MiaAI-Lab/DeepSeek-v4.1-Flash-1017-2x-DGX-Sparks
DeepSeek-V4.1-Flash-1017 EXL3 on 2x DGX Sparks with 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.
DeepSeek-V4.1-Flash-1017 EXL3 on 2x DGX Sparks with TensorFold by Mia's AI Lab Serve DeepSeek-V4.1-Flash-1017 from two NVIDIA DGX Sparks (GB10, 128 GB each, linked by their ConnectX-7 ports) through an OpenAI-compatible API, with the full 1,048,576-token window, image input, tool calling, thinking on or off, and DSpark speculative decoding from the checkpoint's own draft blocks, using TensorFold's native Zig engine and the 2.9 bit EXL3 quantization Mia-AiLab/DeepSeek-V4.1-Flash-1017-EXL3-EAQ-v1-2.9-bpw (196 GiB; its model card has no perplexity or KL measurement against the unquantized model). EAQ v1 is the recommended pack; the experimental EAQ v1.1 pack is opt-in with EAQV11=1. weight, 2 bits in layers 18-22, attention at 5, shared experts at 4-5, head at 6, DSpark blocks at 4; the vision tower and the DSpark blocks are in it; the revision is pinned in scripts/config.sh), plus Engram's two tables, unquantized, from deepseek-ai/DeepSeek-V4.1-Flash-1017 @ 1bae0e10 (shards 47 and 48, 203 GB; the quantization…
Read the full README ↗ · Preview checked 2026-10-28T20:21:39.780Z
Inside the original README — Document outline
- Status
- Requirements
- Quick start
- Abliterated mode
- What start.sh and scripts/prepare.sh do
- Configuration
- Spill v2 and KV sharding (ported, off by default)
- Memory and safety
- Repository layout
- License
- Credits
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: Apache-2.0. A listing is not a grant of reuse or training rights. Confirm the document’s own terms at the source.
What this repository does
DeepSeek-V4.1-Flash-1017 EXL3 on 2x DGX Sparks with TensorFold
Repository facts
- Owner
- MiaAI-Lab
- Primary language
- Python
- Stars
- 289
- Forks
- 47
- Open issues + pull requests
- 23
- License
- Apache-2.0
- Archived
- No
- Default branch
- main
- Created
- 2026-10-18T18:10:59.000Z
- Last push
- 2026-10-24T14:43:47.000Z
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 2026-10-25T12:13:46.029Z. Daily imports are snapshots, not real-time monitoring.
Popularity and source listings do not establish security, suitability, licensing rights or benchmark performance.
Related repositories
EbookFoundation/free-programming-books
donnemartin/system-design-primer