1 views collected around this technical thread.
This article explains how to dramatically speed up PyTorch model training using code optimizations, mixed‑precision, torch.compile, distributed data parallelism, and DeepSpeed, presenting benchmark results that show up to 11.5× acceleration on multiple GPUs while maintaining high accuracy.
This article explains how to train large language models efficiently using LLaMA‑Factory, covering low‑resource training challenges, memory‑saving optimizations for parameters, gradients and activations, framework features, quick‑start guidance, performance tuning, real‑world case studies, and a detailed Q&A.
This article discusses the historical background, major challenges such as massive compute and memory demands, and the technical ecosystem—including data parallelism, pipeline parallelism, and optimization strategies like DeepSpeed and 1F1B—to enable efficient distributed training of large language models.
This article explains the concept, motivations, and step‑by‑step workflow for fine‑tuning large language models—specifically Qwen‑14B—covering data preparation, training commands with DeepSpeed, hyper‑parameter settings, evaluation, and deployment via FastChat, all illustrated with code snippets and configuration details.
This article provides a comprehensive tutorial on fine‑tuning the Qwen‑14B large language model, covering the motivation, fine‑tuning concepts, step‑by‑step workflow, required code, DeepSpeed training parameters, testing scripts, and deployment using FastChat and the 360AI platform.
This article reviews the rapid development of large language models since 2019, outlines the historical background, identifies key challenges such as massive compute demand, memory constraints, and system complexity, and then details distributed training technologies—including data parallelism, pipeline parallelism, and advanced optimization strategies—while also discussing future research directions and answering common questions.
This tutorial walks through fine‑tuning Shanghai AI Lab’s open‑source InternLM models with XTuner, explaining chat‑format conventions, loading and inference (including multimodal InternLM‑XComposer), dataset preparation, configuration sections, DeepSpeed acceleration, and memory‑efficient QLoRA details for 7‑B‑parameter chat models.
Megatron-LLaMA is an open‑source high‑performance training framework for LLaMA models, offering tensor, pipeline, and sequence parallelism, an overlapped optimizer, and near‑linear scalability, achieving up to 176% speedup on 32 GPUs and robust performance even with limited network bandwidth.
DeepSpeed Chat provides a fast, affordable, and scalable system for end‑to‑end RLHF training of ChatGPT‑style large language models, offering one‑click scripts, detailed performance benchmarks across GPU configurations, support for many model families, and a flexible API for custom RLHF pipelines.
This article presents a detailed benchmark of four Transformer models of varying sizes trained on the high‑end Inspur NF5488A5 GPU server, compares its NVSwitch‑based interconnect with a PCIe‑based system, and analyzes the impact of model scale, tensor parallelism, and hardware bandwidth on training efficiency.
This article reviews DeepSpeed‑MoE, an end‑to‑end system that introduces new MoE architectures, model‑compression techniques, and highly optimized inference pipelines, detailing its motivation, design of PR‑MoE (Pyramid‑MoE and Residual‑MoE), distributed parallel strategies, communication and kernel optimizations, and performance gains over dense baselines.