Open HaiFengZeng opened 11 months ago
instead of running in a teacher-forcing in training stage, can we run the model in free running style?:
for i in range(T): X_{t}=RNN(X_{t-1},hidden_state)
since teacher-forcing may have an exposure bias problem.
yes you can
see https://github.com/RWKV/RWKV-infctx-trainer
instead of running in a teacher-forcing in training stage, can we run the model in free running style?:
since teacher-forcing may have an exposure bias problem.