vllm.model_executor.models.arctic

Inference-only Snowflake Arctic model.

logger module-attribute

logger = init_logger(__name__)

ArcticAttention

Bases: Module

Source code in vllm/model_executor/models/arctic.py

class ArcticAttention(nn.Module):

    def __init__(
        self,
        config: ArcticConfig,
        cache_config: Optional[CacheConfig] = None,
        quant_config: Optional[QuantizationConfig] = None,
        prefix: str = "",
    ):
        super().__init__()
        self.config = config
        self.hidden_size = config.hidden_size

        tp_size = get_tensor_model_parallel_world_size()
        self.total_num_heads = config.num_attention_heads
        assert self.total_num_heads % tp_size == 0
        self.num_heads = self.total_num_heads // tp_size
        self.total_num_kv_heads = config.num_key_value_heads
        if self.total_num_kv_heads >= tp_size:
            assert self.total_num_kv_heads % tp_size == 0
        else:
            assert tp_size % self.total_num_kv_heads == 0
        self.num_kv_heads = max(1, self.total_num_kv_heads // tp_size)
        self.head_dim = self.hidden_size // self.total_num_heads
        self.q_size = self.num_heads * self.head_dim
        self.kv_size = self.num_kv_heads * self.head_dim

        self.max_position_embeddings = config.max_position_embeddings
        self.rope_theta = config.rope_theta
        self.scaling = self.head_dim**-0.5

        self.qkv_proj = QKVParallelLinear(self.hidden_size,
                                          self.head_dim,
                                          self.total_num_heads,
                                          self.total_num_kv_heads,
                                          bias=False,
                                          quant_config=quant_config)
        self.o_proj = RowParallelLinear(
            self.total_num_heads * self.head_dim,
            self.hidden_size,
            bias=False,
            reduce_results=True,
            quant_config=quant_config,
        )

        self.rotary_emb = get_rope(
            self.head_dim,
            rotary_dim=self.head_dim,
            max_position=self.max_position_embeddings,
            base=int(self.rope_theta),
            is_neox_style=True,
        )

        self.attn = Attention(self.num_heads,
                              self.head_dim,
                              self.scaling,
                              num_kv_heads=self.num_kv_heads,
                              cache_config=cache_config,
                              quant_config=quant_config,
                              prefix=f"{prefix}.attn")

    def forward(
        self,
        positions: torch.Tensor,
        hidden_states: torch.Tensor,
    ) -> torch.Tensor:
        qkv, _ = self.qkv_proj(hidden_states)
        q, k, v = qkv.split([self.q_size, self.kv_size, self.kv_size], dim=-1)
        q, k = self.rotary_emb(positions, q, k)
        attn_output = self.attn(q, k, v)
        output, _ = self.o_proj(attn_output)
        return output

attn instance-attribute

attn = Attention(
    num_heads,
    head_dim,
    scaling,
    num_kv_heads=num_kv_heads,
    cache_config=cache_config,
    quant_config=quant_config,
    prefix=f"{prefix}.attn",
)

config instance-attribute

config = config

head_dim instance-attribute

head_dim = hidden_size // total_num_heads

hidden_size instance-attribute

hidden_size = hidden_size

kv_size instance-attribute

kv_size = num_kv_heads * head_dim

max_position_embeddings instance-attribute

max_position_embeddings = max_position_embeddings

num_heads instance-attribute

num_heads = total_num_heads // tp_size

num_kv_heads instance-attribute

num_kv_heads = max(1, total_num_kv_heads // tp_size)

o_proj instance-attribute

o_proj = RowParallelLinear(
    total_num_heads * head_dim,
    hidden_size,
    bias=False,
    reduce_results=True,
    quant_config=quant_config,
)

q_size instance-attribute

q_size = num_heads * head_dim

qkv_proj instance-attribute

qkv_proj = QKVParallelLinear(
    hidden_size,
    head_dim,
    total_num_heads,
    total_num_kv_heads,
    bias=False,
    quant_config=quant_config,
)

rope_theta instance-attribute

rope_theta = rope_theta

rotary_emb instance-attribute

rotary_emb = get_rope(
    head_dim,
    rotary_dim=head_dim,
    max_position=max_position_embeddings,
    base=int(rope_theta),
    is_neox_style=True,
)

scaling instance-attribute

scaling = head_dim ** -0.5

total_num_heads instance-attribute

total_num_heads = num_attention_heads

total_num_kv_heads instance-attribute

total_num_kv_heads = num_key_value_heads

__init__

__init__(
    config: ArcticConfig,
    cache_config: Optional[CacheConfig] = None,
    quant_config: Optional[QuantizationConfig] = None,
    prefix: str = "",
)

