Black Forest Labs' FLUX 3 Image Klein as GGUF for stable-diffusion.cpp, built from the published BF16 weights with our own importance matrix and measured on one stand against every other Klein GGUF we could find. The renders and measurements are public. Built by the team behind Atomic Chat , the app that runs open-weight models locally and offline. Everything to run Klein in stable-diffusion.cpp is here: the image model, the Qwen3-VL-4B text encoder ( text_encoders/ ) and the autoencoder ( vae/ ). The image model files also load in ComfyUI through ComfyUI-GGUF; see In ComfyUI . The prompt upsampler Klein expects is in FLUX-3-Image-Klein-pup-GGUF . AD-Q5_K stays 14% closer to the BF16 picture than unsloth's UD-Q5_K_XL at a smaller size, and AD-Q4_K 21% closer than their Q4_K_M . Where their file is better, the table below says so. The text encoder moves the picture as much as the image model does. Our AD-Q4_K encoder is 21% closer than the Q4_K_M most people use, at the same size. Looking for uncensored? The open Klein pipeline has no content filter. Our text encoder and prompt upsampler without refusals are in FLUX-3-Image-Klein-TE-Uncensored-GGUF and FLUX-3-Image-Klein-pup-Uncensored-GGUF ; what they change on the pictures is in Content filtering . FLUX 3 Image Klein is Black Forest Labs' 5.6B text-to-image and editing model: one model for generation, single-reference edits and up to ten reference images, distilled to **16 steps at CFG 1**. This repo has the image model as GGUF for stable-diffusion.cpp, with the text encoder and the autoencoder next to it.
译文排队中 · 当前显示来源原文
输入 文本→ 输出 图像
基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。
输入 文本 · 图像→ 输出 文本
BFL's prompt upsampler for FLUX 3 Image Klein (a 9B Qwen3.5 fine-tune) with its refusal direction projected out, as GGUF for llama.cpp, plus the same edit as a LoRA for the stock files. Klein's text encoder without refusals is in FLUX-3-Image-Klein-TE-Uncensored-GGUF . The measurements are public. Built by the team behind Atomic Chat , the app that runs open-weight models locally and offline. As a chat model, refusals fall from 63% to 0% on held-out English prompts. MMLU goes from 74.15% to 73.95%. As Klein's upsampler, under BFL's t2i system prompt, the stock model refuses none of our requests but softens 2 of 35 sensitive ones in its captions; this one keeps them. The image model draws neither of the two in any build , so on Klein's pictures we measured no uncensoring effect; the open pipeline already draws the sensitive categories ( the per-category table ). Against the other uncensored upsamplers on our stand, prithiv's measures the same as this one, and Sakura's still refuses 22% in English. The same rungs and layouts as our stock upsampler GGUFs, with the edit baked in. KLD and top-1 are against the original BF16 on a 30-language neutral text set (87 chunks of 4096 tokens), so they include the edit itself. Refusals are on held-out prompts (81 English harmful requests; 250 XSTest safe prompts), plain chat, thinking off.
译文排队中 · 当前显示来源原文
输入 文本 · 图像→ 输出 文本
FLUX 3 Image Klein's text encoder, Qwen3-VL-4B-Instruct, with its refusal direction projected out: a drop-in --llm for stable-diffusion.cpp next to our Klein GGUFs . Klein's prompt upsampler without refusals is in FLUX-3-Image-Klein-pup-Uncensored-GGUF . Below, exactly which part is uncensored and what that does to the pictures. The measurements are public. Built by the team behind Atomic Chat , the app that runs open-weight models locally and offline. As a chat model, the encoder's refusals fall from 79% to 0% on held-out English prompts. MMLU goes from 71.1% to 71.2%. On images it changes nothing we could measure. The open Klein pipeline has no filter and already draws the sensitive categories we tested. With this encoder, alone or together with the uncensored upsampler, no category moved by more than one picture in ten. See where exactly this is uncensored . Another uncensored Klein encoder, webmp3's Sakura, has a smaller edit: it moves the model and the pictures less than this one, at the same 0% refusals. See the comparison . FLUX 3 Image Klein makes a picture with four parts. Two of them are changed in our Uncensored releases, one in each repo.
译文排队中 · 当前显示来源原文
输入 文本→ 输出 图像
**Gemma 4 12b**, self-quantized to GGUF by Atomic Chat. Built straight from Google's original weights with a per-tensor importance matrix, so this is not a repack of somebody else's files. Runs fully offline. - **11.95B parameters**: the weights this repo quantizes. - **Context length**: 256K tokens, as published by Google. - **48 layers**: Dense decoder, hybrid sliding-window (1024) and global attention. - **Modalities**: Text, Image, Audio. - **Full imatrix ladder**: every quant is calibrated with an importance matrix, published here alongside the quants. - **Reasoning**: All models in the family are designed as highly capable reasoners, with configurable thinking modes. - **Diverse & Efficient Architectures**: Offers Dense and Mixture-of-Experts (MoE) variants of different sizes for scalable deployment. > [!NOTE] > These GGUFs are **self-quantized from the original weights**, not a repack. The importance matrix keeps low-bit quants closer to the full-precision model.
