Model-Optimizer

mirror of https://github.com/NVIDIA/Model-Optimizer.git synced 2026-10-02 03:14:52 +08:00

Files

T

yeyu-nvidia 651fd223e6 feat: EAGLE3 LoRA co-training improvements (#1607 )

Layer-selective LoRA injection for EAGLE3 co-training, optimizer-stable
warmup (LoRA always in the optimizer; warmup gated by a flag), and an
export+merge+lm_eval evaluation script.

Review feedback addressed:
- trust_remote_code is caller-controlled (TRUST_REMOTE_CODE / --trust_remote_code), default False
- eagle_base_lora_start_layer raises ValueError instead of silently injecting zero adapters
- eval_lora.sh validates HF_MODEL_CKPT / EAGLE_CKPT up front
- added unit tests for start-layer injection and validation (8/8 passing, incl. on-cluster GPU run)

Signed-off-by: Ye Yu <yeyu@nvidia.com>

2026-06-03 17:33:12 +00:00

ar_validate.py

[1/3][Refactor]: File reorg; deprecate ParallelDraft (#1296 )

2026-04-24 14:46:10 -07:00

calibrate_draft_vocab.py

add: DFlash block diffusion speculative decoding (#1211 )

2026-04-14 15:58:39 -07:00

convert_to_vllm_ckpt.py

Feat: Support VLLM one-model eagle ckpt; Add unit tests; (#573 )

2025-11-19 14:26:39 -08:00

eval_lora.sh

feat: EAGLE3 LoRA co-training improvements (#1607 )