Article 10 "Anemll-style" Root-Mean-Square (RMS) Normalization on the Apple Neural Engine: A Simple Hack
ANEMLL-0.3.4 Models build with 0.3.4, improved quality and bug fixes anemll/anemll-Qwen-Qwen3-0.6B-ctx512_0.3.4 Updated Jul 7 • 15 anemll/anemll-meta-llama-Llama-3.2-1B-Instruct-ctx1024_0.3.4 Updated Jul 3 • 16 anemll/anemll-Qwen-Qwen3-0.6B-LUT888-ctx512_0.3.4 Updated Jul 7 • 45
Qwen3 for ANE Initial Support for QWEN3 anemll/anemll-Qwen3-4B-ctx1024_0.3.0 Updated Jun 20 • 48 • 2 anemll/anemll-Qwen3-0.6B-ctx512_0.3.0 Updated Jun 20 • 24
ANEMLL-0.3.4 Models build with 0.3.4, improved quality and bug fixes anemll/anemll-Qwen-Qwen3-0.6B-ctx512_0.3.4 Updated Jul 7 • 15 anemll/anemll-meta-llama-Llama-3.2-1B-Instruct-ctx1024_0.3.4 Updated Jul 3 • 16 anemll/anemll-Qwen-Qwen3-0.6B-LUT888-ctx512_0.3.4 Updated Jul 7 • 45
Qwen3 for ANE Initial Support for QWEN3 anemll/anemll-Qwen3-4B-ctx1024_0.3.0 Updated Jun 20 • 48 • 2 anemll/anemll-Qwen3-0.6B-ctx512_0.3.0 Updated Jun 20 • 24
Runtime error 3 On-Device LLM Throughput Calculator 🚀 Generate throughput plots for LLMs on devices