Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

The output of forward and step are different #500

Open
yjdy opened this issue Jul 29, 2024 · 3 comments
Open

The output of forward and step are different #500

yjdy opened this issue Jul 29, 2024 · 3 comments

Comments

@yjdy
Copy link

yjdy commented Jul 29, 2024

I test forward and step function with same input_ids and intermediate states.
But the output logits and states are quite different.

The code is like this:

for i in range(seqlen):
    hidden_state1,conv_state,ssm_state = model.step(input_ids[i],conv_state, ssm_state)

hidden_state2 = model(input_ids,inference_param)

I think the result should be same, but I do not know the reason.

@WorldEditors
Copy link

The same problem is found, is this problem solved?
I found there might be a bug related to the returned ssm_state (the conv_state is correct according to my investigation)

@chaserd
Copy link

chaserd commented Dec 14, 2024

i guess ssm_state may be different?

@WorldEditors
Copy link

#641

mark the relevant issue, still not solved

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants