Universal Assisted Generation: Faster Decoding with Any Assistant Model
•
50
Nadav Timor
Nadav-Timor
AI & ML interests
None yet
Recent Activity
upvoted
a
paper
about 10 hours ago
MambaByte: Token-free Selective State Space Model
Articles
Organizations
Nadav-Timor's activity
upvoted
a
paper
about 10 hours ago
upvoted
an
article
2 months ago
Article
Faster Assisted Generation with Dynamic Speculation
•
44
upvoted
a
paper
7 months ago
upvoted
a
collection
7 months ago
upvoted
a
paper
7 months ago
authored
a
paper
7 months ago
`llama3p-70b-rc3_vr_mid_3` & `llama3p-7b-rc3_vr_mid_2`?
#2 opened 8 months ago
by
Nadav-Timor
reacted to
julien-c's
post with 👍
10 months ago
Post
What if you could casually access your remote GPU in HF Spaces from the comfort of your local VSCode 🤯
- 8 replies
updated
a
dataset
about 1 year ago
`max_position_embeddings=32768` and `precompute_freqs_cis` with `end=128_000`
1
#6 opened about 1 year ago
by
Nadav-Timor
`max_position_embeddings=32768` with "attention span of 131K tokens"
1
#57 opened about 1 year ago
by
Nadav-Timor
liked
a
dataset
about 1 year ago
liked
a
model
about 1 year ago
liked
a
Space
about 1 year ago
Space isn't working because there is a runtime error
3
#1 opened over 1 year ago
by
Nadav-Timor
liked
a
model
about 1 year ago
Challenges in reproducing the HumanEval scores reported in the paper with BigCode's Eval Harness
6
#2 opened over 1 year ago
by
Nadav-Timor
Padding?
#4 opened over 1 year ago
by
Nadav-Timor
liked
a
dataset
over 1 year ago
updated
a
dataset
over 1 year ago