Add Trinity model family (AfmoeForCausalLM) contrib#55

Open

jimburtoft wants to merge 4 commits intoaws-neuron:mainfrom

jimburtoft:contrib/trinity-model

Commits on Feb 27, 2026

Add Trinity model family (AfmoeForCausalLM) contrib implementation
jimburtoft
committed

Commits on Feb 28, 2026

Fix mixed attention KV cache sizing and add max sequence length results
jimburtoft
committed

Commits on Mar 2, 2026

Add bucketing support with TrinityKVCacheManager for mixed attention
jimburtoft
committed
Add validated bucketing results for Mini and Large
jimburtoft
committed