open-r1/OpenR1-Qwen-7B · About initial Model

I found that the configuration of the openr1-qwen 7b model is somewhat different from the qwen math 7b instruct model (official HF repo), including aspects mentioned in the readme, such as content length and special_tokens_map (although the final results of tokens seem to be the same). I would like to know if it’s possible to change the model-000xxx.safetensors and model.safetensors.index.json from openr1-qwen-7b to the qwen math 7b instruct model (official HF repo) to use it as the initial model for this project.