Commit 1e66b72

committed

Eagle: MLA Based Eagle

Signed-off-by: Izzy Putterman <iputterman@nvidia.com>

1 parent 74832a1 commit 1e66b72Copy full SHA for 1e66b72

2 files changed

+419

-3

lines changed

tensorrt_llm/_torch
- models
  - modeling_speculative.py
- pyexecutor
  - config_utils.py

+419

-3

lines changed

Comments

(0)