A computer, compiled into the weights of a language model. An exact CPU inside Qwen3.5-4B, with Gated DeltaNet state as RAM.
Add ModelScope links next to Hugging Face
Divyansh Agrawal2baf3855fec6
diff --git a/README.md b/README.mdindex b34ee33..c5995e5 100644--- a/README.md+++ b/README.md@@ -2,7 +2,8 @@<img src="docs/qwenputer-banner.svg" alt="Qwenputer: a computer, compiled into the weights of a language model." width="100%"></p>-[Qwenputer model](https://huggingface.co/divagr1925/Qwen3.5-4B-Qwenputer) · [Filesystem model](https://huggingface.co/divagr1925/Qwen3.5-4B-Qwenputer-Filesystem)+Qwenputer model: [Hugging Face](https://huggingface.co/divagr1925/Qwen3.5-4B-Qwenputer) · [ModelScope](https://www.modelscope.ai/models/Spartan1925/Qwen3.5-4B-Qwenputer)+Filesystem model: [Hugging Face](https://huggingface.co/divagr1925/Qwen3.5-4B-Qwenputer-Filesystem) · [ModelScope](https://www.modelscope.ai/models/Spartan1925/Qwen3.5-4B-Qwenputer-Filesystem)**Qwenputer is a new kind of experimental model: Qwen3.5-4B with a 32-bit RISC-V CPU compiled andinserted directly into its weights.** It isn't a plugin, a tool or a sandbox. The CPU is 11@@ -83,7 +84,8 @@ recurrence Qwen3.5 already uses for language.The next logical step was to put a filesystem inside the weights. This is experimental and stillin progress: the model is up as-[`divagr1925/Qwen3.5-4B-Qwenputer-Filesystem`](https://huggingface.co/divagr1925/Qwen3.5-4B-Qwenputer-Filesystem),+[`divagr1925/Qwen3.5-4B-Qwenputer-Filesystem`](https://huggingface.co/divagr1925/Qwen3.5-4B-Qwenputer-Filesystem)+(also on [ModelScope](https://www.modelscope.ai/models/Spartan1925/Qwen3.5-4B-Qwenputer-Filesystem)),and its code lives on the [`rv-store`](https://github.com/divagr18/qwenputer/tree/rv-store/experimental/riscv)branch for now.diff --git a/experimental/riscv/README.md b/experimental/riscv/README.mdindex 54e1a8f..52b61c8 100644--- a/experimental/riscv/README.md+++ b/experimental/riscv/README.md@@ -1,6 +1,6 @@# A RISC-V computer inside Qwen-Model: [divagr1925/Qwen3.5-4B-Qwenputer](https://huggingface.co/divagr1925/Qwen3.5-4B-Qwenputer).+Model: [divagr1925/Qwen3.5-4B-Qwenputer](https://huggingface.co/divagr1925/Qwen3.5-4B-Qwenputer) on Hugging Face, or [Spartan1925/Qwen3.5-4B-Qwenputer](https://www.modelscope.ai/models/Spartan1925/Qwen3.5-4B-Qwenputer) on ModelScope.This folder compiles a 32-bit **RV32IM** CPU into stock Qwen3.5 layers, using the same compiler asthe main project. The CPU also has Zba and a bignum coprocessor. It is a self-contained snapshot: