<h1 id="newbie-image-exp0.1">NewBie image Exp0.1</h1><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/20bec9e6-a22e-421e-a570-8f8e2f1f8a50/width=525/20bec9e6-a22e-421e-a570-8f8e2f1f8a50.jpeg" /><ul><li><a target="_blank" rel="ugc" href="https://github.com/NewBieAI-Lab/NewBie-image-Exp0.1">GitHub (Code)</a></li></ul><ul><li><a target="_blank" rel="ugc" href="https://github.com/NewBieAI-Lab/NewbieLoraTrainer">GitHub (LoRA Trainer)</a></li></ul><ul><li><a target="_blank" rel="ugc" href="https://github.com/E-Anlia/ComfyUI-NewBie">GitHub (ComfyUI)</a></li></ul><ul><li><a target="_blank" rel="ugc" href="https://huggingface.co/NewBie-AI/NewBie-image-Exp0.1">Hugging Face</a></li></ul><ul><li><a target="_blank" rel="ugc" href="https://www.modelscope.cn/models/NewBieAi-lab/NewBie-image-Exp0.1">ModelScope</a></li></ul><h3 id="exp0.1-base">🧱 Exp0.1 Base</h3><ul><li>NewBie image Exp0.1 is a 3.5B parameter DiT model developed through research on the Lumina architecture.Building on these insights, it adopts Next-DiT as the foundation to design a new NewBie architecture tailored for text-to-image generation.The NewBie image Exp0.1 model is trained within this newly constructed system, representing the first experimental release of the NewBie text-to-image generation framework.</li></ul>Text Encoders<ul><li>We use Gemma3-4B-it as the primary text encoder, conditioning on its penultimate-layer token hidden states. We also extract pooled text features from Jina CLIP v2, project them, and fuse them into the time/AdaLN conditioning pathway. Together, Gemma3-4B-it and Jina CLIP v2 provide strong prompt understanding and improved instruction adherence.</li></ul>VAE<ul><li>Use the FLUX.1-dev 16channel VAE to encode images into latents, delivering richer, smoother color rendering and finer texture detail helping safeguard the stunning visual quality of NewBie image Exp0.1.</li></ul>Prompt<ul><li>XML structured prompt</li><li>Natural language prompt</li><li>Tag prompt</li></ul><h3 id="task-type">🖼️ Task type</h3>NewBie image Exp0.1 is pretrain on a large corpus of high-quality anime data, enabling the model to generate remarkably detailed and visually striking anime style images.<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/f0faf6da-f5f3-4969-93ba-e70170aac56d/width=525/f0faf6da-f5f3-4969-93ba-e70170aac56d.jpeg" />We reformatted the dataset text into an XML structured format for our experiments. Empirically, this improved attention binding and attribute/element disentanglement, and also led to faster convergence.Besides that, It also supports natural language and tags inputs.<h3 id="model-zoo">🧰 Model Zoo</h3>NewBie image Exp0.1: <a target="_blank" rel="ugc" href="https://huggingface.co/NewBie-AI/NewBie-image-Exp0.1">Hugging face</a> | <a target="_blank" rel="ugc" href="https://www.modelscope.cn/models/NewBieAi-lab/NewBie-image-Exp0.1">modelscope</a>Gemma3-4B-it: <a target="_blank" rel="ugc" href="https://huggingface.co/google/gemma-3-4b-it">Hugging face</a> | <a target="_blank" rel="ugc" href="https://www.modelscope.cn/models/google/gemma-3-4b-it">modelscope</a>Jina CLIP v2: <a target="_blank" rel="ugc" href="https://huggingface.co/jinaai/jina-clip-v2">Hugging face</a> | <a target="_blank" rel="ugc" href="https://www.modelscope.cn/models/jinaai/jina-clip-v2">modelscope</a>FLUX.1-dev VAE: <a target="_blank" rel="ugc" href="https://huggingface.co/black-forest-labs/FLUX.1-dev/blob/main/vae/diffusion_pytorch_model.safetensors">Hugging face</a> | <a target="_blank" rel="ugc" href="https://www.modelscope.cn/models/black-forest-labs/FLUX.1-dev/tree/master/vae">modelscope</a><h3 id="training-procedure">💪 Training procedure</h3><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/7c61edfd-5558-4129-b780-f2669e67ce53/width=525/7c61edfd-5558-4129-b780-f2669e67ce53.jpeg" /><h3 id="participate">🔬 Participate</h3>Core<ul><li><a target="_blank" rel="ugc" href="https://huggingface.co/E-Anlia">Anlia</a> | <a target="_blank" rel="ugc" href="https://huggingface.co/CreeperMZ">CreeperMZ</a> | <a target="_blank" rel="ugc" href="https://huggingface.co/LAXMAYDAY">L_A_X</a> | <a target="_blank" rel="ugc" href="https://huggingface.co/maikaaomi">maikaaomi</a> | <a target="_blank" rel="ugc" href="https://huggingface.co/xuefei123456">waw1w1</a> | <a target="_blank" rel="ugc" href="https://huggingface.co/LakiCat">LakiCat</a> | <a target="_blank" rel="ugc" href="https://huggingface.co/windsingai">chenkin</a> | <a target="_blank" rel="ugc" href="https://huggingface.co/aplx">aplxaplx</a> | <a target="_blank" rel="ugc" href="https://huggingface.co/GuChen">NULL</a></li></ul>Members<ul><li><a target="_blank" rel="ugc" href="https://huggingface.co/niangao233">niangao233</a> | <a target="_blank" rel="ugc" href="https://huggingface.co/ginkgowm">ginkgowm</a> | <a target="_blank" rel="ugc" href="https://huggingface.co/leafmoone">leafmoone</a> | <a target="_blank" rel="ugc" href="https://huggingface.co/NaviVoid">NaviVoid</a> | <a target="_blank" rel="ugc" href="https://huggingface.co/Emita">Emita</a> | <a target="_blank" rel="ugc" href="https://huggingface.co/TLFZ">TLFZ</a> | <a target="_blank" rel="ugc" href="https://huggingface.co/3HOOO">3HOOO</a></li></ul><h3 id="acknowledgments">✨ Acknowledgments</h3><ul><li>Thanks to the <a target="_blank" rel="ugc" href="https://huggingface.co/Alpha-VLLM">Alpha-VLLM Org</a> for open sourcing the advanced <a target="_blank" rel="ugc" href="https://huggingface.co/collections/Alpha-VLLM/lumina-family">Lumina family</a>. which has been invaluable for our research.</li></ul><ul><li>Thanks to <a target="_blank" rel="ugc" href="https://huggingface.co/google">Google</a> for open sourcing the powerful <a target="_blank" rel="ugc" href="https://huggingface.co/google/gemma-3-4b-it">Gemma3</a> LLM family</li><li>Thanks to the <a target="_blank" rel="ugc" href="https://huggingface.co/jinaai">Jina AI Org</a> for open sourcing the <a target="_blank" rel="ugc" href="https://huggingface.co/jinaai/jina-clip-v2">Jina</a> family, enabling further research.</li><li>Thanks to <a target="_blank" rel="ugc" href="https://huggingface.co/black-forest-labs">Black Forest Labs</a> for open sourcing the <a target="_blank" rel="ugc" href="https://huggingface.co/black-forest-labs/FLUX.1-dev/blob/main/vae">FLUX VAE</a> family. powerful 16channel VAE is one of the key components behind improved image quality.</li></ul><ul><li>Thanks to <a target="_blank" rel="ugc" href="https://huggingface.co/neta-art">Neta.art</a> for fine-tuning and open sourcing the <a target="_blank" rel="ugc" href="https://huggingface.co/Alpha-VLLM/Lumina-Image-2.0">Lumina-image-2.0</a> base model. <a target="_blank" rel="ugc" href="https://huggingface.co/neta-art/Neta-Lumina">Neta-Lumina</a> gives us the opportunity to study the performance of Next-DiT on Anime Types.