dhruvdh

POSTS

• Accelerating LLM Inference with Parallel Draft Models (PARD)

  by dhruvdh on 4/11/2025, 6:10 PM with 0 comments

• Open-sourcing Three EXAONE 3.5 Models: 2.4B, 7.8B, 32B

  by dhruvdh on 12/9/2024, 3:55 PM with 4 comments

• The Tyranny of Possibilities in the Design of Task-Oriented LLM Systems

  by dhruvdh on 1/2/2024, 1:05 PM with 3 comments

• MoAI Platform – Scale PyTorch, TensorFlow, etc. to Thousands of GPU/NPUs

  by dhruvdh on 10/27/2023, 1:30 AM with 0 comments

• Lamini LLM Finetuning on AMD ROCm: A Technical Recipe

  by dhruvdh on 10/25/2023, 6:18 PM with 4 comments

• ModuleFormer: Modularity Emerges from Mixture-of-Experts

  by dhruvdh on 9/17/2023, 1:46 AM with 1 comments