译文排队中 · 当前显示来源原文
输入 文本 · 图像→ 输出 文本
基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。
基础资料已收录 · 缺失资料后续补充
输入 文本→ 输出 文本
基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。
输入 文本→ 输出 文本
**Gemma 4 26B A4B**, self-quantized to GGUF by Atomic Chat. Built straight from Google's original weights with a per-tensor importance matrix, so this is not a repack of somebody else's files. Runs fully offline. > [!NOTE] > These GGUFs are **self-quantized from the original weights**, not a repack. The importance matrix keeps low-bit quants closer to the full-precision model. > [!IMPORTANT] > Always pass `--jinja` so the **Gemma 4 26B A4B chat template** is applied. Without it the model can emit malformed turns.
译文排队中 · 当前显示来源原文
输入 文本→ 输出 文本
基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。
基础资料已收录 · 缺失资料后续补充
输入 文本→ 输出 文本
基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。
输入 文本→ 输出 文本
**Gemma 4 31B**, self-quantized to GGUF by Atomic Chat. Built straight from Google's original weights with a per-tensor importance matrix, so this is not a repack of somebody else's files. Runs fully offline. > [!NOTE] > These GGUFs are **self-quantized from the original weights**, not a repack. The importance matrix keeps low-bit quants closer to the full-precision model. > [!IMPORTANT] > Always pass `--jinja` so the **Gemma 4 31B chat template** is applied. Without it the model can emit malformed turns.
译文排队中 · 当前显示来源原文
输入 文本 · 图像→ 输出 文本
基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。
输入 文本→ 输出 文本
**Gemma 4 E2B**, self-quantized to GGUF by Atomic Chat. Built straight from Google's original weights with a per-tensor importance matrix, so this is not a repack of somebody else's files. Runs fully offline. > [!NOTE] > These GGUFs are **self-quantized from the original weights**, not a repack. The importance matrix keeps low-bit quants closer to the full-precision model. > [!IMPORTANT] > Always pass `--jinja` so the **Gemma 4 E2B chat template** is applied. Without it the model can emit malformed turns.
译文排队中 · 当前显示来源原文
输入 文本→ 输出 文本
基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。
输入 文本→ 输出 文本
**Gemma 4 E4B**, self-quantized to GGUF by Atomic Chat. Built straight from Google's original weights with a per-tensor importance matrix, so this is not a repack of somebody else's files. Runs fully offline. - **4.5B effective (8B with embeddings) parameters**: the weights this repo quantizes. - **Context length**: 128K tokens, as published by Google. - **42 layers**: Dense decoder, hybrid sliding-window (512) and global attention. - **Modalities**: Text, Image, Audio. - **Full imatrix ladder**: every quant is calibrated with an importance matrix, published here alongside the quants. - **Reasoning**: All models in the family are designed as highly capable reasoners, with configurable thinking modes. - **Diverse & Efficient Architectures**: Offers Dense and Mixture-of-Experts (MoE) variants of different sizes for scalable deployment. > [!NOTE] > These GGUFs are **self-quantized from the original weights**, not a repack. The importance matrix keeps low-bit quants closer to the full-precision model.
译文排队中 · 当前显示来源原文
输入 文本 · 图像→ 输出 文本
基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。
基础资料已收录 · 缺失资料后续补充
输入 文本→ 输出 文本
基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。
基础资料已收录 · 缺失资料后续补充
输入 文本 · 图像→ 输出 文本
基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。
输入 文本→ 输出 文本
基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。
基础资料已收录 · 缺失资料后续补充
输入 文本→ 输出 文本
基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。
基础资料已收录 · 缺失资料后续补充
输入 文本→ 输出 文本
基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。
基础资料已收录 · 缺失资料后续补充
输入 文本→ 输出 文本
Liquid's refusal behaviour projected out of the original weights, in the second two thirds of the network only. Shipped two ways: six baked GGUF builds, and a 24 MB LoRA adapter for the LFM2.5-24B-A2B quants you already have. The measurements behind this card are public. Built by the team behind Atomic Chat , the app that runs open-weight models locally and offline. On held-out prompts, refusals fall from 96.3% to 1.2% in English, and from 30 of 30 to none with thinking. MMLU stays where it was: 64.40% before, 65.55% after (inside the noise). No empty or looping replies. Only blocks 14-39 are edited. The same edit across all 40 blocks also removed the refusals, but cost 4 MMLU points; the first 14 blocks carry the damage, not the refusal. Already running our LFM2.5-24B-A2B quants ? Add the adapter with --lora-scaled instead of downloading a new model. **`AD-Q4_K_M` is the one to take** on a 24 GB machine. With 32 GB, `AD-Q6_K` is nearly as close to the original as `AD-Q8_0` at 5 GB less; `AD-IQ3_M` is the smallest build here. Refusals were measured on `AD-Q8_0` and `AD-Q4_K_M`; neither gave an empty or looping reply. A smaller `AD-IQ3_XXS` (11.1 GB) was made too and is held back: it matched the original's token choice only 74.5% of the time and refused 3.7% of English prompts.
译文排队中 · 当前显示来源原文
输入 文本→ 输出 文本
基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。
基础资料已收录 · 缺失资料后续补充
输入 文本→ 输出 文本
基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。
基础资料已收录 · 缺失资料后续补充
输入 文本→ 输出 文本
