This contribution reports the crosscheck results of JVET-AK0078, EE1-2.2 [1]. JVET-AK0078 integrates a proposed Transformer-based reference frame synthesis for VVC inter coding enhancement named TRFS from JVET-AJ0148 [2] in the latest NNVC-11.0 anchor. The TRFS model is also re-trained with a new training set generated on the BVI-DVC dataset, as opposed to Vimeo-90K dataset in the previous contribution. Lastly, there is some complexity reduction applied to the model compared to JVET-AJ0148 [2]. Almost identical results are reported for the inference crosscheck of the trained TRFS model over NNVC-11.0 for the RA configuration of the proponents.