-
Notifications
You must be signed in to change notification settings - Fork 1.1k
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
EAGLE docs #4038
base: main
Are you sure you want to change the base?
EAGLE docs #4038
Conversation
@shuaills @zhaochenyang20 could you take a look at this. |
"* EAGLE predicts the next feature vector, i.e. the last hidden layer of the original LLM, using the feature sequence $(f_1, ..., f_k)$ and the token sequence $(t_2, ..., t_{k+1})$. \n", | ||
"* The next token is than sampled from $p_{k+2}=LMHead(f_{k+1})$. We'll than extend the two sequences and continue.\n", | ||
"\n", | ||
"This enhances drafting accuracy by operating on the features instead of tokens for more regular inputs and passing the tokens from the next timestep additionaly to minimize randomness effects from sampling. For more details see [the paper](https://arxiv.org/abs/2401.15077).\n", |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
We use EAGLE 2 not EAGLE https://arxiv.org/abs/2406.16858
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
i will take a look at the difference tomorrow and adjust accordingly. sorry for that mistake.
@simveit Thanks. Will review it later. |
@zhaochenyang20 @zhyncs I added a little bit more text to include the additional improvment through EAGLE2 in my understanding. Please let me know if it's correct. |
Motivation
More details and explanation in EAGLE docs.
Checklist