# SPDX-License-Identifier: Apache-2.0 # SPDX-FileCopyrightText: Copyright contributors to the vLLM project """Inference-only GraniteMoeShared model. The architecture is the same as granitemoe but with the addition of shared experts. Also serves the `granitemoe_swa` checkpoints (`GraniteMoeSWAForCausalLM`), which add the same per-layer sliding window, attention sink and per-layer RoPE support as `granite_swa` (see `granite.py`). """ from collections.abc import Iterable from itertools import islice import torch from torch import nn from transformers.models.granitemoeshared import GraniteMoeSharedConfig from vllm.compilation.decorators import support_torch_compile from vllm.config import CacheConfig, VllmConfig from vllm.distributed import get_pp_group from vllm.model_executor.layers.activation import SiluAndMul from vllm.model_executor.layers.layernorm import RMSNorm from vllm.model_executor.layers.linear import ( MergedColumnParallelLinear, RowParallelLinear, ) from vllm.model_executor.layers.logits_processor import LogitsProcessor from vllm.model_executor.layers.quantization import QuantizationConfig from vllm.model_executor.layers.vocab_parallel_embedding import ( ParallelLMHead, VocabParallelEmbedding, ) from vllm.sequence import IntermediateTensors from .granitemoe import GraniteMoeAttention, GraniteMoeModel, GraniteMoeMoE from .interfaces import EagleModelMixin, SupportsEagle3, SupportsLoRA, SupportsPP from .utils import AutoWeightsLoader, make_layers, maybe_prefix class GraniteMoeSharedMLP(nn.Module): def __init__( self, config: GraniteMoeSharedConfig, quant_config: QuantizationConfig | None = None, prefix: str = "", ): super().__init__() self.input_size = config.hidden_size self.hidden_size = config.shared_intermediate_size self.input_linear = MergedColumnParallelLinear( input_size=self.input_size, output_sizes=[self.hidden_size] * 2, bias=False, quant_config=quant_config, prefix=f"{prefix}.input_linear", ) self.output_linear = RowParallelLinear( self.hidden_size, self.input_size, bias=False, quant_config=quant_config, prefix=f"{prefix}.output_linear", ) if config.hidden_act != "silu": raise ValueError( f"Unsupported activation: {config.hidden_act}. " "Only silu is supported for now." ) self.act_fn = SiluAndMul() def forward(self, hidden_states: torch.Tensor) -> torch.Tensor: hidden_states, _ = self.input_linear(hidden_states) hidden_states = self.act_fn(hidden_states) hidden_states, _ = self.output_linear(hidden_states) return hidden_states class GraniteMoeSharedDecoderLayer(nn.Module): def __init__( self, config: GraniteMoeSharedConfig, cache_config: CacheConfig | None = None, quant_config: QuantizationConfig | None = None, prefix: str = "", ) -> None: super().__init__() self.hidden_size = config.hidden_size self.self_attn = GraniteMoeAttention( config=config, hidden_size=self.hidden_size, num_heads=config.num_attention_heads, max_position=config.max_position_embeddings, num_kv_heads=config.num_key_value_heads, cache_config=cache_config, quant_config=quant_config, prefix=f"{prefix}.self_attn", attention_multiplier=config.attention_multiplier, ) self.block_sparse_moe = GraniteMoeMoE( num_experts=config.num_local_experts, top_k=config.num_experts_per_tok, hidden_size=config.hidden_size, intermediate_size=config.intermediate_size, quant_config=quant_config, prefix=f"{prefix}.block_sparse_moe", ) self.shared_mlp = ( None if getattr(config, "shared_intermediate_size", 0) == 0 else GraniteMoeSharedMLP( config, quant_config=quant_config, prefix=f"{prefix}.shared_mlp" ) ) self.input_layernorm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps) self.post_attention_layernorm = RMSNorm( config.hidden_size, eps=config.rms_norm_eps ) self.residual_multiplier = config.residual_multiplier def forward( self, positions: torch.Tensor, hidden_states: torch.Tensor, ) -> torch.Tensor: # Self Attention residual = hidden_states hidden_states = self.input_layernorm(hidden_states) hidden_states = self.self_attn( positions=positions, hidden_states=hidden_states, ) hidden_states = residual + hidden_states * self.residual_multiplier residual = hidden_states hidden_states = self.post_attention_layernorm(hidden_states) if self.shared_mlp is None: hidden_states = self.block_sparse_moe(hidden_states) else: # create a copy since block_sparse_moe modifies in-place moe_hidden_states = hidden_states.clone() moe_hidden_states = self.block_sparse_moe(moe_hidden_states) hidden_states = moe_hidden_states + self.shared_mlp(hidden_states) del moe_hidden_states hidden_states = residual + hidden_states * self.residual_multiplier return hidden_states @support_torch_compile class GraniteMoeSharedModel(nn.Module, EagleModelMixin): def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""): super().