vllm.model_executor.layers

Modules:

Name	Description
`activation`	Custom activation functions.
`fused_moe`
`layernorm`	Custom normalization layers.
`lightning_attn`
`linear`
`logits_processor`	A layer that compute logits from hidden_stats.
`mamba`
`pooler`
`quantization`
`rejection_sampler`
`resampler`	Shared resampler perceiver network used in multimodal models and
`rotary_embedding`	Rotary Positional Embeddings.
`sampler`	A layer that samples the next tokens from the model's outputs.
`spec_decode_base_sampler`
`typical_acceptance_sampler`
`utils`	Utility methods for model layers.
`vocab_parallel_embedding`