vllm.model_executor.layers ¶

Modules:

Name	Description
`activation`	Custom activation functions.
`attention_layer_base`	Base class for attention-like layers.
`batch_invariant`
`fla`
`fused_moe`
`layernorm`	Custom normalization layers.
`lightning_attn`
`linear`
`logits_processor`	A layer that compute logits from hidden_stats.
`mamba`
`mla`
`pooler`
`quantization`
`resampler`	Shared resampler perceiver network used in multimodal models and
`rotary_embedding`	Rotary Positional Embeddings.
`shared_fused_moe`
`utils`	Utility methods for model layers.
`vocab_parallel_embedding`