eval_loss = 1.1393344551324844 f1_score = 0.8489477495153341 precision = 0.8877870053737177 recall = 0.8133643646779752