I agree that this site is using cookies. You can find further informations here.
X
Awan, Ammar Ahmad
15  results:
OpenAccess-fulltext X
?
1

DeepSpeed-FastGen: High-throughput Text Generation for LLMs..:

Holmes, Connor ; Tanaka, Masahiro ; Wyatt, Michael...
http://arxiv.org/abs/2401.08671.  , 2024
Link: http://arxiv.org/abs/240..
 
?
2

MCR-DL: Mix-and-Match Communication Runtime for Deep Learni..:

Anthony, Quentin ; Awan, Ammar Ahmad ; Rasley, Jeff...
http://arxiv.org/abs/2303.08374.  , 2023
Link: http://arxiv.org/abs/230..
 
?
3

DeepSpeed4Science Initiative: Enabling Large-Scale Scientif..:

Song, Shuaiwen Leon ; Kruft, Bonnie ; Zhang, Minjia...
http://arxiv.org/abs/2310.04610.  , 2023
Link: http://arxiv.org/abs/231..
 
?
4

A Hybrid Tensor-Expert-Data Parallelism Approach to Optimiz..:

Singh, Siddharth ; Ruwase, Olatunji ; Awan, Ammar Ahmad...
http://arxiv.org/abs/2303.06318.  , 2023
Link: http://arxiv.org/abs/230..
 
?
5

DeepSpeed-VisualChat: Multi-Round Multi-Image Interleave Ch..:

Yao, Zhewei ; Wu, Xiaoxia ; Li, Conglong...
http://arxiv.org/abs/2309.14327.  , 2023
Link: http://arxiv.org/abs/230..
 
?
6

DeepSpeed-Chat: Easy, Fast and Affordable RLHF Training of ..:

Yao, Zhewei ; Aminabadi, Reza Yazdani ; Ruwase, Olatunji...
http://arxiv.org/abs/2308.01320.  , 2023
Link: http://arxiv.org/abs/230..
 
?
7

DeepSpeed Inference: Enabling Efficient Inference of Transf..:

Aminabadi, Reza Yazdani ; Rajbhandari, Samyam ; Zhang, Minjia...
http://arxiv.org/abs/2207.00032.  , 2022
Link: http://arxiv.org/abs/220..
 
?
8

DeepSpeed-MoE: Advancing Mixture-of-Experts Inference and T..:

Rajbhandari, Samyam ; Li, Conglong ; Yao, Zhewei...
http://arxiv.org/abs/2201.05596.  , 2022
Link: http://arxiv.org/abs/220..
 
?
9

1-bit Adam: Communication Efficient Large-Scale Training wi..:

Tang, Hanlin ; Gan, Shaoduo ; Awan, Ammar Ahmad...
http://arxiv.org/abs/2102.02888.  , 2021
Link: http://arxiv.org/abs/210..
 
?
10

1-bit LAMB: Communication Efficient Large-Scale Large-Batch..:

Li, Conglong ; Awan, Ammar Ahmad ; Tang, Hanlin..
http://arxiv.org/abs/2104.06069.  , 2021
Link: http://arxiv.org/abs/210..
 
?
11

Scalable and Efficient MoE Training for Multitask Multiling..:

Kim, Young Jin ; Awan, Ammar Ahmad ; Muzio, Alexandre...
http://arxiv.org/abs/2109.10465.  , 2021
Link: http://arxiv.org/abs/210..
 
?
12

HyPar-Flow: Exploiting MPI and Keras for Scalable Hybrid-Pa..:

Awan, Ammar Ahmad ; Jain, Arpan ; Anthony, Quentin..
http://arxiv.org/abs/1911.05146.  , 2019
Link: http://arxiv.org/abs/191..
 
?
13

Scalable Distributed DNN Training using TensorFlow and CUDA..:

Awan, Ammar Ahmad ; Bedorf, Jeroen ; Chu, Ching-Hsiang..
http://arxiv.org/abs/1810.11112.  , 2018
Link: http://arxiv.org/abs/181..
 
?
14

Optimized Broadcast for Deep Learning Workloads on Dense-GP..:

Awan, Ammar Ahmad ; Chu, Ching-Hsiang ; Subramoni, Hari.
http://arxiv.org/abs/1707.09414.  , 2017
Link: http://arxiv.org/abs/170..
 
?
15

1-bit LAMB: Communication Efficient Large-Scale Large-Batch..:

Conglong Li ; Ammar Ahmad Awan ; Hanlin Tang..
doi:10.5281/zenodo.6629249.  , 2022
Link: https://zenodo.org/recor..
 
1-15