This article analyzes the Mixture‑of‑Subspaces in Low‑Rank Adaptation (MoSLoRA) paper, explaining its motivation, design choices that replace LoRA's gate with a mixer matrix, connections to multi‑head attention, experimental findings on LLaMA‑3 fine‑tuning, and theoretical proofs of its re‑parameterization properties.
This guide walks you through enabling Alibaba Cloud Function Compute and NAS, claiming free quotas, deploying the Flux text‑to‑image model with a LoRA, and using ComfyUI to create detailed, ultra‑realistic illustrations of classic Journey to the West scenes.
This article provides a comprehensive, step‑by‑step guide to training customized large language models, covering industry‑specific needs, data privacy, meticulous data cleaning, optimal data‑ratio balancing, token budgeting, GPU memory accounting, LoRA fine‑tuning techniques, and practical evaluation metrics for robust AI deployment.
The article introduces a multi‑LoRA strategy that lets many scenario‑specific adapters share a single base LLM, dramatically cutting GPU usage and cost while preserving performance, and explains how to fine‑tune with LoRA, merge adapters, and serve them efficiently using VLLM.
This guide walks you through using Alibaba Cloud PAI and the open‑source LLaMA Factory framework to fine‑tune a Llama‑3 8B model for Chinese dialogue and role‑playing, create a “Zhuge Liang” chatbot, evaluate its performance, and clean up resources.
This guide introduces the fundamentals of AI image generation with Stable Diffusion, covering three main usage methods, the Draw Things desktop app, model types, samplers, prompts, and post‑processing techniques to create high‑quality promotional graphics for events like the 618 sale.
This article outlines a five‑week large‑model training curriculum, detailing weekly topics such as transformer fundamentals, encoder‑decoder architectures, self‑attention, LoRA fine‑tuning, and quantization, along with associated video lectures and PDF slide decks for developers.
EasyAnimate, an open‑source DiT‑based video generation framework from Alibaba Cloud AI Platform PAI, offers a complete pipeline—including data preprocessing, VAE and DiT training, LoRA fine‑tuning, motion‑module integration, and scalable inference up to 768×768 resolution and 144 frames—leveraging Diffusion Transformers to produce longer, higher‑quality videos.
This article systematically analyzes popular parameter‑efficient fine‑tuning (PEFT) techniques for large language models—including Adapter Tuning, Prefix Tuning, LoRA, QLoRA, AdaLoRA, and SoRA—detailing their principles, implementation code, experimental results on NLU tasks, and practical acceleration using the Unsloth library.
This article explains how AI‑generated content (AIGC) technologies such as text‑to‑image diffusion models, Lora fine‑tuning, and IP‑Adapter style transfer are applied to interior design, dramatically reducing design time, cutting costs, and enabling personalized, high‑quality visualizations for both consumers and furniture merchants.
This article explains how to use a large language model to automatically convert long‑form texts into Alpaca‑style question‑answer pairs, build a LangChain processing chain, and then fine‑tune a model such as Phi‑3‑mini‑4k‑instruct with LoRA, providing full Python code examples.
OPPO details how it deploys multimodal pretrained models on resource‑constrained edge devices by compressing CLIP‑based image‑text retrieval, adapting Chinese text‑to‑image generation with LoRA and adapters, and lightweighting diffusion models through layer pruning and progressive distillation, achieving sub‑3‑second generation while preserving cloud‑level quality.
This article provides concise technical explanations for 25 common questions about Transformer models, covering scaled dot‑product attention scaling, separate Q/K projections, feed‑forward network design, attention variants, normalization, LoRA versus full‑parameter training, KV‑cache, pre‑ and post‑norm, computational cost analysis, and advanced position‑encoding techniques.
Leveraging AI image generation and LoRA model fine‑tuning, Baidu’s MEUX team reimagined the traditional 24 solar‑term cards in a fresh New‑Chinese aesthetic, blending cultural heritage, artistic composition, and modern visual effects to boost design efficiency and create a cohesive brand experience.
