Model	techniques
DeepSeek-V4.1-Flash used	Batch-level embedding prefetch usedOffline checkpoint merging used—
NVIDIA-Nemotron-3-Ultra-550B-A55B used	Asymmetric local-read and single-writer cache paths usedAsynchronous checkpointing usedCaching the distributed checkpoint save plan usedIn-flight weight updates usedOffline checkpoint merging usedOverlapping NCCL transfers with device-to-host copies usedPersistent checkpoint worker processes usedPersistent shared-storage cache for compiled artifacts usedSeeding node-local storage from a warm shared cache used—
MiMo-V2.6-Flash used	CPU-resident optimizer states used—
DeepSeek-V4-Flash core	Tensor-level automatic-differentiation activation checkpointing coreMinimal recomputation-graph extraction by backward traversal usedSelective tensor recomputation for activation-memory reduction used—
GLM-5.2 used	Heartbeat-driven fault tolerance used—
MiniMax-M3 used	Autotune configuration generation used—
DeepSeek-V4-Pro core	Tensor-level automatic-differentiation activation checkpointing coreMinimal recomputation-graph extraction by backward traversal usedSelective tensor recomputation for activation-memory reduction used—
Gemma 4 31B used	Slice-Granularity Elasticity used—
Kimi K3 core	Composable activation storage policies coreUnified pluggable activation storage manager coreCross-rank remote activation offloading usedDouble-buffered chunked streaming of reference-model weights usedElement-wise activation recomputation usedPolicy-model gradient-buffer reuse for reference-model weights usedPost-iteration NVMe offloading of training states used—
Laguna-S-2.1 core	In-flight recovery system coreGPU-to-GPU weight synchronization over GPUDirect RDMA usedJob-level eviction and reclaim usedPre-admission hardware stress testing usedSame-node sticky pod respawn used—
MiMo-V2.6-Pro used	CPU-resident optimizer states used—

training runtime

In this branch 29

filed here 29

By model