MiaAI-Lab/Qwen3.8-Flash-Next-Single-DGX-Spark-TensorFold
Qwen3.8 Flash Next on one DGX Spark (TensorFold)
Older snapshot: check the original source for current information.
From the publisher
README & documentation
Source previewRead the project’s overview, installation instructions and usage examples. The original README is the source of truth.
Qwen3.8 Flash Next on one DGX Spark (TensorFold) by Mia's AI Lab Serve Qwen3.8 Flash Next from a single NVIDIA DGX Spark (GB10, 128 GB) through an OpenAI-compatible API. It runs TensorFold's Zig engine, tensorfold-native (release v1.1.4, commit eab85828, with this recipe's patches on top), in NVIDIA's PyTorch container. One GPU, with image and video input (VISION=1). VISION=0 serves text only. The default checkpoint is NVIDIA's FP4 QAD export: routed experts in NVFP4 (group size 16, CSF-compressed scale factors), attention and shared-expert linears in MXFP8, the MTP experts and the vision tower quantized too (MXFP8 attention, W4A16 NVFP4 MLP), an FP8 n-gram table, and all 512 routed experts with 10 per token (the INT4 cut keeps 5). The INT4-AutoRound checkpoint (served by the db281878 recipe on main, not by this v1.1.4 branch) uses 5 experts per token (4.8B active parameters) and scores about 10% lower general capability on its authors'…
Read the full README ↗ · Preview checked 2026-11-15T15:34:58.173Z
Inside the original README — Document outline
- Performance
- Checks on the current build
- Requirements
- Quick start
- Choosing the checkpoint
- ABLIT (opt-in)
- Images and video
- What start.sh and scripts/prepare.sh do
- KV and memory
- Kernel NVERRNOMEMORY messages under load (issue #33)
- Configuration
- Thinking and sampling
Headings are captured from the source. Links open the publisher’s document, not a locally hosted copy.
Links from the README
References supplied by the publisher, not independently verified endorsements. Check the destination before downloading files or entering credentials.
Documentation belongs to its respective authors. Reported project/model license: Other. A listing is not a grant of reuse or training rights. Confirm the document’s own terms at the source.
What this repository does
Qwen3.8 Flash Next on one DGX Spark (TensorFold)
Repository facts
- Owner
- MiaAI-Lab
- Primary language
- Shell
- Stars
- 178
- Forks
- 27
- Open issues + pull requests
- 13
- License
- Other
- Archived
- No
- Default branch
- main
- Created
- 2026-09-29T04:51:21.000Z
- Last push
- 2026-10-02T10:12:45.000Z
Topics and intended use
No topics were included in the latest source metadata.
Review the README for scope, installation, examples and limitations. We do not run repository code or certify it.
Evaluate before installing
Review licensing and dependencies, inspect recent commits and unresolved issues, and test in an isolated environment before production use. Stars and forks alone cannot answer these questions.
Source and freshness
Source: GitHub. Metadata observed 2026-10-05T12:05:52.305Z. Daily imports are snapshots, not real-time monitoring.
Popularity and source listings do not establish security, suitability, licensing rights or benchmark performance.