Source code in vllm/model_executor/models/arctic.py

def __init__(
    self,
    config: ArcticConfig,
    cache_config: Optional[CacheConfig] = None,
    quant_config: Optional[QuantizationConfig] = None,
    prefix: str = "",
):
    super().__init__()
    self.config = config
    self.hidden_size = config.hidden_size

    tp_size = get_tensor_model_parallel_world_size()
    self.total_num_heads = config.num_attention_heads
    assert self.total_num_heads % tp_size == 0
    self.num_heads = self.total_num_heads // tp_size
    self.total_num_kv_heads = config.num_key_value_heads
    if self.total_num_kv_heads >= tp_size:
        assert self.total_num_kv_heads % tp_size == 0
    else:
        assert tp_size % self.total_num_kv_heads == 0
    self.num_kv_heads = max(1, self.total_num_kv_heads // tp_size)
    self.head_dim = self.hidden_size // self.total_num_heads
    self.q_size = self.num_heads * self.head_dim
    self.kv_size = self.num_kv_heads * self.head_dim

    self.max_position_embeddings = config.max_position_embeddings
    self.rope_theta = config.rope_theta
    self.scaling = self.head_dim**-0.5

    self.qkv_proj = QKVParallelLinear(self.hidden_size,
                                      self.head_dim,
                                      self.total_num_heads,
                                      self.total_num_kv_heads,
                                      bias=False,
                                      quant_config=quant_config)
    self.o_proj = RowParallelLinear(
        self.total_num_heads * self.head_dim,
        self.hidden_size,
        bias=False,
        reduce_results=True,
        quant_config=quant_config,
    )

    self.rotary_emb = get_rope(
        self.head_dim,
        rotary_dim=self.head_dim,
        max_position=self.max_position_embeddings,
        base=int(self.rope_theta),
        is_neox_style=True,
    )

    self.attn = Attention(self.num_heads,
                          self.head_dim,
                          self.scaling,
                          num_kv_heads=self.num_kv_heads,
                          cache_config=cache_config,
                          quant_config=quant_config,
                          prefix=f"{prefix}.attn")

forward

forward(positions: Tensor, hidden_states: Tensor) -> Tensor

Source code in vllm/model_executor/models/arctic.py

def forward(
    self,
    positions: torch.Tensor,
    hidden_states: torch.Tensor,
) -> torch.Tensor:
    qkv, _ = self.qkv_proj(hidden_states)
    q, k, v = qkv.split([self.q_size, self.kv_size, self.kv_size], dim=-1)
    q, k = self.rotary_emb(positions, q, k)
    attn_output = self.attn(q, k, v)
    output, _ = self.o_proj(attn_output)
    return output

ArcticDecoderLayer

Bases: Module

Source code in vllm/model_executor/models/arctic.py

class ArcticDecoderLayer(nn.Module):

    def __init__(
        self,
        config: ArcticConfig,
        cache_config: Optional[CacheConfig] = None,
        quant_config: Optional[QuantizationConfig] = None,
        prefix: str = "",
    ) -> None:
        super().__init__()
        self.hidden_size = config.hidden_size
        layer_idx = extract_layer_index(prefix)
        is_moe_layer = (layer_idx + 1) % config.moe_layer_frequency == 0
        self.use_residual = config.use_residual and is_moe_layer
        self.self_attn = ArcticAttention(config,
                                         cache_config,
                                         quant_config=quant_config,
                                         prefix=f"{prefix}.self_attn")
        self.block_sparse_moe = ArcticMoE(
            config,
            quant_config=quant_config,
            reduce_results=(not self.use_residual),
            prefix=f"{prefix}.block_sparse_moe",
        )

        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)

        if self.use_residual:
            self.residual_layernorm = RMSNorm(config.hidden_size,
                                              eps=config.rms_norm_eps)
            self.residual_mlp = ArcticMLP(config,
                                          is_residual_mlp=True,
                                          reduce_results=False,
                                          prefix=f"{prefix}.residual_mlp")

    def forward(
        self,
        positions: torch.Tensor,
        hidden_states: torch.Tensor,
    ) -> torch.Tensor:
        residual_input = hidden_states
        hidden_states = self.input_layernorm(hidden_states)
        hidden_states = self.self_attn(
            positions=positions,
            hidden_states=hidden_states,
        )
        hidden_states = residual_input + hidden_states

        residual_attn = hidden_states
        if self.use_residual:
            hidden_states = self.residual_layernorm(hidden_states)
            hidden_states = self.residual_mlp(hidden_states)
            residual_mlp = hidden_states
            hidden_states = self.post_attention_layernorm(residual_input)
            hidden_states = self.block_sparse_moe(hidden_states)
            hidden_states = residual_mlp + hidden_states
            hidden_states = tensor_model_parallel_all_reduce(hidden_states)
            hidden_states = residual_attn + hidden_states
        else:
            hidden_states = self.post_attention_layernorm(hidden_states)
            hidden_states = self.block_sparse_moe(hidden_states)
            hidden_states = residual_attn + hidden_states
        return hidden_states

