⚡️ Speed up function gen_encoder_output_proposals by 14% in PR #1250 (feature/inference-v1-models)

⚡️ Speed up function gen_encoder_output_proposals by 14% in PR #1250 (feature/inference-v1-models) #171