Buckets:
| <meta charset="utf-8" /><meta name="hf:doc:metadata" content="{"title":"BFL (fp8r / nvfp4)","local":"bfl-fp8r--nvfp4","sections":[{"title":"Kernels","local":"kernels","sections":[],"depth":2}],"depth":1}"/> | |
| <link href="/docs/diffusers/main/en/_app/immutable/entry/start.XW2G_w3j.js" rel="modulepreload"> | |
| <link href="/docs/diffusers/main/en/_app/immutable/chunks/CAM1Godo.js" rel="modulepreload"> | |
| <link href="/docs/diffusers/main/en/_app/immutable/chunks/Bb-LL0eD.js" rel="modulepreload"> | |
| <link href="/docs/diffusers/main/en/_app/immutable/entry/app.rg_Uuwoj.js" rel="modulepreload"> | |
| <link href="/docs/diffusers/main/en/_app/immutable/chunks/CY-Wb_o3.js" rel="modulepreload"> | |
| <link href="/docs/diffusers/main/en/_app/immutable/chunks/DnVIJfC5.js" rel="modulepreload"> | |
| <link href="/docs/diffusers/main/en/_app/immutable/chunks/DsnmJJEf.js" rel="modulepreload"> | |
| <link href="/docs/diffusers/main/en/_app/immutable/chunks/Cy2JwhB0.js" rel="modulepreload"> | |
| <link href="/docs/diffusers/main/en/_app/immutable/nodes/0.C1B0vLqo.js" rel="modulepreload"> | |
| <link href="/docs/diffusers/main/en/_app/immutable/chunks/DLds7x_i.js" rel="modulepreload"> | |
| <link href="/docs/diffusers/main/en/_app/immutable/nodes/15.D4qUgu09.js" rel="modulepreload"> | |
| <link href="/docs/diffusers/main/en/_app/immutable/chunks/CeW8NfJL.js" rel="modulepreload"> | |
| <!--7gq4hb--><meta name="hf:doc:metadata" content="{"title":"BFL (fp8r / nvfp4)","local":"bfl-fp8r--nvfp4","sections":[{"title":"Kernels","local":"kernels","sections":[],"depth":2}],"depth":1}"/><!----> | |
| <link href="/docs/diffusers/main/en/_app/immutable/assets/0.tn0RQdqM.css" rel="modulepreload"> <!--[--><!--[0--><!--[--><!--[0--><!--[--><p></p> <div class="items-center shrink-0 min-w-[100px] max-sm:min-w-[50px] justify-end ml-auto flex" style="float: right; margin-left: 10px; display: inline-flex; position: relative; z-index: 10;"><div class="inline-flex rounded-md max-sm:rounded-sm"><button class="inline-flex items-center gap-1 h-7 max-sm:h-7 px-2 max-sm:px-1.5 text-sm font-medium text-gray-800 border border-r-0 rounded-l-md max-sm:rounded-l-sm border-gray-200 bg-white hover:shadow-inner dark:border-gray-850 dark:bg-gray-950 dark:text-gray-200 dark:hover:bg-gray-800" aria-live="polite"><span class="inline-flex items-center justify-center rounded-md p-0.5 max-sm:p-0 hover:text-gray-800 dark:hover:text-gray-200"><svg class="sm:size-3.5 size-3" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg><!----></span> <span>Copy page</span></button> <button class="inline-flex items-center justify-center w-6 max-sm:w-5 h-7 max-sm:h-7 disabled:pointer-events-none text-sm text-gray-500 hover:text-gray-700 dark:hover:text-white rounded-r-md max-sm:rounded-r-sm border border-l transition border-gray-200 bg-white hover:shadow-inner dark:border-gray-850 dark:bg-gray-950 dark:text-gray-200 dark:hover:bg-gray-800" aria-haspopup="menu" aria-expanded="false" aria-label="Open copy menu"><svg class="transition-transform text-gray-400 overflow-visible sm:size-3.5 size-3 rotate-0" width="1em" height="1em" viewBox="0 0 12 7" fill="none" xmlns="http://www.w3.org/2000/svg"><path d="M1 1L6 6L11 1" stroke="currentColor"></path></svg><!----></button></div> <!--[-1--><!--]--></div><!----> <!--[0--><h1 class="relative group"><a id="bfl-fp8r--nvfp4" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#bfl-fp8r--nvfp4"><span><svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg><!----></span></a> <span>BFL (fp8r / nvfp4)</span></h1><!--]--><!