block_sparse_moe instance-attribute

block_sparse_moe = ArcticMoE(
    config,
    quant_config=quant_config,
    reduce_results=not use_residual,
    prefix=f"{prefix}.block_sparse_moe",
)

hidden_size instance-attribute

hidden_size = hidden_size

input_layernorm instance-attribute

input_layernorm = RMSNorm(hidden_size, eps=rms_norm_eps)

post_attention_layernorm instance-attribute

post_attention_layernorm = RMSNorm(
    hidden_size, eps=rms_norm_eps
)

residual_layernorm instance-attribute

residual_layernorm = RMSNorm(hidden_size, eps=rms_norm_eps)

residual_mlp instance-attribute

residual_mlp = ArcticMLP(
    config,
    is_residual_mlp=True,
    reduce_results=False,
    prefix=f"{prefix}.residual_mlp",
)

self_attn instance-attribute

self_attn = ArcticAttention(
    config,
    cache_config,
    quant_config=quant_config,
    prefix=f"{prefix}.self_attn",
)

use_residual instance-attribute

use_residual = use_residual and is_moe_layer

__init__

__init__(
    config: ArcticConfig,
    cache_config: Optional[CacheConfig] = None,
    quant_config: Optional[QuantizationConfig] = None,
    prefix: str = "",
) -> None

Source code in vllm/model_executor/models/arctic.py

def __init__(
    self,
    config: ArcticConfig,
    cache_config: Optional[CacheConfig] = None,
    quant_config: Optional[QuantizationConfig] = None,
    prefix: str = "",
) -> None:
    super().__init__()
    self.hidden_size = config.hidden_size
    layer_idx = extract_layer_index(prefix)
    is_moe_layer = (layer_idx + 1) % config.moe_layer_frequency == 0
    self.use_residual = config.use_residual and is_moe_layer
    self.self_attn = ArcticAttention(config,
                                     cache_config,
                                     quant_config=quant_config,
                                     prefix=f"{prefix}.self_attn")
    self.block_sparse_moe = ArcticMoE(
        config,
        quant_config=quant_config,
        reduce_results=(not self.use_residual),
        prefix=f"{prefix}.block_sparse_moe",
    )

    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)

    if self.use_residual:
        self.residual_layernorm = RMSNorm(config.hidden_size,
                                          eps=config.rms_norm_eps)
        self.residual_mlp = ArcticMLP(config,
                                      is_residual_mlp=True,
                                      reduce_results=False,
                                      prefix=f"{prefix}.residual_mlp")

forward

forward(positions: Tensor, hidden_states: Tensor) -> Tensor

Source code in vllm/model_executor/models/arctic.py

def forward(
    self,
    positions: torch.Tensor,
    hidden_states: torch.Tensor,
) -> torch.Tensor:
    residual_input = hidden_states
    hidden_states = self.input_layernorm(hidden_states)
    hidden_states = self.self_attn(
        positions=positions,
        hidden_states=hidden_states,
    )
    hidden_states = residual_input + hidden_states

    residual_attn = hidden_states
    if self.use_residual:
        hidden_states = self.residual_layernorm(hidden_states)
        hidden_states = self.residual_mlp(hidden_states)
        residual_mlp = hidden_states
        hidden_states = self.post_attention_layernorm(residual_input)
        hidden_states = self.block_sparse_moe(hidden_states)
        hidden_states = residual_mlp + hidden_states
        hidden_states = tensor_model_parallel_all_reduce(hidden_states)
        hidden_states = residual_attn + hidden_states
    else:
        hidden_states = self.post_attention_layernorm(hidden_states)
        hidden_states = self.block_sparse_moe(hidden_states)
        hidden_states = residual_attn + hidden_states
    return hidden_states

ArcticForCausalLM

Bases: Module, SupportsPP, SupportsQuant

Source code in vllm/model_executor/models/arctic.py

class ArcticForCausalLM(nn.Module, SupportsPP, SupportsQuant):
    packed_modules_mapping = {"qkv_proj": ["q_proj", "k_proj", "v_proj"]}