__init__() config = vllm_config.model_config.hf_config cache_config = vllm_config.cache_config quant_config = vllm_config.quant_config self.config = config self.quant_config = quant_config # Required by MixtralModel self.vocab_size = config.vocab_size self.embed_tokens = VocabParallelEmbedding( self.vocab_size, config.hidden_size, quant_config=quant_config, ) self.embedding_multiplier = config.embedding_multiplier self.start_layer, self.end_layer, self.layers = make_layers( config.num_hidden_layers, lambda prefix: GraniteMoeSharedDecoderLayer( config, cache_config, quant_config=quant_config, prefix=prefix ), prefix=f"{prefix}.layers", ) self.norm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps) def embed_input_ids(self, input_ids: torch.Tensor) -> torch.Tensor: return self.embed_tokens(input_ids) def forward( self, input_ids: torch.Tensor | None, positions: torch.Tensor, intermediate_tensors: IntermediateTensors | None, inputs_embeds: torch.Tensor | None = None, ) -> torch.Tensor | IntermediateTensors | tuple[torch.Tensor, list[torch.Tensor]]: if get_pp_group().is_first_rank: if inputs_embeds is not None: hidden_states = inputs_embeds else: hidden_states = self.embed_input_ids(input_ids) hidden_states *= self.embedding_multiplier else: assert intermediate_tensors is not None hidden_states = intermediate_tensors["hidden_states"] aux_hidden_states = self._maybe_add_hidden_state([], 0, hidden_states, None) for idx, layer in enumerate( islice(self.layers, self.start_layer, self.end_layer) ): hidden_states = layer(positions, hidden_states) self._maybe_add_hidden_state( aux_hidden_states, idx + 1, hidden_states, None ) if not get_pp_group().is_last_rank: return IntermediateTensors( { "hidden_states": hidden_states, } ) hidden_states = self.norm(hidden_states) if len(aux_hidden_states) > 0: return hidden_states, aux_hidden_states return hidden_states hf_to_vllm_mapper = GraniteMoeModel.hf_to_vllm_mapper load_weights = GraniteMoeModel.load_weights class GraniteMoeSharedForCausalLM(nn.Module, SupportsLoRA, SupportsPP, SupportsEagle3): fall_back_to_pt_during_load = False packed_modules_mapping = { "qkv_proj": [ "q_proj", "k_proj", "v_proj", ], } # LoRA specific attributes embedding_modules = { "embed_tokens": "input_embeddings", "lm_head": "output_embeddings", } def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""): super().__init__() config = vllm_config.model_config.hf_config quant_config = vllm_config.quant_config self.config = config self.model = GraniteMoeSharedModel( vllm_config=vllm_config, prefix=maybe_prefix(prefix, "model") ) self.lm_head = ParallelLMHead( config.vocab_size, config.hidden_size, quant_config=quant_config, prefix=maybe_prefix(prefix, "lm_head"), ) if config.tie_word_embeddings: self.lm_head = self.lm_head.tie_weights(self.model.embed_tokens) self.logits_processor = LogitsProcessor( config.vocab_size, config.vocab_size, scale=1 / self.config.logits_scaling, ) def embed_input_ids(self, input_ids: torch.Tensor) -> torch.Tensor: return self.model.embed_input_ids(input_ids) def forward( self, input_ids: torch.Tensor | None, positions: torch.Tensor, intermediate_tensors: IntermediateTensors | None = None, inputs_embeds: torch.Tensor | None = None, ) -> torch.Tensor: hidden_states = self.model( input_ids, positions, intermediate_tensors, inputs_embeds ) return hidden_states def compute_logits(self, hidden_states: torch.Tensor) -> torch.Tensor | None: logits = self.logits_processor(self.lm_head, hidden_states) return logits def make_empty_intermediate_tensors( self, batch_size: int, dtype: torch.dtype, device: torch.device ) -> IntermediateTensors: return IntermediateTensors( { "hidden_states": torch.zeros( (batch_size, self.config.hidden_size), dtype=dtype, device=device ), } ) def load_weights(self, weights: Iterable[tuple[str, torch.Tensor]]) -> set[str]: loader = AutoWeightsLoader(self) return loader.load_weights(weights)