----> <p><a href="https://huggingface.co/black-forest-labs" rel="nofollow">Black Forest Labs</a> ships quantized variants of its <a href="../api/pipelines/flux3">Flux3</a> transformers in two schemes, stored as single-file checkpoints next to the original bf16 weights:</p> <ul><li><code>fp8r</code>: rowwise FP8. Each quantized linear stores an E4M3 <code>weight</code> and an fp32 per-row <code>weight_scale</code>. Activations are quantized to E4M3 per row at runtime.</li> <li><code>nvfp4</code>: NVFP4 in the <a href="https://github.com/NVIDIA/Model-Optimizer" rel="nofollow">ModelOpt</a> layout. Each quantized linear stores a packed E2M1 <code>weight</code> (two values per byte), E4M3 block scales of 16 (<code>weight_scale</code>), a global fp32 <code>weight_scale_2</code>, and a calibrated fp32 activation scale <code>input_scale</code>.</li></ul> <p>Diffusers loads both through <code>~FromSingleFileMixin.from_single_file</code> with a <a href="/docs/diffusers/pr_15081/en/api/quantization#diffusers.BFLQuantizationConfig">BFLQuantizationConfig</a>. The scheme is read from the checkpoint, so the same config works for either file. Only loading prequantized checkpoints is supported; pipeline-level loading is not.</p> <p>The example below loads the <code>fp8r</code> checkpoint of <a href="https://huggingface.co/black-forest-labs/flux-3-image-klein-base" rel="nofollow">FLUX 3 Image Klein Base</a>, which lives at <code>variants/fp8r/model.safetensors</code> in the model repository. Pass <code>config</code> so that the model config is read from the Diffusers folder of that repository, then load the other components by name (see <a href="../api/pipelines/flux3#single-file-loading">Single-file loading</a>).</p> <div class="code-block relative "><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg><!----> <div class=" absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0 "><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent;"></div> Copied</div><!----></button><!----></div> <pre class="language-python "><!----><span class="hljs-keyword">import</span> torch | |
| <span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> BFLQuantizationConfig, Flux3Transformer2DModel, ModularPipeline | |
| ckpt_path = <span class="hljs-string">"https://huggingface.co/black-forest-labs/flux-3-image-klein-base/blob/main/variants/fp8r/model.safetensors"</span> | |
| transformer = Flux3Transformer2DModel.from_single_file( | |
| ckpt_path, | |
| config=<span class="hljs-string">"black-forest-labs/flux-3-image-klein-base"</span>, | |
| subfolder=<span class="hljs-string">"transformer"</span>, | |
| quantization_config=BFLQuantizationConfig(compute_dtype=torch.bfloat16), | |
| dtype=torch.bfloat16, | |
| ) | |
| pipe = ModularPipeline.from_pretrained(<span class="hljs-string">"black-forest-labs/flux-3-image-klein-base"</span>) | |
| pipe.load_components(dtype=torch.bfloat16) | |
| pipe.update_components(transformer=transformer) | |
| pipe.to(<span class="hljs-string">"cuda"</span>) | |
| prompt = <span class="hljs-string">"A wide-angle wildlife photograph of a red fox standing alert in a snow-covered forest clearing at golden hour."</span> | |
| image = pipe(prompt=prompt, generator=torch.Generator().manual_seed(<span class="hljs-number">0</span>), output=<span class="hljs-string">"images"</span>)[<span class="hljs-number">0</span>] | |
