SergiioB / intel-arc-pro-b70-inference-cookbook
Pinned Intel Arc Pro B60/B70 inference recipes, engine patches and measurement records: Qwen3.8-27B EXL3 on one card (262,144 context, 63 tok/s; MTP3 measured lossy on 27% of a fixed 329-item set), Nemotron Bolt 74B, Flash-Next, dual-card routes, and a machine-checked catalog where each speculative row states whether the drafter changes the answer.
Why this job: says: cookbook
What it does
Pinned vLLM XPU, llama.cpp SYCL and EXL3-XPU recipes, failure notes, and measurement records for Intel Arc Pro B60/B70. Every published benchmark row carries its exact workload, a commit-pinned evidence link, and — for routes that decode with a drafter — whether the drafter changes the model's greedy answer. Start with what worked, what did not, and what to test next. For a first local run, use START.md.
Get started
No single install command. Read the steps in the README (opens in new tab)
- Maintenance
- Active
- Last commit yesterday
- Stars
- 195
- +4 this week
- Licence
- MIT
- Allows business use
- Downloads for
- None listed
- No release files
- Latest version
- v1.11.0
- Released 4 days ago
- Built with
- Python
- 66% of the code
Is it looked after?
- Contributors
- 2
- Last 90 days
- Issues
- 83% closed
- 2 still open
- Pull requests
- 57 merged
- 0 waiting
- Releases
- 23
- 23 in the last year
- No contributing guide
- No security policy
- No code of conduct
More: languages, topics and project record
Languages
- Python 66%
- C++ 23%
- Shell 5%
- HTML 3%
Topics
- battlemage
- intel-arc
- intel-arc-b70
- intel-gpu
- llama-cpp
- llm-inference
- local-ai
- moe
- mtp
- speculative-decoding
- sycl
- vllm
Project record
- JobLists and guides
- StatusListed
- Commits395
- Forks13
- Watching12
- Stars in 28 days+19
- CreatedAug 6, 2026
- First seen by usNov 7, 2026
- CheckedNov 9, 2026