embed_tokens.weight'], unexpected_keys=[]) `
Among them, T5. safetensors is obtained by merging two safetensors files in the text_decoder_2 directory of the huggingface model.


Start by reproducing Flux LoRA training with the T5.safetensors file assembled from the two files in the text_decoder_2 directory. Inspect the loader output containing the missing encoder.embed_tokens.weight key and compare it with the expected model contents. Done means establishing whether the message is expected or identifying a reproducible loading problem and the affected handling.

Written by the indexing model from the issue text.