|
Download README.md from hiru2005/Sync-S: direct link, hf CLI and curl.
- Browser
- Download file 6.2 kB
-
https://huggingface.co/hiru2005/Sync-S/resolve/main/README.md
- Command line
-
hf download hf://hiru2005/Sync-S/README.md
-
curl -L -o README.md https://huggingface.co/hiru2005/Sync-S/resolve/main/README.md
6.2 kB
| license: other | |
| license_name: qwen-community-1.0 | |
| license_link: LICENSE | |
| base_model: | |
| - Qwen/Qwen3.8-Flash-Next | |
| base_model_relation: quantized | |
| tags: | |
| - qwen3.8-flash-next | |
| - moe | |
| - sync-strata | |
| - experimental | |
| # Qwen3.8-Flash-Next Sync-S (experimental) | |
| A re-quantized expert pack of **Qwen3.8-Flash-Next** for one consumer NVIDIA GPU, made to run faster than upstream | |
| Strata's Gyro. It runs only on **[Sync-Strata](https://github.com/hiruhiru2005/sync-strata)** (a fork of | |
| [Strata](https://github.com/Niko1221/Strata)). **This is not a GGUF: it does not load in llama.cpp or upstream Strata.** | |
| Recommended for: **PCs with 32 GB of RAM**. | |
| ## What was done | |
| - The routed experts were re-quantized to a trellis format of the Gyro / QTIP family, with the bit width chosen per | |
| expert. The pack also carries the dense weights that the engine reads from it. | |
| - The remaining weights (embeddings, the n-gram/PLE table and so on) are not in this repository: they are in the | |
| shared parts, [hiru2005/Sync-shared](https://huggingface.co/hiru2005/Sync-shared), which every Sync model uses. | |
| ## Files | |
| Total: **28.14 GB (26.21 GiB)**, plus the shared parts. | |
| | File | Size | What | | |
| |---|---:|---| | |
| | `MANIFEST.json` | 1,835 B | Index of the pack (file names, sizes, checksums) | | |
| | `compat-bf16.json` | 36,116 B | Layout and conversion tables read by the engine | | |
| | `conversions.json` | 151,849 B | Layout and conversion tables read by the engine | | |
| | `dense_nq.txt` | 2,013 B | Layout and conversion tables read by the engine | | |
| | `experts_k.txt` | 358,331 B | Layout and conversion tables read by the engine | | |
| | `hadamard.txt` | 8,938 B | Layout and conversion tables read by the engine | | |
| | `index.txt` | 93,669 B | Layout and conversion tables read by the engine | | |
| | `native_experts.txt` | 3,543 B | Layout and conversion tables read by the engine | | |
| | `r1q2.txt` | 3,818 B | Layout and conversion tables read by the engine | | |
| | `dense_nq.bin` | 755,548,160 B | Dense weights read by the engine from the pack (second part) | | |
| | `dense.bin` | 1,538,035,200 B | Dense weights read by the engine from the pack | | |
| | `experts.v2.bin` | 25,846,112,256 B | Routed experts (re-quantized) | | |
| | `LICENSE` | | Qwen Community License 1.0 | | |
| ## Requirements | |
| - Windows 10/11 or 64-bit Linux. Supported GPUs: NVIDIA RTX 30/40 series (driver R580 or newer); RTX 20, GTX 16, RTX 50 are expected to work, and GTX 10, Radeon RX 6000/7000/9000 and the integrated Radeon of Ryzen AI Max (Strix Halo: 8060S, 8050S, 8040S; Windows and Linux) are experimental, all not checked on real hardware. Other integrated GPUs and Intel GPUs are not supported ([the GPU table](https://github.com/hiruhiru2005/sync-strata#supported-gpus)). A CPU with AVX2. | |
| - RAM and VRAM: see "Recommended for" above. The app picks the model by the RAM installed in the PC; if the free RAM is short it warns you instead of choosing a smaller model. The minimum amounts of RAM and VRAM are to be measured. | |