| image.save(<span class="hljs-string">"flux3-klein-base-fp8r.png"</span>)<!----></pre></div><!----> <p>Replace <code>variants/fp8r</code> with <code>variants/nvfp4</code> to load the <code>nvfp4</code> variant. The step-distilled <a href="https://huggingface.co/black-forest-labs/flux-3-image-klein" rel="nofollow">FLUX 3 Image Klein</a> ships the same two variants; load them with its repository id as <code>config</code> and pipeline.</p> <!--[1--><h2 class="relative group"><a id="kernels" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#kernels"><span><svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg><!----></span></a> <span>Kernels</span></h2><!--]--><!----> <p>The quantized weights stay in their low-precision storage dtype. Which GEMM runs depends on the device:</p> <table><thead><tr><th>Scheme</th><th>Fused kernel</th><th>Requirements</th></tr></thead><tbody><tr><td><code>fp8r</code></td><td><code>torch._scaled_mm</code> with rowwise scales</td><td>CUDA GPU with compute capability 8.9 or higher</td></tr><tr><td><code>nvfp4</code></td><td><code>flashinfer.mm_fp4</code> (CUTLASS) with <code>flashinfer.nvfp4_quantize</code> for the activations</td><td>Blackwell GPU (compute capability 10.0 or higher), a CUDA toolkit with <code>nvcc</code>, and <code>flashinfer-python</code> with <code>flashinfer-cubin</code> at the same version</td></tr></tbody></table> <p>Install the two FlashInfer packages at the same version, for example <code>pip install flashinfer-python==0.6.13 flashinfer-cubin==0.6.13</code>. The <code>flashinfer-cubin</code> release can lag behind <code>flashinfer-python</code>, and FlashInfer refuses to import when the versions differ.</p> <p>FlashInfer compiles its NVFP4 kernels on first use with the CUDA toolkit of the environment and picks the target architecture from the visible GPU (or from <code>FLASHINFER_CUDA_ARCH_LIST</code> when set). Consumer and workstation Blackwell GPUs (compute capability 12.0) need a CUDA toolkit of 12.9 or newer. With an older toolkit FlashInfer finds no eligible architecture and the first NVFP4 call fails with <code>FlashInfer requires GPUs with sm75 or higher</code>.</p> <p>On any other device the weights are dequantized to <code>compute_dtype</code> in each forward pass and the layer runs as a regular linear. In that case the activations are not quantized, so outputs differ slightly from the fused kernels.</p> <a class="!text-gray-400 !no-underline text-sm flex items-center not-prose mt-4" href="https://github.com/huggingface/diffusers/blob/main/docs/source/en/quantization/bfl.md" target="_blank"><svg class="mr-1" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M31,16l-7,7l-1.41-1.41L28.17,16l-5.58-5.59L24,9l7,7z"></path><path d="M1,16l7-7l1.41,1.41L3.83,16l5.58,5.59L8,23l-7-7z"></path><path d="M12.419,25.484L17.639,6.552l1.932,0.518L14.351,26.002z"></path></svg><!----> <span><span class="underline">Update</span> on GitHub</span></a><!----> <p></p><!--]--><!----><!--]--><!--]--><!--]--> <!--[-1--><!--]--><!--]--> | |
| <script> | |
| { | |
| __sveltekit_mpy3v7 = { | |
| base: "/docs/diffusers/main/en", | |
| assets: "/docs/diffusers/main/en" | |
| }; | |
| const element = document.currentScript.parentElement; | |
| Promise.all([ | |
| import("/docs/diffusers/main/en/_app/immutable/entry/start.XW2G_w3j.js"), | |
| import("/docs/diffusers/main/en/_app/immutable/entry/app.rg_Uuwoj.js") | |
| ]).then(([kit, app]) => { | |
| kit.start(app, element, { | |
| node_ids: [0, 15], | |
| data: [null,null], | |
| form: null, | |
| error: null | |
| }); | |
| }); | |
| } | |
| </script> | |
Xet Storage Details
- Size:
- 13.8 kB
- Xet hash:
- 2351ab075ce6a242a504b96693d64c6a07a45176fae650ccb9bcccfc844fea8e
·
Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.