{"id":14018,"date":"2026-11-17T11:02:51","date_gmt":"2026-11-17T02:02:51","guid":{"rendered":"https:\/\/localmodelwatch.tsuchitsuchi.com\/?p=14018"},"modified":"2026-11-17T16:12:06","modified_gmt":"2026-11-17T07:12:06","slug":"stable-diffusion-3-5-medium-amd-npu","status":"publish","type":"post","link":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/11\/17\/stable-diffusion-3-5-medium-amd-npu\/","title":{"rendered":"stable-diffusion-3.5-medium-amdnpu-medusa: File List"},"content":{"rendered":"<p><!-- lmw:facts --><\/p>\n<h2>At a Glance<\/h2>\n<div class=\"lmw-table-scroll\" tabindex=\"0\" style=\"overflow-x:auto;-webkit-overflow-scrolling:touch;max-width:100%;\">\n<table style=\"width:max-content;min-width:100%;border-collapse:collapse;\">\n<thead>\n<tr>\n<th>Item<\/th>\n<th>Value<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>Repository<\/td>\n<td><a href=\"https:\/\/huggingface.co\/stabilityai\/stable-diffusion-3.5-medium-amdnpu-medusa\">stabilityai\/stable-diffusion-3.5-medium-amdnpu-medusa<\/a><\/td>\n<\/tr>\n<tr>\n<td>Publisher guide<\/td>\n<td><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/publisher-stability-en\/\">Stability AI: models and licenses<\/a><\/td>\n<\/tr>\n<tr>\n<td>Published<\/td>\n<td>2026-10-27<\/td>\n<\/tr>\n<tr>\n<td>License<\/td>\n<td>stabilityai-ai-community<\/td>\n<\/tr>\n<tr>\n<td>Formats<\/td>\n<td><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/format-onnx-en\/\">ONNX<\/a><\/td>\n<\/tr>\n<tr>\n<td>Source type<\/td>\n<td>Primary source (the publisher itself)<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<p><em>Values determined by this site&#8217;s code when the information was collected. Dates are JST.<\/em><\/p>\n<p><!-- \/lmw:facts --><\/p>\n<h2>Overview<\/h2>\n<p>Stable Diffusion 3.5 Medium, optimized for operation on AMD NPUs, has been released. This is a text-to-image generation model employing the Multimodal Diffusion Transformer (MMDiT) architecture, which generates images based on text prompt inputs.<\/p>\n<h2>Specifications<\/h2>\n<p>This model is specifically optimized for operation on AMD NPU accelerators. The main specifications and configuration are as follows:<\/p>\n<ul>\n<li>Architecture: Multimodal Diffusion Transformer (MMDiT)<\/li>\n<li>Model Type: Diffusion-based text-to-image generation model<\/li>\n<li>Optimization Target: AMD NPU<\/li>\n<li>Distribution Format: ONNX<\/li>\n<li>License: Stability AI Community<\/li>\n<li>Variations: This repository includes the &#8220;M1.0-optimized NPU model&#8221; derived from <code>amd\/stable-diffusion-3.5-medium-amdnpu<\/code>.<\/li>\n<\/ul>\n<p>The repository structure is organized into a <code>common<\/code> folder for shared components and a <code>normal<\/code> folder for text-to-image conversion.<\/p>\n<ul>\n<li>Transformer: Located in the <code>normal\/transformer<\/code> folder, it is exported and structured to utilize AMD NPU accelerators during the denoising steps.<\/li>\n<li>VAE decoder: Located in the <code>common\/vae_decoder<\/code> folder, it is structured to use NPU accelerators for efficient image reconstruction.<\/li>\n<li>Text encoders: Includes <code>common\/text_encoder<\/code>, <code>text_encoder_2<\/code>, <code>text_encoder_3<\/code>, and <code>text_encoder_3_int4<\/code>. These are CLIP and T5 text encoders (including an INT4 version of T5) used to convert input prompts into conditioning embeddings.<\/li>\n<li>Tokenizers: Includes <code>common\/tokenizer<\/code>, <code>tokenizer_2<\/code>, and <code>tokenizer_3<\/code>, containing configurations and vocabulary files for preprocessing prompts for each text encoder.<\/li>\n<li>Scheduler: The <code>common\/scheduler<\/code> folder contains scheduler settings used in the diffusion sampling process.<\/li>\n<li>VAE encoder: The <code>common\/vae_encoder<\/code> folder contains the model for converting images into latent representations.<\/li>\n<\/ul>\n<p>While the Transformer and VAE decoder are optimized for NPU operation, other components (text encoders, tokenizers, scheduler) are designed to be shared across GPU and NPU pipelines.