use bloom-350m to train reward model in step2

[70557dzqc]

I want to train bloom_350m in chinese dataset, and run run_350m.sh, change the model_name_or_path. But the loss is nan, how should I solve it? Is the argument "num_padding_at_beginning" cause this?

[70557dzqc]

sorry, it's bloom-560m, not bloom_350m

[LuciusMos]

@panxb833 Hi! I met the same problem as you. Do you know how to solve it?

[Ablustrund]

@panxb833 @LuciusMos I met the same problem as you. Do you know how to solve it?

[xlinsz]

working for me, what's your error?

[scarydemon2]

same error

[scarydemon2]

same error

I have fixed this problem. ref:#571

[robotsp]

@scarydemon2 I have the same problem. Do we need to modify the code in reward_model forward_value function?