    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 = ArcticModel(vllm_config=vllm_config,
                                 prefix=maybe_prefix(prefix, "model"))
        self.vocab_size = config.vocab_size
        self.lm_head = ParallelLMHead(
            self.vocab_size,
            config.hidden_size,
            quant_config=quant_config,
        )
        if self.config.tie_word_embeddings:
            self.lm_head.weight = self.model.embed_tokens.weight
        self.num_experts = config.num_local_experts
        self.num_experts_per_tok = config.num_experts_per_tok
        self.unpadded_vocab_size = config.vocab_size
        self.logits_processor = LogitsProcessor(self.unpadded_vocab_size,
                                                config.vocab_size)
        self.make_empty_intermediate_tensors = (
            self.model.make_empty_intermediate_tensors)

    def get_input_embeddings(self, input_ids: torch.Tensor) -> torch.Tensor:
        return self.model.get_input_embeddings(input_ids)

    def forward(
        self,
        input_ids: torch.Tensor,
        positions: torch.Tensor,
        intermediate_tensors: Optional[IntermediateTensors] = None,
        inputs_embeds: Optional[torch.Tensor] = None,
    ) -> Union[torch.Tensor, IntermediateTensors]:
        hidden_states = self.model(input_ids, positions, intermediate_tensors,
                                   inputs_embeds)
        return hidden_states

    def compute_logits(
        self,
        hidden_states: torch.Tensor,
        sampling_metadata: SamplingMetadata,
    ) -> Optional[torch.Tensor]:
        logits = self.logits_processor(self.lm_head, hidden_states,
                                       sampling_metadata)
        return logits

    def load_weights(self, weights: Iterable[tuple[str,
                                                   torch.Tensor]]) -> set[str]:
        stacked_params_mapping = [
            # (param_name, shard_name, shard_id)
            ("qkv_proj", "q_proj", "q"),
            ("qkv_proj", "k_proj", "k"),
            ("qkv_proj", "v_proj", "v"),
        ]

        mlp_params_mapping: list[tuple[str, str, int]] = []
        expert_params_mapping: list[tuple[str, str, int]] = []
        num_layers = self.config.num_hidden_layers

        for layer in range(num_layers):
            mlp_params_mapping.append(
                (f"layers.{layer}.residual_mlp.w13.weight",
                 f"layers.{layer}.residual_mlp.w1.weight", 0))
            mlp_params_mapping.append(
                (f"layers.{layer}.residual_mlp.w13.weight",
                 f"layers.{layer}.residual_mlp.w3.weight", 1))
            if layer % 2 == 0:
                # MLP layers
                mlp_params_mapping.append(
                    (f"layers.{layer}.block_sparse_moe.mlp.w13.weight",
                     f"layers.{layer}.block_sparse_moe.mlp.w1.weight", 0))
                mlp_params_mapping.append(
                    (f"layers.{layer}.block_sparse_moe.mlp.w13.weight",
                     f"layers.{layer}.block_sparse_moe.mlp.w3.weight", 1))
            else:
                # MoE layers
                for expert_id in range(self.config.num_local_experts):
                    expert_params_mapping.append(
                        ("ws", f"experts.{expert_id}.w1.weight", expert_id))
                    expert_params_mapping.append(
                        ("w2s", f"experts.{expert_id}.w2.weight", expert_id))
                    expert_params_mapping.append(
                        ("ws", f"experts.{expert_id}.w3.weight", expert_id))

        params_dict = dict(self.named_parameters())
        loaded_params: set[str] = set()

        logger.info(
            "It will take ~10 minutes loading from the 16-bit weights. "
            "Alternatively, use the prequantized 8-bit weights of arctic "
            "and set load-format to `sharded_state` will accelerate loading.")
        for name, loaded_weight in weights:
            for (param_name, weight_name, shard_id) in stacked_params_mapping:
                if weight_name not in name:
                    continue
                name = name.replace(weight_name, param_name)
                # Skip loading extra bias for GPTQ models.
                if name.endswith(".bias") and name not in params_dict:
                    continue
                if is_pp_missing_parameter(name, self):
                    continue
                param = params_dict[name]
                weight_loader = param.weight_loader
                weight_loader(param, loaded_weight, shard_id)
                break
            else:
                for param_name, weight_name, shard_id in mlp_params_mapping:
                    if weight_name not in name:
                        continue
                    name = name.replace(weight_name, param_name)
                    if is_pp_missing_parameter(name, self):
                        continue
                    param = params_dict[name]
                    weight_loader = param.weight_loader
                    weight_loader(param, loaded_weight, shard_id)
                    break
                else:
                    for param_name, weight_name, shard_id \
                            in expert_params_mapping:
                        if weight_name not in name:
                            continue
                        name = name.replace(weight_name, param_name)
                        if is_pp_missing_parameter(name, self):
                            continue
                        param = params_dict[name]
                        weight_loader = param.weight_loader
                        weight_loader(param,
                                      loaded_weight,
                                      weight_name,
                                      expert_id=shard_id)
                        break
                    else:
                        if name.endswith(".bias") and name not in params_dict:
                            continue
                        if is_pp_missing_parameter(name, self):
                            continue
                        param = params_dict[name]