<\/p>\n<h2>Performance<\/h2>\n<p>The <code>stable-diffusion-3.5-medium<\/code> model, which serves as the base for this version, is positioned as an improved version of the previous SD3 Medium.<\/p>\n<p>According to the model card, this model is reported to have improved prompt adherence, generation quality, and output diversity. It is also stated that improvements in typography (text rendering) and prompt understanding have been achieved.<\/p>\n<h2>Strengths and Use Cases<\/h2>\n<p>This model is suitable for generating images based on text prompts or modifying existing images. Inheriting the characteristics of the base Stable Diffusion 3.5 Medium model, it enables high-fidelity generation relative to prompts and high-quality, diverse outputs. In particular, it is well-suited for image generation involving complex instructions due to improved typography accuracy and prompt understanding.<\/p>\n<p>The most significant feature is its optimization for AMD NPU accelerators. This is intended to enable efficient text-to-image generation processes in environments such as PCs equipped with compatible hardware.<\/p>\n<p><!-- lmw:files --><\/p>\n<h2>Distributed Files<\/h2>\n<p><em>Weight files published in <a href=\"https:\/\/huggingface.co\/stabilityai\/stable-diffusion-3.5-medium-amdnpu-medusa\/tree\/main\">stabilityai\/stable-diffusion-3.5-medium-amdnpu-medusa<\/a>, listed by this site from the Hugging Face API. Sizes are the actual file sizes.<\/em><\/p>\n<div class=\"lmw-table-scroll\" tabindex=\"0\" style=\"overflow-x:auto;-webkit-overflow-scrolling:touch;max-width:100%;\">\n<table style=\"width:max-content;min-width:100%;border-collapse:collapse;\">\n<thead>\n<tr>\n<th>File<\/th>\n<th>Size<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td><code>common\/text_encoder\/model.onnx<\/code><\/td>\n<td>248MB<\/td>\n<\/tr>\n<tr>\n<td><code>common\/text_encoder_2\/model.onnx<\/code><\/td>\n<td>1.39GB<\/td>\n<\/tr>\n<tr>\n<td><code>common\/text_encoder_3\/model.onnx<\/code><\/td>\n<td>1MB<\/td>\n<\/tr>\n<tr>\n<td><code>common\/text_encoder_3_int4\/model.onnx<\/code><\/td>\n<td>1MB<\/td>\n<\/tr>\n<tr>\n<td><code>common\/tokenizer\/model.onnx<\/code><\/td>\n<td>1MB<\/td>\n<\/tr>\n<tr>\n<td><code>common\/tokenizer_2\/model.onnx<\/code><\/td>\n<td>1MB<\/td>\n<\/tr>\n<tr>\n<td><code>common\/tokenizer_3\/model.onnx<\/code><\/td>\n<td>1MB<\/td>\n<\/tr>\n<tr>\n<td><code>common\/vae_decoder\/dynamic\/dd\/replaced.onnx<\/code><\/td>\n<td>0MB<\/td>\n<\/tr>\n<tr>\n<td><code>common\/vae_encoder\/model.onnx<\/code><\/td>\n<td>69MB<\/td>\n<\/tr>\n<tr>\n<td><code>common\/vae_encoder\/optimized.onnx<\/code><\/td>\n<td>69MB<\/td>\n<\/tr>\n<tr>\n<td><code>common\/vae_encoder\/optimized_1024x1024.onnx<\/code><\/td>\n<td>69MB<\/td>\n<\/tr>\n<tr>\n<td><code>common\/vae_encoder\/optimized_512x512.onnx<\/code><\/td>\n<td>69MB<\/td>\n<\/tr>\n<tr>\n<td><code>normal\/transformer\/dynamic\/dd\/replaced.onnx<\/code><\/td>\n<td>2MB<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<p><!-- \/lmw:files --><\/p>\n<h2>How to Get It<\/h2>\n<p>This model is distributed in ONNX format. It can be obtained from the following Hugging Face repository:<\/p>\n<ul>\n<li>Repository: <code>stabilityai\/stable-diffusion-3.5-medium-amdnpu-medusa<\/code><\/li>\n<\/ul>\n<p>Additionally, the <code>sd-sandbox<\/code> repository provided by AMD is recommended as an environment for verifying the operation of this model.<br \/>\n&#8211; <a href=\"https:\/\/github.com\/amd\/sd-sandbox\">github.com\/amd\/sd-sandbox<\/a><\/p>\n<p>Since this model is not gated, no license agreement procedure is required for downloading.<\/p>\n<p><!-- lmw:runnability --><\/p>\n<h2>Can You Run It Locally?<\/h2>\n<p>The publisher distributes this model as ONNX.<\/p>\n<p><strong>License \u2014 <code>stabilityai-ai-community<\/code>:<\/strong> A custom license from the publisher. Check the original terms directly, including whether commercial use is permitted.