A repackaged nvidia/Gemma-4-31B-IT-NVFP4 that is **68% smaller** in GPU memory and **~2.5× faster** than the base model, while retaining **nearly identical quality** (1-3% loss). Fits on a *single* RTX 5090 (🎉). It fully leverages NVIDIA Blackwell FP4 tensor cores (RTX 5090, RTX PRO 6000, B200, and other SM 12.0+ GPUs with ≥20 GB VRAM) for **~2× higher concurrent throughput** than other quants like prithivMLmods/gemma-4-31B-it-NVFP4 or cyankiwi/gemma-4-31B-it-AWQ-4bit. This variant is **text-only**, video/audio weights and encoders have been stripped. If you need video/audio support -> open an issue or PR.

译文排队中 · 当前显示来源原文

文本生成

输入文本→ 输出文本

价格待核验多模态与渠道费用须分别核对

查看详情 ↗

加入对比来源 / 文档 ↗

flint-abliterated-GGUF

lilcheaty

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。

文本生成

输入文本→ 输出文本

价格待核验多模态与渠道费用须分别核对

查看详情 ↗

加入对比来源 / 文档 ↗

JANKUTrainedChenkinNoobai_v777

LillyCherry

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。

图像生成

输入文本→ 输出图像

价格待核验多模态与渠道费用须分别核对

查看详情 ↗

加入对比来源 / 文档 ↗

JANKUTrainedNoobaiRouwei_v69

LillyCherry

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

图像生成

输入文本→ 输出图像

价格待核验多模态与渠道费用须分别核对

查看详情 ↗

加入对比来源 / 文档 ↗

abcdef

lilmeaty

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

文本生成

输入文本→ 输出文本

价格待核验多模态与渠道费用须分别核对

查看详情 ↗

加入对比来源 / 文档 ↗

SmolLM2-135M-Instruct-GPTQ

lilmeaty

1. Model Summary 2. Limitations 3. Training 4. License 5. Citation SmolLM2 is a family of compact language models available in three size: 135M, 360M, and 1.7B parameters. They are capable of solving a wide range of tasks while being lightweight enough to run on-device. SmolLM2 demonstrates significant advances over its predecessor SmolLM1, particularly in instruction following, knowledge, reasoning. The 135M model was trained on 2 trillion tokens using a diverse dataset combination: FineWeb-Edu, DCLM, The Stack, along with new filtered datasets we curated and will release soon. We developed the instruct version through supervised fine-tuning (SFT) using a combination of public datasets and our own curated datasets. We then applied Direct Preference Optimization (DPO) using UltraFeedback.

译文排队中 · 当前显示来源原文

文本生成

输入文本→ 输出文本

价格待核验多模态与渠道费用须分别核对

查看详情 ↗

加入对比来源 / 文档 ↗

MiniMax-H3

LILUMY

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

分类待补充

输入待核验→ 输出待核验

价格待核验多模态与渠道费用须分别核对

查看详情 ↗

加入对比来源 / 文档 ↗

AnyPose

lilylilith

Creating poses for Qwen Image Edit (2511) or its variants can be difficult. Even with openpose being baked into the model, it can still look like something is off with the result, such as the depth being incorrect or the pose not fully matching the character, with distortions that you wouldn't expect. Additionally, creating them by hand can take a bit too long, especially if you are going the blender openpose rig route. What if instead you could just skip openpose entirely and write a simple (but albeit lengthy) prompt to copy the pose of any image on your computer? Well, this is what the AnyPose LoRAs attempt to do. Made in mind with the new Qwen Image Edit 2511 lightning LoRA for fast inference, with just a single reference image as a pose guide, you can pilot any image to follow that pose. No control net needed. Use 4-step lightning LoRA for Qwen Image Edit 2511. Set both of the AnyPose LoRA weights (base & helper) strength to 0.7. Upload two images, one with the initial input character, and another with the pose you want copied to the first character. Use the following prompt, "Make the person in image 1 do the exact same pose of the person in image 2. Changing the style and background of the image of the person in image 1 is undesirable, so don't do it. The new pose should be pixel accurate to the pose we are trying to copy. The position of the arms and head and legs should be the same as the pose we are trying to copy. Change the field of view and angle to match exactly image 2. Head tilt and eye gaze pose should match the person in image 2." If additional context is needed (such as clothing that is unseen from the model), prompt it in at the end. Background not staying the same as the initial input image? Append the following to the end "Remove the background of image 2, and replace it with the background of image 1."