                        weight_loader = getattr(param, "weight_loader",
                                                default_weight_loader)
                        weight_loader(param, loaded_weight)
            loaded_params.add(name)
        return loaded_params

config instance-attribute

config = config

lm_head instance-attribute

lm_head = ParallelLMHead(
    vocab_size, hidden_size, quant_config=quant_config
)

logits_processor instance-attribute

logits_processor = LogitsProcessor(
    unpadded_vocab_size, vocab_size
)

make_empty_intermediate_tensors instance-attribute

make_empty_intermediate_tensors = (
    make_empty_intermediate_tensors
)

model instance-attribute

model = ArcticModel(
    vllm_config=vllm_config,
    prefix=maybe_prefix(prefix, "model"),
)

num_experts instance-attribute

num_experts = num_local_experts

num_experts_per_tok instance-attribute

num_experts_per_tok = num_experts_per_tok

packed_modules_mapping class-attribute instance-attribute

packed_modules_mapping = {
    "qkv_proj": ["q_proj", "k_proj", "v_proj"]
}

unpadded_vocab_size instance-attribute

unpadded_vocab_size = vocab_size

vocab_size instance-attribute

vocab_size = vocab_size

__init__

__init__(*, vllm_config: VllmConfig, prefix: str = '')

Source code in vllm/model_executor/models/arctic.py

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 = ArcticModel(vllm_config=vllm_config,
                             prefix=maybe_prefix(prefix, "model"))
    self.vocab_size = config.vocab_size
    self.lm_head = ParallelLMHead(
        self.vocab_size,
        config.hidden_size,
        quant_config=quant_config,
    )
    if self.config.tie_word_embeddings:
        self.lm_head.weight = self.model.embed_tokens.weight
    self.num_experts = config.num_local_experts
    self.num_experts_per_tok = config.num_experts_per_tok
    self.unpadded_vocab_size = config.vocab_size
    self.logits_processor = LogitsProcessor(self.unpadded_vocab_size,
                                            config.vocab_size)
    self.make_empty_intermediate_tensors = (
        self.model.make_empty_intermediate_tensors)

compute_logits

compute_logits(
    hidden_states: Tensor,
    sampling_metadata: SamplingMetadata,
) -> Optional[Tensor]

Source code in vllm/model_executor/models/arctic.py

def compute_logits(
    self,
    hidden_states: torch.Tensor,
    sampling_metadata: SamplingMetadata,
) -> Optional[torch.Tensor]:
    logits = self.logits_processor(self.lm_head, hidden_states,
                                   sampling_metadata)
    return logits

forward

forward(
    input_ids: Tensor,
    positions: Tensor,
    intermediate_tensors: Optional[
        IntermediateTensors
    ] = None,
    inputs_embeds: Optional[Tensor] = None,
) -> Union[Tensor, IntermediateTensors]

Source code in vllm/model_executor/models/arctic.py

def forward(
    self,
    input_ids: torch.Tensor,
    positions: torch.Tensor,
    intermediate_tensors: Optional[IntermediateTensors] = None,
    inputs_embeds: Optional[torch.Tensor] = None,
) -> Union[torch.Tensor, IntermediateTensors]:
    hidden_states = self.model(input_ids, positions, intermediate_tensors,
                               inputs_embeds)
    return hidden_states

get_input_embeddings

get_input_embeddings(input_ids: Tensor) -> Tensor

Source code in vllm/model_executor/models/arctic.py

def get_input_embeddings(self, input_ids: torch.Tensor) -> torch.Tensor:
    return self.model.get_input_embeddings(input_ids)

load_weights

load_weights(
    weights: Iterable[tuple[str, Tensor]],
) -> set[str]

Source code in vllm/model_executor/models/arctic.py

def load_weights(self, weights: Iterable[tuple[str,
                                               torch.Tensor]]) -> set[str]:
    stacked_params_mapping = [
        # (param_name, shard_name, shard_id)
        ("qkv_proj", "q_proj", "q"),
        ("qkv_proj", "k_proj", "k"),
        ("qkv_proj", "v_proj", "v"),
    ]

    mlp_params_mapping: list[tuple[str, str, int]] = []
    expert_params_mapping: list[tuple[str, str, int]] = []
    num_layers = self.config.num_hidden_layers

