Yifeng Ding

code generation (3) instruction tuning (2) self-supervised learning (1) knowledge transfer (1) knowledge distillation (1) model merging (1) parameter-efficient fine-tuning (1) parameter efficient fine-tuning (1) next-token prediction (1) mixture of expert (1) pretrained model (1) cross-modal fusion (1) code model (1) planning capability (1) code reasoning (1) large language model (1) sparse upcycling (1) upcycled expert (1) code infilling (1) horizon-length prediction (1)