Picture for Leqi Zou

Leqi Zou

MegaScale: Scaling Large Language Model Training to More Than 10,000 GPUs

Add code
Feb 23, 2024
Figure 1 for MegaScale: Scaling Large Language Model Training to More Than 10,000 GPUs
Figure 2 for MegaScale: Scaling Large Language Model Training to More Than 10,000 GPUs
Figure 3 for MegaScale: Scaling Large Language Model Training to More Than 10,000 GPUs
Figure 4 for MegaScale: Scaling Large Language Model Training to More Than 10,000 GPUs
Viaarxiv icon

Monolith: Real Time Recommendation System With Collisionless Embedding Table

Add code
Sep 27, 2022
Figure 1 for Monolith: Real Time Recommendation System With Collisionless Embedding Table
Figure 2 for Monolith: Real Time Recommendation System With Collisionless Embedding Table
Figure 3 for Monolith: Real Time Recommendation System With Collisionless Embedding Table
Figure 4 for Monolith: Real Time Recommendation System With Collisionless Embedding Table
Viaarxiv icon

CowClip: Reducing CTR Prediction Model Training Time from 12 hours to 10 minutes on 1 GPU

Add code
Apr 22, 2022
Figure 1 for CowClip: Reducing CTR Prediction Model Training Time from 12 hours to 10 minutes on 1 GPU
Figure 2 for CowClip: Reducing CTR Prediction Model Training Time from 12 hours to 10 minutes on 1 GPU
Figure 3 for CowClip: Reducing CTR Prediction Model Training Time from 12 hours to 10 minutes on 1 GPU
Figure 4 for CowClip: Reducing CTR Prediction Model Training Time from 12 hours to 10 minutes on 1 GPU
Viaarxiv icon