    for layer in range(num_layers):
        mlp_params_mapping.append(
            (f"layers.{layer}.residual_mlp.w13.weight",
             f"layers.{layer}.residual_mlp.w1.weight", 0))
        mlp_params_mapping.append(
            (f"layers.{layer}.residual_mlp.w13.weight",
             f"layers.{layer}.residual_mlp.w3.weight", 1))
        if layer % 2 == 0:
            # MLP layers
            mlp_params_mapping.append(
                (f"layers.{layer}.block_sparse_moe.mlp.w13.weight",
                 f"layers.{layer}.block_sparse_moe.mlp.w1.weight", 0))
            mlp_params_mapping.append(
                (f"layers.{layer}.block_sparse_moe.mlp.w13.weight",
                 f"layers.{layer}.block_sparse_moe.mlp.w3.weight", 1))
        else:
            # MoE layers
            for expert_id in range(self.config.num_local_experts):
                expert_params_mapping.append(
                    ("ws", f"experts.{expert_id}.w1.weight", expert_id))
                expert_params_mapping.append(
                    ("w2s", f"experts.{expert_id}.w2.weight", expert_id))
                expert_params_mapping.append(
                    ("ws", f"experts.{expert_id}.w3.weight", expert_id))

    params_dict = dict(self.named_parameters())
    loaded_params: set[str] = set()

    logger.info(
        "It will take ~10 minutes loading from the 16-bit weights. "
        "Alternatively, use the prequantized 8-bit weights of arctic "
        "and set load-format to `sharded_state` will accelerate loading.")
    for name, loaded_weight in weights:
        for (param_name, weight_name, shard_id) in stacked_params_mapping:
            if weight_name not in name:
                continue
            name = name.replace(weight_name, param_name)
            # Skip loading extra bias for GPTQ models.
            if name.endswith(".bias") and name not in params_dict:
                continue
            if is_pp_missing_parameter(name, self):
                continue
            param = params_dict[name]
            weight_loader = param.weight_loader
            weight_loader(param, loaded_weight, shard_id)
            break
        else:
            for param_name, weight_name, shard_id in mlp_params_mapping:
                if weight_name not in name:
                    continue
                name = name.replace(weight_name, param_name)
                if is_pp_missing_parameter(name, self):
                    continue
                param = params_dict[name]
                weight_loader = param.weight_loader
                weight_loader(param, loaded_weight, shard_id)
                break
            else:
                for param_name, weight_name, shard_id \
                        in expert_params_mapping:
                    if weight_name not in name:
                        continue
                    name = name.replace(weight_name, param_name)
                    if is_pp_missing_parameter(name, self):
                        continue
                    param = params_dict[name]
                    weight_loader = param.weight_loader
                    weight_loader(param,
                                  loaded_weight,
                                  weight_name,
                                  expert_id=shard_id)
                    break
                else:
                    if name.endswith(".bias") and name not in params_dict:
                        continue
                    if is_pp_missing_parameter(name, self):
                        continue
                    param = params_dict[name]

                    weight_loader = getattr(param, "weight_loader",
                                            default_weight_loader)
                    weight_loader(param, loaded_weight)
        loaded_params.add(name)
    return loaded_params

ArcticMLP

Bases: Module

Source code in vllm/model_executor/models/arctic.py

class ArcticMLP(nn.Module):

    def __init__(self,
                 config: ArcticConfig,
                 expert_id: int = -1,
                 is_residual_mlp: bool = False,
                 quant_config: Optional[QuantizationConfig] = None,
                 reduce_results: bool = True,
                 prefix: str = ""):
        super().__init__()
        self.hidden_size = config.hidden_size
        self.expert_id = expert_id

        self.ffn_dim = config.intermediate_size if not is_residual_mlp \
            else self.hidden_size

        self.w13 = MergedColumnParallelLinear(self.hidden_size,
                                              [self.ffn_dim] * 2,
                                              bias=False,
                                              quant_config=quant_config)
        self.w2 = RowParallelLinear(self.ffn_dim,
                                    self.hidden_size,
                                    bias=False,
                                    reduce_results=reduce_results,
                                    quant_config=quant_config)
        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):
        gate_up, _ = self.w13(hidden_states)
        hidden_states = self.act_fn(gate_up)
        hidden_states, _ = self.w2(hidden_states)
        return hidden_states

act_fn instance-attribute

act_fn = SiluAndMul()

expert_id instance-attribute

expert_id = expert_id

ffn_dim instance-attribute

ffn_dim = (
    intermediate_size
    if not is_residual_mlp
    else hidden_size
)

hidden_size instance-attribute

hidden_size = hidden_size

w13 instance-attribute

w13 = MergedColumnParallelLinear(
    hidden_size,
    [ffn_dim] * 2,
    bias=False,
    quant_config=quant_config,
)

w2 instance-attribute

w2 = RowParallelLinear(
    ffn_dim,
    hidden_size,
    bias=False,
    reduce_results=reduce_results,
    quant_config=quant_config,
)

__init__

__init__(
    config: ArcticConfig,
    expert_id: int = -1,
    is_residual_mlp: bool = False,
    quant_config: Optional[QuantizationConfig] = None,
    reduce_results: bool = True,
    prefix: str = "",
)

