CogniTech Systems

Pinned Loading

1. qwen3.8-claude-code qwen3.8-claude-code Public

   Alibaba's Qwen3.8-27B running fully local on Apple Silicon: llama.cpp setup, measured benchmarks, speculative decoding, and a Claude Code bridge via LiteLLM.

   Python 1 1

2. muse-glimmer-claude-code muse-glimmer-claude-code Public

   Meta's Muse-Glimmer-30B running fully local on Apple Silicon: llama.cpp setup, measured benchmarks, speculative decoding, and a Claude Code bridge via LiteLLM.

   Python 9 1

3. moe-on-arc moe-on-arc Public

   Benchmarking Qwen3.6 MoE on Intel integrated graphics. Measuring accuracy, not just throughput.

   Python 2

4. EdgeImpulse EdgeImpulse Public

   A repository of Edge Impulse examples, using the Edge Impulse platform to create intelligent IoT devices.

   1