译文排队中 · 当前显示来源原文

图像编辑

输入图像→ 输出图像

价格待核验多模态与渠道费用须分别核对

查看详情 ↗

加入对比来源 / 文档 ↗

QIE-2511-MP-AnyLight

lilylilith

Lighting is the hardest thing about AI image generation by far. Even if you specify the lighting direction with keywords, the level of control we have for lighting is primitive to say the least. It is very hard to get the desired lighting that you are actually envisioning, not just generic lighting setups like from the left, right, etc, but actual complex lighting setups that are hard to specify in just words. Professionals and people who are serious with getting the right results require it, because currently lighting is for all intents and purposes completely random. You cannot specify the distance the AI generates the light (eg. 40ft away), nor the specific X,Y,Z angle of the light. There have been other attempts to light up a scene, such as Lighting Remap Alpha0.2, which help position lights in various positions. This does help, however, it still falls to the pitfalls mentioned previously since it's based on a 2d canvas. Wouldn't it be better to have some sort of semi real-time lighting setup where you can roughly make sure the character is lit up the way that you want it while not being bound to two-dimensions? Well, this is what this LoRA attempts to do. How is this possible? Well, with a mix of a couple LoRAs (Material Preview style), Trellis2 model generation (or Hunyuan 2.1/3), and Blender, we can have more precise control over the lighting in the scene.

译文排队中 · 当前显示来源原文

图像编辑

输入图像→ 输出图像

价格待核验多模态与渠道费用须分别核对

查看详情 ↗

加入对比来源 / 文档 ↗

pixel-art-lora

Limbicnation

A LoRA adapter trained on FLUX.2-klein-4B for generating pixel art character sprites. Optimized for game-ready assets with transparent backgrounds. 2. Load FLUX.2-klein-4B as the base checkpoint 3. Add a **Load LoRA** node, point to `pytorch_lora_weights.safetensors` 4. Set LoRA strength: **0.85 - 1.4** (tested range) 5. Use trigger words in your positive prompt This model was trained with full data provenance tracking. All source images are CC0 public domain, making the outputs safe for commercial game development, Steam releases, and B2B studio use.

译文排队中 · 当前显示来源原文

图像生成

输入文本→ 输出图像

价格待核验多模态与渠道费用须分别核对

查看详情 ↗

加入对比来源 / 文档 ↗

qwen2-5-7b-image-prompt-lora-v2

Limbicnation

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

分类待补充

输入待核验→ 输出待核验

价格待核验多模态与渠道费用须分别核对

查看详情 ↗

加入对比来源 / 文档 ↗

Qwen3.5-4B-GGUF

Limeboi

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。

视觉问答

输入文本 · 图像→ 输出文本

价格待核验多模态与渠道费用须分别核对

查看详情 ↗

加入对比来源 / 文档 ↗

M7-7b

liminerity

assistant: "Here's an ASCII art attempting to symbolize the essence of an LLM (Large Language Model): assistant: "In this depiction, the LLM is characterized by its vast memory, instinctual roots, and the ability to learn from vast amounts of text—represented by the woven words in the lower part. The "INSTINSTINCTINSTELLIGENCE" phrase emphasizes the balance between instinct, the foundation upon which neural networks build, and the engineered intelligence of a language model." Multiverse-Experiment-slerp-7b is a merge of the following models using mergekit: * liminerity/merge4 * liminerity/merge2

译文排队中 · 当前显示来源原文

文本生成

输入文本→ 输出文本

价格待核验多模态与渠道费用须分别核对

查看详情 ↗

加入对比来源 / 文档 ↗