Source code in vllm/model_executor/models/arctic.py

def __init__(self,
             config: ArcticConfig,
             expert_id: int = -1,
             is_residual_mlp: bool = False,
             quant_config: Optional[QuantizationConfig] = None,
             reduce_results: bool = True,
             prefix: str = ""):
    super().__init__()
    self.hidden_size = config.hidden_size
    self.expert_id = expert_id

    self.ffn_dim = config.intermediate_size if not is_residual_mlp \
        else self.hidden_size

    self.w13 = MergedColumnParallelLinear(self.hidden_size,
                                          [self.ffn_dim] * 2,
                                          bias=False,
                                          quant_config=quant_config)
    self.w2 = RowParallelLinear(self.ffn_dim,
                                self.hidden_size,
                                bias=False,
                                reduce_results=reduce_results,
                                quant_config=quant_config)
    if config.hidden_act != "silu":
        raise ValueError(f"Unsupported activation: {config.hidden_act}. "
                         "Only silu is supported for now.")
    self.act_fn = SiluAndMul()

forward

forward(hidden_states)

Source code in vllm/model_executor/models/arctic.py

def forward(self, hidden_states):
    gate_up, _ = self.w13(hidden_states)
    hidden_states = self.act_fn(gate_up)
    hidden_states, _ = self.w2(hidden_states)
    return hidden_states

ArcticMoE

Bases: Module

Model-parallel implementation of Arctic MoE Layer.

Source code in vllm/model_executor/models/arctic.py

class ArcticMoE(nn.Module):
    """
    Model-parallel implementation of Arctic MoE Layer.
    """

    def __init__(self,
                 config: ArcticConfig,
                 tp_size: Optional[int] = None,
                 params_dtype: Optional[torch.dtype] = None,
                 quant_config: Optional[QuantizationConfig] = None,
                 reduce_results: bool = True,
                 prefix: str = ""):
        super().__init__()

        layer_id = extract_layer_index(prefix)
        self.tp_size = tp_size or get_tensor_model_parallel_world_size()
        self.hidden_size = config.hidden_size
        self.num_experts = config.num_local_experts
        self.layer_id = layer_id
        self.top_k = config.num_experts_per_tok
        self.intermediate_size = config.intermediate_size // self.tp_size

        self.is_moe_layer = (layer_id + 1) % config.moe_layer_frequency == 0
        self.is_quant = isinstance(quant_config, DeepSpeedFPConfig)
        self.reduce_results = reduce_results
        # Some other parameters
        if params_dtype is None:
            params_dtype = torch.get_default_dtype()
        self.params_dtype = params_dtype

        if not self.is_moe_layer:
            self.mlp = ArcticMLP(config,
                                 quant_config=quant_config,
                                 reduce_results=reduce_results,
                                 prefix=f"{prefix}.mlp")
        else:
            self.gate = ReplicatedLinear(self.hidden_size,
                                         self.num_experts,
                                         bias=False,
                                         params_dtype=self.params_dtype,
                                         quant_config=quant_config,
                                         prefix=f"{prefix}.gate")
            if self.is_quant:
                self.ws = DeepSpeedFPParameter(
                    torch.Size((self.num_experts, 2 * self.intermediate_size,
                                self.hidden_size)),
                    params_dtype=params_dtype,
                    quant_config=quant_config,
                )
                self.w2s = DeepSpeedFPParameter(
                    torch.Size((self.num_experts, self.hidden_size,
                                self.intermediate_size)),
                    params_dtype=params_dtype,
                    quant_config=quant_config,
                )
            else:
                self.ws = nn.Parameter(
                    torch.empty(self.num_experts,
                                2 * self.intermediate_size,
                                self.hidden_size,
                                device=current_platform.device_type,
                                dtype=self.params_dtype))
                self.w2s = nn.Parameter(
                    torch.empty(self.num_experts,
                                self.hidden_size,
                                self.intermediate_size,
                                device=current_platform.device_type,
                                dtype=self.params_dtype))
            set_weight_attrs(self.ws, {
                "weight_loader": self.weight_loader,
            })
            set_weight_attrs(self.w2s, {
                "weight_loader": self.weight_loader,
            })

    def weight_loader(self, param: nn.Parameter, loaded_weight: torch.Tensor,
                      weight_name: str, expert_id: int):
        tp_rank = get_tensor_model_parallel_rank()
        param_data = param.ds_dequantize() if self.is_quant else param.data
        shard_size = self.intermediate_size
        shard = slice(tp_rank * shard_size, (tp_rank + 1) * shard_size)
        if weight_name.endswith("w1.weight"):
            param_data[expert_id, 0:shard_size, :] = loaded_weight[shard, :]
        if weight_name.endswith("w3.weight"):
            param_data[expert_id,
                       shard_size:2 * shard_size, :] = loaded_weight[shard, :]
        if weight_name.endswith("w2.weight"):
            param_data[expert_id, :, :] = loaded_weight[:, shard]
        if self.is_quant:
            param.ds_quantize_(param_data)