The guide walks through fine‑tuning InternLM‑Chat‑7B with XTuner, showing how to build a DataLoader from a HuggingFace Dataset, convert a LoRA .pth checkpoint to HuggingFace format, merge the adapter into the base model, run inference, and adapt the process for custom datasets and 4‑bit quantization experiments.
The AI Clothing Graffiti Project in Taobao Life 2 leverages Stable Diffusion, ControlNet, and LoRA to let users generate and stylize clothing designs via text‑image prompts, employing parallel processing, face repair, and content filtering, and has launched successfully, inviting algorithm engineers to join the team.
This article shares hands‑on guidance on using continual pre‑training (CPT), supervised fine‑tuning (SFT), and LoRA adapters for large language models, covering dataset size requirements, learning‑rate scheduling, warm‑up ratios, epoch strategies, and practical routing choices based on real‑world experiments.
This guide explains how AI image generators such as Stable Diffusion and DALL·E 3 turn text prompts into pictures by using diffusion models, CLIP‑aligned embeddings, and optional controls like negative prompts, fine‑tuned LoRA checkpoints and ControlNet conditioning, highlighting their differences, workflow, and practical customization.
This article explains the motivation, dictionary design, bigram table, encoding rules, and real‑world compression results of the Smaz2 library, a space‑optimized C/Python compressor for short LoRa messages on microcontrollers with less than 2 KB RAM.
This article explains the concepts of pretraining and fine‑tuning for large language models, compares full‑parameter, LoRA and QLoRA approaches, discusses resource consumption, introduces the ModelScope SWIFT framework with code examples, and shows how fine‑tuning can improve data‑visualisation tasks while reducing token usage.
This article explains the principle of LoRA (Low‑Rank Adaptation) for large language models, demonstrates how to decompose weight updates into low‑rank matrices, and provides a complete PyTorch implementation that fine‑tunes a small VGG‑19 network on a custom goldfish dataset.
This article walks through a complete methodology for creating brand symbol posters using AI, covering basic and advanced Stable Diffusion techniques such as ControlNet, depth‑map generation, semantic segmentation, LoRA integration, and post‑processing to achieve high‑quality, efficient visual assets.
This guide walks you through preparing the environment, downloading Qwen‑72B‑Chat, performing LoRA fine‑tuning on PAI‑DSW, merging weights, and deploying the model for offline inference, web UI, API, and PAI SDK services on Alibaba Cloud.
This article details how 58.com’s design team leveraged AI tools such as Stable Diffusion, LoRA models, and prompt engineering to create a visually rich September recruitment landing page, outlining the project background, design strategy, technical exploration, and measurable results.
Explore how AI tools like Stable Diffusion and Runway GEN‑2 can accelerate marketing design workflows, from converting 2D sketches to 3D assets, generating dynamic video posters, training custom LoRA models for travel backgrounds, and creating personalized avatar styles, with step‑by‑step case studies and practical tips.
SWIFT is an open‑source, PyTorch‑based framework that integrates multiple efficient fine‑tuning methods such as LoRA, QLoRA, Adapter, and the proprietary ResTuning, enabling developers to fine‑tune large language and multimodal models on consumer‑grade GPUs with significantly reduced memory and compute requirements.
This article introduces parameter‑efficient fine‑tuning (PEFT) techniques—including additive adapters, soft‑prompt methods, selection‑based BitFit, and re‑parameterization approaches like LoRA and AdaLoRA—explains their architectures, experimental results, and provides end‑to‑end code for fine‑tuning ChatGLM2‑6B on a Chinese medical QA dataset.
A leaked Google internal document titled “We have no moat, OpenAI also has none” reveals that both companies are losing the AI arms race to rapidly advancing open‑source models, which achieve comparable performance at a fraction of the cost, prompting a strategic rethink for Google.
This article introduces EasyPhoto, an open‑source SD‑WebUI plugin that lets users upload a handful of personal photos, quickly train a LoRA model, and generate high‑quality AI portraits using Stable Diffusion, ControlNet and face‑ID techniques, with step‑by‑step installation and usage guidance.