| - Disk: this pack plus the shared parts (`native-gguf`, `mtp-rt`, `tokenizer`, `mmproj` (optional)). An NVMe SSD is recommended. | |
| ## Use with Sync-Strata | |
| The Sync-Strata app downloads this model for you: open the app and pick **Sync-S** on the first screen (or let | |
| "auto" choose). To download by hand, put the files in a folder named `sync-s-sb` inside the app's models folder (`models_dir` in `settings.json`), together with the shared parts. | |
| ``` | |
| sync settings set model sync-s | |
| sync up | |
| ``` | |
| ## Quality | |
| Measured on Colab with the production eval_v4 (8 domains, 1,400 windows; the teacher is the Q8_0 GGUF of the original model). | |
| | | Sync-S | | |
| |---|---:| | |
| | KL to the original model (all positions, lower is better) | 0.2310 | | |
| This value was measured on the earlier form of Sync-S (table-lookup points, the same dense part). The form in this repository (no table lookup) has not been measured yet; for Sync-XS and Sync-M the two forms were on par, within the measurement noise. | |
| For comparisons with other models (upstream Gyro-S, Gyro-M, IQ3_XXS, Sync-XS), see the [Sync-XS page](https://huggingface.co/hiru2005/Sync-XS). | |
| ## Speed | |
| Measured on the author's PC with Sync-Strata 0.1.0 (Ryzen 7 7800X3D, 32 GB DDR5, RTX 3080 12 GB, NVMe SSD; Windows 11 and Linux). The values in the table are for Windows; the Linux values are to be measured. No numbers for GTX 10 or Radeon: no hardware to measure on. | |
| | | tail skip | jbench decode (tok/s) | dsh chat (tok/s) | | |
| |---|---|---:|---:| | |
| | Sync-S | R4 | 144.0 | 104.1 | | |
| | Sync-S | R7 | 127.5 | 90.7 | | |
| | Sync-S | R10 | 104.6 | 80.1 | | |
| Speed depends strongly on the PC. The values of the other models, including upstream Gyro, are in the table of the [Sync-Strata README](https://github.com/hiruhiru2005/sync-strata#benchmarks). | |
| ## Calibration | |
| The quantization was calibrated on public text (code, documentation, encyclopedia articles, several languages, agent | |
| traces) and on some text made from the author's own chat logs. The calibration text is not distributed. | |
| ## Credits | |
| - Base model: [Qwen/Qwen3.8-Flash-Next](https://huggingface.co/Qwen/Qwen3.8-Flash-Next) by the Qwen team. | |
| - [agentionai](https://huggingface.co/agentionai): the Gyro trellis format and kernels, and the | |
| [Gyro-S GGUF](https://huggingface.co/agentionai/Qwen3.8-Flash-Next-Gyro-GGUF) that the shared parts are cut from. | |
| - [Strata](https://github.com/Niko1221/Strata) by Niko1221 and the Strata contributors (MIT), which Sync-Strata forks. | |
| - QTIP (Tseng et al., 2024); [llama.cpp / ggml](https://github.com/ggml-org/llama.cpp) (MIT). | |
| ## Support (optional) | |
| Sync-Strata and the Sync models are free. A donation page is at [Ko-fi](https://ko-fi.com/hiruhiru2005): a thank-you only, with nothing in return (no perks, no priority support, no feature promises), and it does not change the license terms. | |
| ## License | |
| These files are derived from Qwen3.8-Flash-Next and are under the **Qwen Community License 1.0** (`LICENSE`, | |
| Copyright (c) 2026 Qwen). Keep `LICENSE` with every copy. Commercial products with more than 100 million monthly active | |
| users or US$20 million monthly revenue must show the model name, and a Model-as-a-Service or AI work assistant business | |
| needs a separate license from Qwen for commercial use. Read the full text. | |