mkleinegger
/

apex

mixture-of-experts

Model card Files Files and versions

658 MB

Ctrl+K

Ctrl+K

1 contributor

History: 4 commits

mkleinegger's picture

Add v4 transfer matrices (ridge only, no intercept) for Qwen3.8-Flash-Next-FP8 and Qwen3.6-35B-A3B: final calibration mix incl. thinking-off code set

2e9c72a verified 9 days ago

deepseek-v4.1-flash
Add APEX-MoE ridge transfer priors for Qwen3.6-35B-A3B, Qwen3.8-Flash-Next-FP8, DeepSeek-V4.1-Flash 15 days ago
qwen3.6-35b-a3b
Add APEX-MoE ridge transfer priors for Qwen3.6-35B-A3B, Qwen3.8-Flash-Next-FP8, DeepSeek-V4.1-Flash 15 days ago
qwen3.8-flash-next-fp8
Add APEX-MoE ridge transfer priors for Qwen3.6-35B-A3B, Qwen3.8-Flash-Next-FP8, DeepSeek-V4.1-Flash 15 days ago
v3
Add v3 ridge-only priors (Qwen3.6, Qwen3.8) with window ridges for dynamic re-selection 11 days ago
v4
Add v4 transfer matrices (ridge only, no intercept) for Qwen3.8-Flash-Next-FP8 and Qwen3.6-35B-A3B: final calibration mix incl. thinking-off code set 9 days ago
.gitattributes

1.52 kB
initial commit 15 days ago
README.md

6.58 kB
Add v4 transfer matrices (ridge only, no intercept) for Qwen3.8-Flash-Next-FP8 and Qwen3.6-35B-A3B: final calibration mix incl. thinking-off code set 9 days ago