# Tier-Flow/LoopVL

> LoopVL is a vision-language model that reuses Transformer modules across recurrent computation (L→L→L→H schedule, 128 effective layer applications per forward pass). The repo provides inference code, evaluation on 16 benchmarks, and experiments on visual attention.

- Magnitude: 3.3 out of 10 — Early signal
- Stars: 96 total · +3 stars measured 2026-10-05, 10:59–12:35 UTC, ≈ 16 by evening
- Star trust: star growth looks organic
- Category: Language models · Language: Python · License: Apache-2.0 · Created: 2026-10-01 · Last push: 2026-10-01
- GitHub: https://github.com/Tier-Flow/LoopVL · Page: https://gitnova.dev/en/r/Tier-Flow/LoopVL

## Useful for

- Ask a question about an image via runtime/infer.py with a token budget
- Run the 16-dataset benchmark suite (MMStar, POPE, etc.) on one or more GPUs
- Reproduce paper experiments: attention plots and hidden-state dynamics

## Why it’s here

- Star-counter measurements on 2026-10-05 (UTC), 10:59–12:35: 93 → 96 stars (+3). This is the change over that interval.
- Estimated end-of-day forecast: about +16 stars, using observed gains and the previous day.
- The repository is 4 days old and already has 96 stars. With less than two weeks of history, there's no usual pace to compare the spike against yet.

## Star trust

Star growth looks organic.

Star-trust labels are heuristics based on the repository’s behavior, not a check of every stargazer.

## Numbers

- Forks: 1
- Issues and pull requests: 1
- Watchers: 2
- Average over the last week: 21 per day
- Usual pace: too little history (under two weeks)
- Stars in the last hour (measured): 1
- Latest release: benchmark-assets-v1 (2026-10-01)

## Stars per day, last 9 days (oldest → newest, today is partial)

2026-09-27 … 2026-10-05: 0, 0, 0, 1, 3, 2, 31, 51, 3

## More in this category

1. **Niko1221/Strata** — 8.2 · Breakout · Language models · C++ · +1,613 stars measured 2026-10-05, 00:06–12:33 UTC, ≈ 3,024 by evening
   A C++ local inference engine that runs the 125B MoE model Qwen3.8-Flash-Next on a regular PC with one NVIDIA GPU (12-24 GB) and 64 GB RAM, exposing an OpenAI/Anthropic-compatible API on localhost.
   Full card: https://gitnova.dev/en/r/Niko1221/Strata.md
2. **StayLameBro/backburner** — 6.2 · Early signal · Language models · Python · +160 stars measured 2026-10-05, 00:06–12:33 UTC, ≈ 294 by evening
   A llama.cpp fork that plugs an iPhone into a Mac over USB-C and splits the work: the Mac runs layers 1-40, the iPhone runs 41-64 on its GPU, speeding up prefill and allowing context up to 196k-229k tokens. Aimed at people running…
   Full card: https://gitnova.dev/en/r/StayLameBro/backburner.md
3. **antirez/ds4** — 5.9 · Breakout · Language models · C · +91 stars measured 2026-10-05, 00:06–12:33 UTC, ≈ 196 by evening
   DwarfStar (ds4) is a native C inference engine for running a few specific large LLMs (DeepSeek V4 Flash/PRO, GLM 5.x, Qwen3.8 Flash Next) on Metal, CUDA and ROCm. Targets consumer hardware like MacBooks, DGX Spark and Strix Halo, with SSD…
   Full card: https://gitnova.dev/en/r/antirez/ds4.md
4. **empero-org/brewery-ai** — 5.2 · Early signal · Language models · Python · +24 stars measured 2026-10-05, 07:56–12:33 UTC, ≈ 58 by evening
   A console agent that walks users through the full fine-tuning cycle for language and image models via a chat with an AI guide: from picking a base model and data to training on a GPU and publishing to Hugging Face.
   Full card: https://gitnova.dev/en/r/empero-org/brewery-ai.md
5. **Edge0-AI/Edge0** — 4.9 · Breakout · Language models · Python · +199 stars measured 2026-10-05, 00:07–12:33 UTC, ≈ 344 by evening
   An open-source streaming MoE inference framework: expert weights are offloaded from SSD on demand while a trained prerouter predicts routing ahead of time. Runs on Apple Silicon via MLX and ships with two ready-to-run model tiers (35B and…
   Full card: https://gitnova.dev/en/r/Edge0-AI/Edge0.md

---

Magnitude (0–10) measures how fast and how unusually interest in a repository is growing right now. It is not a quality score.
Days are UTC. “So far today” is a fact; “expected by the end of the day” is a forecast.
Summaries and use cases are written by an LLM (DeepSeek V4.1 Flash) from the README and may be inaccurate: verify specific claims (benchmarks, speed, hardware) in the repository itself.
Data as of 2026-10-05 12:40 UTC, updated every 30 minutes.