rush_tabesh Profile Banner
Rush Tabesh Profile
Rush Tabesh

@rush_tabesh

Followers
23
Following
17
Media
0
Statuses
4

Rush (Soroush) Tabesh | Ph.D. Student @ ISTAustria | Efficiency in Deep Learning

Vienna, Austria
Joined August 2023
Don't wanna be here? Send us removal request.
@rush_tabesh
Rush Tabesh
4 months
RT @DAlistarh: Our QuEST paper was selected for Oral Presentation at ICLR @sparseLLMs workshop! .QuEST is the first algorithm with Pareto-o….
Tweet card summary image
github.com
Work in progress. Contribute to IST-DASLab/QuEST development by creating an account on GitHub.
0
9
0
@rush_tabesh
Rush Tabesh
5 months
RT @egor_zverev_ai: (1/n) In our #ICLR2025 paper, we explore a fundamental issue that enables prompt injections: π‹π‹πŒπ¬β€™ π’π§πšπ›π’π₯𝐒𝐭𝐲 𝐭𝐨 π¬πžπ©πšπ«πšβ€¦.
0
13
0
@rush_tabesh
Rush Tabesh
5 months
Happy to introduce #HALO lower-precision fine-tuning for LLMs. With proper Hadamard transforms, #HALO enables accurate INT8/FP6 fine-tuningβ€”lossless speedups up to 1.41Γ—. πŸ“„ Paper: πŸ’» Code: #LLM #Quantization.
@AshkboosSaleh
Saleh Ashkboos
6 months
Happy to release #HALO, a Hadamard-Assisted Lower-Precision scheme that enables INT8/FP6 full #finetuning (FFT) of LLMs. with @mmnnn76 @rush_tabesh @RobertoL_Castro @thoefler @DAlistarh. Paper: Code: 1/4
Tweet media one
0
3
4
@rush_tabesh
Rush Tabesh
1 year
RT @DAlistarh: The code for RoSA: Accurate Parameter-Efficient Fine-Tuning via Sparse + Low-Rank Adapters is now available: ..
0
5
0