</li><li>Thanks to <a target="_blank" rel="ugc" href="https://huggingface.co/deepghs">DeepGHS</a>/<a target="_blank" rel="ugc" href="https://huggingface.co/narugo1992">narugo1992</a>/<a target="_blank" rel="ugc" href="https://huggingface.co/SumomoLee">SumomoLee</a> for providing high-quality Anime Datasets.</li><li>Thanks to <a target="_blank" rel="ugc" href="https://huggingface.co/nyanko7">Nyanko</a> for the early help and support.</li><li>Thanks to <a target="_blank" rel="ugc" href="https://huggingface.co/woctordho">woctordho</a> for helping improve NewBie’s compatibility with community tools.</li></ul><h3 id="contribute">📖 Contribute</h3><ul><li>Neko, 衡鲍, XiaoLxl, xChenNing, Hapless, Lius</li><li>WindySea, 秋麒麟热茶, 古柯, Rnglg2, Ly, GHOSTLXH</li><li>Sarara, Seina, KKT机器人, NoirAlmondL, 天满, 暂时</li><li>Wenaka喵, ZhiHu, BounDless, DetaDT, 紫影のソナーニル</li><li>花火流光, R3DeK, 圣人A, 王王玉, 乾坤君Sennke, 砚青</li><li>Heathcliff01, 无音, MonitaChan, WhyPing, TangRenLan</li><li>HomemDesgraca, EPIC, ARKBIRD, Talan, 448, Hugs288</li></ul><h3 id="community-guide">🧭 Community Guide</h3>Getting Started Guide<ul><li><a target="_blank" rel="ugc" href="https://ai.feishu.cn/wiki/NZl9wm7V1iuNzmkRKCUcb1USnsh">English</a></li><li><a target="_blank" rel="ugc" href="https://ai.feishu.cn/wiki/P3sgwUUjWih8ZWkpr0WcwXSMnTb">中文</a></li></ul>LoRa Trainer<ul><li><a target="_blank" rel="ugc" href="https://www.notion.so/Newbie-AI-lora-training-tutorial-English-2c2e4ae984ab8177b312e318827657e6?source=copy_link">English</a></li><li><a target="_blank" rel="ugc" href="https://www.notion.so/Newbie-AI-lora-2b84f7496d81803db524f5fc4a9c94b9?source=copy_link">中文</a></li></ul><h3 id="communication">💬 Communication</h3><ul><li><a target="_blank" rel="ugc" href="https://discord.gg/bDJjy7rBGm">Discord</a></li><li><a target="_blank" rel="ugc" href="https://pd.qq.com/s/a79to55q6">解构原典</a></li><li><a target="_blank" rel="ugc" href="https://qm.qq.com/q/qnHFwN9fSE">ChatGroup</a></li></ul><h3 id="license">📜 License</h3><ul><li>Model Weights: Newbie Non-Commercial Community License (Newbie-NC-1.0).Applies to: model weights/parameters/configs and derivatives (fine-tunes, LoRA, merges, quantized variants, etc.)For Non Commercial use only, and must be shared under the same license.See <a target="_blank" rel="ugc" href="https://huggingface.co/NewBie-AI/NewBie-image-Exp0.1/blob/main/LICENSE.md">NewBie-image-Exp0.1 LICENSE.md</a></li></ul><ul><li>Code: Apache License 2.0.Applies to: training/inference scripts and related source code in this project.See <a target="_blank" rel="ugc" href="https://www.apache.org/licenses/LICENSE-2.0">Apache-2.0</a></li></ul><h3 id="disclaimer">⚠️ Disclaimer</h3>This model may produce unexpected or harmful outputs. Users are solely responsible for any risks and potential consequences arising from its use.

NewBie image

Christmas is just around the corner 🎄 Today we (NewBieAi-Lab) are excited to release NewBie image Exp0.1! Also, I’d love to recommend a song I really enjoy: “52 Hearts.” It fits the season perfectly. Wishing you all a wonderful Christmas and a relaxing holiday—have fun and enjoy! ✨

Files with SHA256: b8ed0f26c0b90b70a5a2ef0f69041529af7fbfef206f773c07d6cd1d56d77ef8