ivo's home

StateBench is a Hugging Face repository that ships with over 10,000 synthetic documents up to 200k tokens, designed to test how language models track entity state across long contexts. It defines four difficulty levels—from r‑trivial to general finite—and provides pre‑trained results for transformers and hybrid Gated DeltaNet models, all runnable on a single RTX 3090 or any CPU. The benchmark replaces ad‑hoc unit tests that only check the last update by offering a systematic, reproducible evaluation pipeline.

If you need concrete, level‑based numbers to compare how different architectures handle long‑range state changes, StateBench is worth a look. It is limited to synthetic patterns and doesn’t cover real‑world discourse or multimodal scenarios, so it may not be the final test for production deployments.

more: https://huggingface.co/blog/allenai/statebench

my other side: ivo.blogspot.com

Diffusers is a pip‑installable library that bundles state‑of‑the‑art pretrained diffusion models for images, audio and 3D molecules. It provides a handful of pipelines that run on CUDA GPUs or Apple Silicon via MPS, and it includes over 30,000 checkpoints from the Hugging Face Hub. The library is designed as a drop‑in replacement for custom diffusion codebases that you would otherwise have to build from scratch.

If you need to prototype or fine‑tune diffusion models quickly, Diffusers is worth a look. The API is straightforward but it sacrifices raw performance for ease of use, and it may not be suitable for production workloads that demand the absolute lowest latency or integration with non‑PyTorch backends.

more: https://github.com/huggingface/diffusers

my other side: ivo.blogspot.com

A Node 24 CLI tool that runs coding agents on a local schedule or event trigger. It keeps the agent process between runs, waking only to handle new tasks or send approval requests via Slack or Feishu. Configuration is a simple JSON file mapping events to agents, requiring no cloud infrastructure or public ports.

It is a practical solution for developers who want hands-off automation on a single machine. The main limitation is the lack of Windows support, which is currently in development. It is not suitable for teams needing a centralized dashboard, as it relies entirely on local file configuration and command-line status checks.

more: https://github.com/KKKKhazix/agent-autoloop

my other side: ivo.blogspot.com

This client runs the DeepSeek Harness engine on Android 8.0+ via PRoot without needing Root or Termux. It can execute locally on your phone or connect remotely to a server via SSH. The interface supports chat, tool approval, and mobile screen control in both native and web modes.

Useful if you want a dedicated mobile frontend for DSH without managing a separate server. The main catch is that local execution requires about 1 GB of storage and the APK is not on the Play Store. It is not for users expecting offline model inference, as requests still go to your configured provider.

more: https://github.com/ccch1mneyyy/dsh-Go

my other side: ivo.blogspot.com