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
rename HF repo id to Edge0/Edge0-35B-A3B-preview (capital B/A)
linyubupafc51cf43d917
diff --git a/README.md b/README.mdindex bdca4d2..39dd2b5 100644--- a/README.md+++ b/README.md@@ -6,7 +6,7 @@**An open-source streaming MoE inference framework — SSD expert offload + parallel LoRA + prerouter routing prediction.**-[](https://huggingface.co/Edge0/Edge0-35b-a3b-preview)+[](https://huggingface.co/Edge0/Edge0-35B-A3B-preview)[](https://huggingface.co/Edge0/Edge0-8B-A1B-preview)[](https://github.com/Edge0-AI/edge0)[](LICENSE)@@ -28,7 +28,7 @@ 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 |+| `edge0-35b` | [`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 |Both checkpoints are built on open sparse-MoE base models (Qwen3.5-MoE@@ -106,7 +106,7 @@ The two tiers are published on Hugging Face — 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/Edge0-35B-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)```bash@@ -115,7 +115,7 @@ directory**, so a single download is a ready-to-run model:.venv/bin/python scripts/fetch_models.py --tier edge0-8b --target-dir models# or directly with the CLI:-.venv/bin/huggingface-cli download Edge0/Edge0-35b-a3b-preview --local-dir models/edge0-35b+.venv/bin/huggingface-cli download Edge0/Edge0-35B-A3B-preview --local-dir models/edge0-35b.venv/bin/huggingface-cli download Edge0/Edge0-8B-A1B-preview --local-dir models/edge0-8b```diff --git a/README_zh.md b/README_zh.mdindex 8b3fe1a..b1f8232 100644--- a/README_zh.md+++ b/README_zh.md@@ -6,7 +6,7 @@**开源流式 MoE 推理框架 —— SSD 专家 offload + 并行 LoRA + prerouter 路由预判**-[](https://huggingface.co/Edge0/Edge0-35b-a3b-preview)+[](https://huggingface.co/Edge0/Edge0-35B-A3B-preview)[](https://huggingface.co/Edge0/Edge0-8B-A1B-preview)[](https://github.com/Edge0-AI/edge0)[](LICENSE)@@ -24,7 +24,7 @@| 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 |+| `edge0-35b` | [`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 |两个 checkpoint 均基于开源稀疏 MoE 基座(分别为 Qwen3.5-MoE 35B-A3B@@ -87,7 +87,7 @@ python3.12 -m venv .venv && .venv/bin/pip install -e '.[dev,fetch]'两个档位发布在 Hugging Face——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/Edge0-35B-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)```bash@@ -96,7 +96,7 @@ LoRA + prerouter 适配器打包在**同一目录**,一次下载即为可运.venv/bin/python scripts/fetch_models.py --tier edge0-8b --target-dir models# 或直接用 CLI:-.venv/bin/huggingface-cli download Edge0/Edge0-35b-a3b-preview \+.venv/bin/huggingface-cli download Edge0/Edge0-35B-A3B-preview \--local-dir models/edge0-35b.venv/bin/huggingface-cli download Edge0/Edge0-8B-A1B-preview \--local-dir models/edge0-8bdiff --git a/scripts/fetch_models.py b/scripts/fetch_models.pyindex fa2377b..67d1302 100755--- a/scripts/fetch_models.py+++ b/scripts/fetch_models.py@@ -12,7 +12,7 @@ Requirements:Repo ids default to the official releases (override via the environmentif you mirror them):- export EDGE0_35B_REPO=Edge0/Edge0-35b-a3b-preview # default+ export EDGE0_35B_REPO=Edge0/Edge0-35B-A3B-preview # defaultexport EDGE0_8B_REPO=Edge0/Edge0-8B-A1B-preview # defaultUsage:@@ -33,7 +33,7 @@ import sysfrom pathlib import PathTIER_REPOS = {- "edge0-35b": ("EDGE0_35B_REPO", "Edge0/Edge0-35b-a3b-preview"),+ "edge0-35b": ("EDGE0_35B_REPO", "Edge0/Edge0-35B-A3B-preview"),"edge0-8b": ("EDGE0_8B_REPO", "Edge0/Edge0-8B-A1B-preview"),}diff --git a/scripts/upload_hf.py b/scripts/upload_hf.pyindex 0355683..56f10e4 100644--- a/scripts/upload_hf.py+++ b/scripts/upload_hf.py@@ -72,7 +72,7 @@ HfApi._validate_yaml = lambda self, content, repo_type, token=None: NoneREPOS = {"edge0-8b": ("EDGE0_8B_MODEL", "Edge0/Edge0-8B-A1B-preview"),- "edge0-35b": ("EDGE0_35B_MODEL", "Edge0/Edge0-35b-a3b-preview"),+ "edge0-35b": ("EDGE0_35B_MODEL", "Edge0/Edge0-35B-A3B-preview"),}