<\/p>\n<p><em>Compiled by this site&#8217;s code from the published formats and the license field. License summaries are not legal advice \u2014 check the publisher&#8217;s original terms before relying on them.<\/em><\/p>\n<p><!-- \/lmw:runnability --><\/p>\n<p><!-- lmw:lab --><\/p>\n<h2>Our Own Measurements<\/h2>\n<h3>Measurements We Did Not Take<\/h3>\n<p>We have not confirmed that this site meets the commercial-use terms of this model&#8217;s license (<code>stabilityai-ai-community<\/code>). Because this site carries advertising, we did not run the model (no CPU run, answers, quantization comparison, conversion or generation).<\/p>\n<p><!-- \/lmw:lab --><\/p>\n<p><!-- lmw:same-task --><\/p>\n<h2>Other Models for the Same Task<\/h2>\n<p><em>Recent image generation models covered by Local Model Watch, newest first. Grouped by the task each publisher declares on Hugging Face (pipeline_tag); the smallest VRAM tier is this site&#8217;s estimate.<\/em><\/p>\n<ul>\n<li><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/11\/17\/minimax-h3-image-character-swap-2\/\">MiniMax-H3-Image-Character-Swap Image Editing Model: 4GB+ VRAM<\/a> (4GB)<\/li>\n<li><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/11\/17\/comfy-org-releases-world-model-pan2-for-comfyui\/\">PAN2 World Model for Simulating Physical Changes: ComfyUI Paths<\/a><\/li>\n<li><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/11\/17\/prunaai-qwen-image-2-1-remove-object\/\">PrunaAI Releases Object Removal LoRA for Qwen-Image-2.1<\/a> (7.1B)<\/li>\n<li><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/11\/17\/comfy-org-minimax-h3-image-released\/\">MiniMax-H3-Image 16-Megapixel Image Generation Model: 4GB+ VRAM<\/a> (4GB)<\/li>\n<\/ul>\n<p><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/models-by-task-en\/#task-image\">See all image generation models \u2192<\/a><\/p>\n<p><!-- \/lmw:same-task --><\/p>\n<p><!-- lmw:next-steps --><\/p>\n<h2>What to Read Next<\/h2>\n<ul>\n<li><strong>Formats this model is available in<\/strong> \u2192 <a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/format-onnx-en\/\">ONNX format guide and models<\/a><\/li>\n<li><strong>Learn about the publisher<\/strong> \u2192 <a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/publisher-stability-en\/\">Stability AI: models, licenses and articles<\/a><\/li>\n<\/ul>\n<p><!-- \/lmw:next-steps --><\/p>\n<h2>Sources<\/h2>\n<ul>\n<li><a href=\"https:\/\/huggingface.co\/stabilityai\/stable-diffusion-3.5-medium-amdnpu-medusa\">stabilityai\/stable-diffusion-3.5-medium-amdnpu-medusa<\/a><\/li>\n<\/ul>\n","protected":false},"excerpt":{"rendered":"<p>Stability AI releases Stable Diffusion 3.5 Medium optimized for AMD NPUs in ONNX format, improving prompt adherence and image quality.<\/p>\n","protected":false},"author":1,"featured_media":14017,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[203],"tags":[503,4282,1693,4284,4286,4288,1547,753],"class_list":["post-14018","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-image-video-and-audio","tag-amd-en","tag-npu-en","tag-onnx-en","tag-stability-ai-en","tag-stable-diffusion-3-5-en","tag-stable-diffusion-3-5-medium-amdnpu-medusa-en","tag-verified","tag--en"],"lang":"en","translations":{"en":14018,"ja":14016},"_links":{"self":[{"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/posts\/14018","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/comments?post=14018"}],"version-history":[{"count":3,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/posts\/14018\/revisions"}],"predecessor-version":[{"id":18257,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/posts\/14018\/revisions\/18257"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/media\/14017"}],"wp:attachment":[{"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/media?parent=14018"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/categories?post=14018"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/tags?post=14018"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}