This study evaluates virtual clothing try‑on by fine‑tuning LoRA models on a handful of garment images and integrating Stable Diffusion with ControlNet, Inpainting, and Segment Anything, showing that the AI‑generated pipeline matches or exceeds traditional geometric‑matching VTON in visual fidelity, especially for upper‑body apparel, while running locally on an RTX 3060.
This guide compares Stable Diffusion fine‑tuning methods, shows why LoRA offers the best size‑and‑speed trade‑off, and provides a step‑by‑step workflow—from dataset collection and preprocessing to parameter tuning, 20‑minute training on a single GPU, and practical tips for successful custom model generation.
This article explains how designers can shorten 3D design cycles by using AIGC tools to train LoRA adapters for Stable Diffusion, covering the theory, required plugins, data preparation, labeling rules, training steps, result tweaking, and practical applications.
This article introduces the open‑source FaceChain AI portrait generation project, detailing background, demo results, environment setup on ModelScope notebooks, step‑by‑step usage instructions, and an in‑depth explanation of the Stable Diffusion‑based LoRA training and inference pipeline.
This practical guide walks users through installing Stable Diffusion WebUI, explains the differences between base, LoRA, VAE, and ControlNet models, shows how to derive prompts with CLIP or DeepBooru, and provides detailed text‑to‑image and image‑to‑image examples for effective prompt engineering.
This article details the step‑by‑step process of installing GPU drivers, setting up a Python environment, deploying the open‑source Alpaca‑LoRA large language model, fine‑tuning it with Chinese data on a multi‑GPU server, and running inference, while discussing practical challenges and performance observations.
This guide walks you through using Alibaba Cloud's PAI platform to quickly fine‑tune Llama 2 with LoRA or full‑parameter methods, deploy the models as online inference services, and launch an interactive WebUI, covering preparation, data formatting, training jobs, and deployment details.
The article explains how LoRA and its 4‑bit QLoRA extension dramatically reduce trainable parameters and GPU memory for fine‑tuning large language models, while GPTQ post‑training quantization compresses weights for cheap inference, and shows how KubeAI integrates these techniques into a one‑click workflow for 7 B, 13 B, and 33 B models from data upload to API deployment.
This article explains how to use PAI-Blade to accelerate Stable Diffusion inference by optimizing LoRA and ControlNet components, detailing configuration steps, code modifications, benchmark results on A100/A10 GPUs, and integration with both Diffusers and the popular Stable-Diffusion-webui, highlighting performance gains and memory savings.
This article introduces the upgraded PAI‑Diffusion Chinese models, highlighting major improvements in image quality and style diversity, detailing lightweight fine‑tuning methods such as LoRA and Textual Inversion, showcasing controllable editing, scenario‑specific customization, and providing step‑by‑step usage instructions on popular platforms.
An analysis of a leaked Google internal document reveals how open‑source large language models, low‑cost fine‑tuning techniques like LoRA, and rapid community innovation are reshaping the AI competition, challenging the dominance of both Google and OpenAI and prompting a strategic rethink.
This guide explains how AI creates photorealistic images, tracing the shift from VAEs and GANs to diffusion models, detailing latent diffusion, ControlNet conditioning, CLIP text‑image alignment, and lightweight fine‑tuning methods like DreamBooth and LoRA, plus practical tips for higher‑resolution results.
Tencent’s rapid‑deployment project creates a volunteer‑hosted, low‑cost atmospheric monitoring network using LoRa, NB‑IoT and Wi‑Fi, with edge‑agnostic data collection, cloud‑based device normalization, and business processing modules that manage devices, analyze PM2.5 sensor data, and visualize real‑time air quality across five community sites.
The article describes a complete LoRa‑based environmental monitoring and control system for data centers, built on TencentOS Tiny and featuring a LoRa gateway, sensor‑rich node, LCD display, relay and LED modules, cloud integration via Tencent Cloud IoT Explorer, and a WeChat mini‑program for remote management.
The article reviews the latest IoT and edge‑computing trends, outlines Tencent’s multi‑layer IoT platform, edge solutions, LoRa network deployment, and how AI‑driven data analysis will transform the fragmented IoT market into a smarter, more integrated ecosystem.
