Run 35B MoE models in ~2.5 GB of RAM. Weights stream from SSD; a trained router predicts loads a step ahead. Mac, iPhone, Android, Windows supported
Python
docs(readme): add the ModelScope mirrors next to the Hugging Face links
linyubupaf8c4cb2c1871
diff --git a/README.md b/README.mdindex 20c587e..4e1ed13 100644--- a/README.md+++ b/README.md@@ -8,6 +8,8 @@[](https://huggingface.co/Edge0/Edge0-35B-A3B-preview)[](https://huggingface.co/Edge0/Edge0-8B-A1B-preview)+[](https://www.modelscope.cn/models/Edge0/Edge0-35B-A3B-preview)+[](https://www.modelscope.cn/models/Edge0/Edge0-8B-A1B-preview)[](https://github.com/Edge0-AI/Edge0)[](LICENSE)@@ -28,8 +30,8 @@ trained prerouter heads work together as one unit.| Tier | Released checkpoint | Inference profile ||---|---|---|-| `edge0-35b` | [`Edge0/Edge0-35B-A3B-preview`](https://huggingface.co/Edge0/Edge0-35B-A3B-preview) | 4-bit, 40 layers, 256 experts, prerouter K=4 |-| `edge0-8b` | [`Edge0/Edge0-8B-A1B-preview`](https://huggingface.co/Edge0/Edge0-8B-A1B-preview) | 4-bit, 24 layers, 128 experts, prerouter K=8 |+| `edge0-35b` | [`Edge0/Edge0-35B-A3B-preview`](https://huggingface.co/Edge0/Edge0-35B-A3B-preview) · [ModelScope](https://www.modelscope.cn/models/Edge0/Edge0-35B-A3B-preview) | 4-bit, 40 layers, 256 experts, prerouter K=4 |+| `edge0-8b` | [`Edge0/Edge0-8B-A1B-preview`](https://huggingface.co/Edge0/Edge0-8B-A1B-preview) · [ModelScope](https://www.modelscope.cn/models/Edge0/Edge0-8B-A1B-preview) | 4-bit, 24 layers, 128 experts, prerouter K=8 |Both checkpoints are built on open sparse-MoE base models (Qwen3.6-35B-A3Band the Ling 3.0 bailing hybrid respectively) and ship with the@@ -96,12 +98,12 @@ python3.12 -m venv .venv && .venv/bin/pip install -e '.[dev,fetch]'### 2) Download a model-The two tiers are published on Hugging Face — each repo bundles the+The two tiers are published on Hugging Face and ModelScope — each repo bundles thebase checkpoint and the trained LoRA + prerouter adapters in **onedirectory**, so a single download is a ready-to-run model:-- [`Edge0/Edge0-35B-A3B-preview`](https://huggingface.co/Edge0/Edge0-35B-A3B-preview) (~23 GB)-- [`Edge0/Edge0-8B-A1B-preview`](https://huggingface.co/Edge0/Edge0-8B-A1B-preview) (~4.2 GB)+- [`Edge0/Edge0-35B-A3B-preview`](https://huggingface.co/Edge0/Edge0-35B-A3B-preview) (~23 GB) · [ModelScope mirror](https://www.modelscope.cn/models/Edge0/Edge0-35B-A3B-preview)+- [`Edge0/Edge0-8B-A1B-preview`](https://huggingface.co/Edge0/Edge0-8B-A1B-preview) (~4.2 GB) · [ModelScope mirror](https://www.modelscope.cn/models/Edge0/Edge0-8B-A1B-preview)```bash# with the repo's helper (defaults to the two repos above):diff --git a/README_zh.md b/README_zh.mdindex 74d2af7..e946e82 100644--- a/README_zh.md+++ b/README_zh.md@@ -8,6 +8,8 @@[](https://huggingface.co/Edge0/Edge0-35B-A3B-preview)[](https://huggingface.co/Edge0/Edge0-8B-A1B-preview)+[](https://www.modelscope.cn/models/Edge0/Edge0-35B-A3B-preview)+[](https://www.modelscope.cn/models/Edge0/Edge0-8B-A1B-preview)[](https://github.com/Edge0-AI/Edge0)[](LICENSE)@@ -24,8 +26,8 @@ Recover-LoRA + prerouter 路由预判」抽象成可扩展的通用框架。后| 档位 | 发布 checkpoint | 推理档 ||---|---|---|-| `edge0-35b` | [`Edge0/Edge0-35B-A3B-preview`](https://huggingface.co/Edge0/Edge0-35B-A3B-preview) | 4bit,40 层,256 专家,prerouter K=4 |-| `edge0-8b` | [`Edge0/Edge0-8B-A1B-preview`](https://huggingface.co/Edge0/Edge0-8B-A1B-preview) | 4bit,24 层,128 专家,prerouter K=8 |+| `edge0-35b` | [`Edge0/Edge0-35B-A3B-preview`](https://huggingface.co/Edge0/Edge0-35B-A3B-preview) · [ModelScope](https://www.modelscope.cn/models/Edge0/Edge0-35B-A3B-preview) | 4bit,40 层,256 专家,prerouter K=4 |+| `edge0-8b` | [`Edge0/Edge0-8B-A1B-preview`](https://huggingface.co/Edge0/Edge0-8B-A1B-preview) · [ModelScope](https://www.modelscope.cn/models/Edge0/Edge0-8B-A1B-preview) | 4bit,24 层,128 专家,prerouter K=8 |两个 checkpoint 均基于开源稀疏 MoE 基座(分别为 Qwen3.6-35B-A3B与 Ling 3.0 混合架构),并携带为本框架训练的 LoRA 与 prerouter 权重——@@ -84,11 +86,11 @@ python3.12 -m venv .venv && .venv/bin/pip install -e '.[dev,fetch]'### 2) 下载模型-两个档位发布在 Hugging Face——每个仓库把基模 checkpoint 与训练好的+两个档位发布在 Hugging Face 与 ModelScope——每个仓库把基模 checkpoint 与训练好的LoRA + prerouter 适配器打包在**同一目录**,一次下载即为可运行的模型:-- [`Edge0/Edge0-35B-A3B-preview`](https://huggingface.co/Edge0/Edge0-35B-A3B-preview)(约 23 GB)-- [`Edge0/Edge0-8B-A1B-preview`](https://huggingface.co/Edge0/Edge0-8B-A1B-preview)(约 4.2 GB)+- [`Edge0/Edge0-35B-A3B-preview`](https://huggingface.co/Edge0/Edge0-35B-A3B-preview)(约 23 GB) · [ModelScope 镜像](https://www.modelscope.cn/models/Edge0/Edge0-35B-A3B-preview)+- [`Edge0/Edge0-8B-A1B-preview`](https://huggingface.co/Edge0/Edge0-8B-A1B-preview)(约 4.2 GB) · [ModelScope 镜像](https://www.modelscope.cn/models/Edge0/Edge0-8B-A1B-preview)```bash# 用仓库自带脚本(默认即上述两个仓库):