`nemo_automodel.components.speculative.eagle.registry`#

Dispatch registry mapping target architecture -> EAGLE draft model.

Mirrors SpecForge’s modeling/auto.py pattern but keyed by HF architectures string instead of by config class. The string key avoids a hard import dependency on every individual HF *Config class (which can differ between transformers versions) and matches NeMo AutoModel’s existing _transformers/registry.MODEL_ARCH_MAPPING style.

Today the dense draft (LlamaEagle3DraftModel / LlamaEagleDraftModel) covers all registered architectures: the implementation is config-driven and reads attention_bias, mlp_bias, head_dim, rope_theta / rope_scaling, and rms_norm_eps directly from the target config. Adding an architecture that fits this shape is a one-line registry append. Architectures that need a different draft (e.g. gpt-oss with its alternating sliding-window attention and non-standard RoPE) get a new draft_cls entry pointing at a dedicated draft module.

Module Contents#

Classes#

DraftSpec

How to build an EAGLE draft model for a particular target architecture.

Functions#

`_resolve`	Return the first registered draft spec matching any architecture in the list.
`resolve_eagle3_draft_spec`	Resolve the EAGLE-3 draft spec for a target’s `config.architectures` field.
`resolve_eagle1_draft_spec`	Resolve the EAGLE-1 / EAGLE-2 draft spec for a target’s `config.architectures` field.

Data#

`_DENSE_ARCHITECTURES`
`EAGLE3_DRAFT_REGISTRY`
`EAGLE1_DRAFT_REGISTRY`

API#

class nemo_automodel.components.speculative.eagle.registry.DraftSpec#