Commit History
29 days ago
-
refactor: consolidate pure 32L Dense architecture, purge MoE files, audit hardening
Verified signature3a5d3e5 -
Audit fix: Update README.md for ViuMini-Dense-360M
Verified signature1d7d841 -
Update README.md for ViuMini-Dense-360M 32L architecture
Verified signature333904e
-
Fix CUDA OOM on T4: cast model to amp_dtype directly, batch_size=2, safe RMSNorm
Verified signature9d0f9ca
-
sync local: enforce_mix sampler, tokenizer v2 (mask fix + 35/30/35), mix optional docs, repair scripts
Verified signature991af26
-
init: 28L MoE 241M audit-fixed, smoke pass
Verified signature140f9f6