Hugging Face 热门模型第 26 名:ccore/Qwen3.5-4B-3D-Coder,166 赞,近期下载 3544 次。
标签:gguf, qwen3_5, code, threejs, 3d, preview, base_model:Qwen/Qwen3.5-4B, base_model:finetune:Qwen/Qwen3.5-4B, license:apache-2.0, endpoints_compatible, region:us, conversational
README:
Qwen3.5-4B-3D-Coder
早期预览 · 通用编码,侧重于 3D
一项探索小型语言模型在通用编码和高质量 3D 场景生成方面可行性的实验。
这是我仍在训练的 4B 模型的早期预览版本,特别侧重于使用 Three.js 创建交互式 3D 场景。初步的实际测试结果在通用编码方面令人鼓舞,尤其是在生成 3D 场景方面,因此我决定在开发继续的同时分享这一进展。
这是由 CCORE 进行的独立微调,基于 Qwen/Qwen3.5-4B。这不是官方的 Qwen 发布版本。
更新 — 2026 年 10 月 26 日
Q4_K_M 和 Q8_0 均已更新。如果您下载了较早的版本,请重新下载文件以获取此更新。
- 修复了部分用户报告的与层相关的问题。
- 在我最新的测试中,此次更新似乎能产生更多成功的单次场景生成结果。
下图展示了我在测试中的一个示例对比。

训练和发布时间表
预计训练将于 2026 年 10 月 28 日 UTC 时间 07:30 完成。
在 2026 年 10 月 26 日和 27 日,我将把 Q8_0 GGUF 更新到最新可用的训练检查点。
训练完成后,我将发布:
- 完整的 FP16 模型权重。
- 全套量化版本。
- 训练数据集,以便其他人可以使用它来训练或微调其他模型。
- 评估报告,包含各种提示词,展示模型在整个训练过程中在这些提示词上的表现。
早期示例
精选渲染后的 Three.js 场景截图:通过 HTML 和 JavaScript 构建的风格化建筑、花园、体素环境、家具、车间内部和机械物体。
点击任何图片可查看原始全屏截图。
樱花寺花园
日落村庄
龙骑士巢穴
体素发射场
装订工作台
LGA 3647 插槽
鹈鹕自行车 No. 07
弓形前部秘书桌
光琳庭园
樱花盆景
生成设置
对 Q4_K_M 和 Q8 版本使用以下设置:
| 参数 |
值 |
| 温度 (Temperature) |
0.6–0.8 |
| Top P |
0.90–0.95 |
| 重复惩罚 (Repeat penalty) |
1.0 |
| Top K |
20–40 |
LM Studio:禁用 MTP
重要提示: 如果您在 LM Studio 中运行此模型,请禁用多令牌预测 (MTP),否则模型无法正常工作。LM Studio 默认启用此选项。
Three.js 场景的系统提示词
仅使用来自
https://cdnjs.cloudflare.com/ajax/libs/three.js/0.160.0/three.module.min.js
的 Three.js 核心 ES 模块。
不要导入附加组件或 legacy examples/js 路径;
将任何相机控制直接实现在 HTML 中。
示例提示词
使用上述系统提示词,然后尝试以下用户提示词之一。
大象康复栖息地
请创建一个自包含的 HTML 文件,使用来自 cdnjs 的 Three.js 展示
一个大象康复栖息地。
一个宽阔的有遮蔽庭院通过无障碍通道连接水域、阴凉区和喂食区。请使用大象的尺寸和运动方式
查看原文
Hugging Face 热门模型第 26 名:ccore/Qwen3.5-4B-3D-Coder,166 赞,近期下载 3544 次。
标签:gguf, qwen3_5, code, threejs, 3d, preview, base_model:Qwen/Qwen3.5-4B, base_model:finetune:Qwen/Qwen3.5-4B, license:apache-2.0, endpoints_compatible, region:us, conversational
README:
Qwen3.5-4B-3D-Coder
Early preview · General coding with a focus on 3D
An experiment exploring the feasibility of small language models for general-purpose coding and high-quality 3D scene generation.
This is an early preview of a 4B model I am still training, with a particular focus on creating interactive 3D scenes with Three.js. Initial hands-on results have been promising for coding in general, and especially for generating 3D scenes, so I decided to share the progress while development continues.
This is an independent fine-tune by CCORE, based on Qwen/Qwen3.5-4B. It is not an official Qwen release.
Update — October 26, 2026
Both Q4_K_M and Q8_0 have been updated. If you downloaded an earlier version, please download the files again to get this update.
- Fixed the layer-related issue reported by some users.
- In my latest tests, this update appears to produce more successful one-shot scene generations.
The image below shows one example comparison from my tests.

Training and release schedule
Training is expected to finish on October 28, 2026, at 07:30 UTC.
On October 26 and 27, 2026, I will update the Q8_0 GGUF to the latest available training checkpoint.
After training finishes, I will publish:
- The full FP16 model weights.
- The full set of quantized versions.
- The training dataset, so others can use it to train or fine-tune other models.
- An evaluation report with a variety of prompts, showing how the model performed on them throughout training.
Early examples
Selected screenshots of rendered Three.js scenes: stylized architecture, gardens, voxel environments, furniture, workshop interiors, and mechanical objects, built through HTML and JavaScript.
Click any image to view the original screenshot at full size.
Sakura Temple Garden
Sunset Village
Dragonrider Aerie
Voxel launch site
The Bookbinding Bench
LGA 3647 Socket
Pelican Bike No. 07
Bow-Front Secretary Desk
Kōrin Garden
Sakura bonsai
Generation settings
Use the following settings for both the Q4_K_M and Q8 versions:
| Parameter |
Value |
| Temperature |
0.6–0.8 |
| Top P |
0.90–0.95 |
| Repeat penalty |
1.0 |
| Top K |
20–40 |
LM Studio: disable MTP
Important: If you're running this model in LM Studio, disable multi-token prediction (MTP) for the model to work properly. LM Studio enables this option by default.
System prompt for Three.js scenes
Use only the Three.js core ES module from
https://cdnjs.cloudflare.com/ajax/libs/three.js/0.160.0/three.module.min.js.
Do not import addons or legacy examples/js paths;
implement any camera controls directly inside the HTML.
Example prompts
Use the system prompt above, then try one of these user prompts.
Elephant rehabilitation habitat
Please create one self-contained HTML file with Three.js from cdnjs showing
an elephant rehabilitation habitat.
A broad sheltered yard connects water, shade, and feeding areas through
accessible paths. Use the elephant's scale and movement cl
原文版权归 Hugging Face 热门模型 及原作者所有,本站转载仅供阅读。如需删除,请联系 support@aisa.one。