    def local_moe_fused(self, hidden_states: torch.Tensor) -> torch.Tensor:
        num_tokens, hidden_size = hidden_states.shape
        hidden_states = hidden_states.view(-1, self.hidden_size)
        # router_logits: (num_tokens, n_experts)
        router_logits, _ = self.gate(hidden_states)
        do_normalize = self.top_k > 1
        topk_weights, topk_ids, token_expert_indices = fused_topk(
            hidden_states, router_logits, self.top_k, renormalize=do_normalize)
        # topk_ids: (num_tokens, k)
        if self.is_quant:
            if 2 * num_tokens <= self.num_experts:
                # If much fewer tokens than experts, use selective dequantize.
                ws_dequantized = self.ws.ds_selective_dequantize(
                    topk_ids.flatten())
                w2s_dequantized = self.w2s.ds_selective_dequantize(
                    topk_ids.flatten())
                # We gathered the experts to the tokens so update the mapping.
                topk_ids = torch.arange(
                    0,
                    topk_ids.numel(),
                    device=topk_ids.device,
                ).reshape(topk_ids.shape)
            else:
                ws_dequantized = self.ws.ds_dequantize()
                w2s_dequantized = self.w2s.ds_dequantize()

        final_hidden_states = fused_experts(
            hidden_states,
            ws_dequantized if self.is_quant else self.ws,
            w2s_dequantized if self.is_quant else self.w2s,
            topk_weights,
            topk_ids,
            inplace=True)
        if self.reduce_results and self.tp_size > 1:
            final_hidden_states = tensor_model_parallel_all_reduce(
                final_hidden_states)
        return final_hidden_states.view(num_tokens, hidden_size)

    def forward(self, hidden_states: torch.Tensor):
        if self.is_moe_layer:
            final_hidden_states = self.local_moe_fused(hidden_states)
        else:
            final_hidden_states = self.mlp(hidden_states)
        return final_hidden_states

gate instance-attribute

gate = ReplicatedLinear(
    hidden_size,
    num_experts,
    bias=False,
    params_dtype=params_dtype,
    quant_config=quant_config,
    prefix=f"{prefix}.gate",
)

hidden_size instance-attribute

hidden_size = hidden_size

intermediate_size instance-attribute

intermediate_size = intermediate_size // tp_size

is_moe_layer instance-attribute

is_moe_layer = layer_id + 1 % moe_layer_frequency == 0

is_quant instance-attribute

is_quant = isinstance(quant_config, DeepSpeedFPConfig)

layer_id instance-attribute

layer_id = layer_id

mlp instance-attribute

mlp = ArcticMLP(
    config,
    quant_config=quant_config,
    reduce_results=reduce_results,
    prefix=f"{prefix}.mlp",
)

num_experts instance-attribute

num_experts = num_local_experts

params_dtype instance-attribute

params_dtype = params_dtype

reduce_results instance-attribute

reduce_results = reduce_results

top_k instance-attribute

top_k = num_experts_per_tok

tp_size instance-attribute

tp_size = tp_size or get_tensor_model_parallel_world_size()

w2s instance-attribute

w2s = DeepSpeedFPParameter(
    Size((num_experts, hidden_size, intermediate_size)),
    params_dtype=params_dtype,
    quant_config=quant_config,
)

ws instance-attribute

ws = DeepSpeedFPParameter(
    Size((num_experts, 2 * intermediate_size, hidden_size)),
    params_dtype=params_dtype,
    quant_config=quant_config,
)

__init__

__init__(
    config: ArcticConfig,
    tp_size: Optional[int] = None,
    params_dtype: Optional[dtype] = None,
    quant_config: Optional[QuantizationConfig] = None,
    reduce_results: bool = True,
    prefix: str = "",
)

Source code in vllm/model_executor/models/arctic.py

def __init__(self,
             config: ArcticConfig,
             tp_size: Optional[int] = None,
             params_dtype: Optional[torch.dtype] = None,
             quant_config: Optional[QuantizationConfig] = None,
             reduce_results: bool = True,
             prefix: str = ""):
    super().__init__()

    layer_id = extract_layer_index(prefix)
    self.tp_size = tp_size or get_tensor_model_parallel_world_size()
    self.hidden_size = config.hidden_size
    self.num_experts = config.num_local_experts
    self.layer_id = layer_id
    self.top_k = config.num_experts_per_tok
    self.intermediate_size = config.intermediate_size // self.tp_size

    self.is_moe_layer = (layer_id + 1) % config.moe_layer_frequency == 0
    self.is_quant = isinstance(quant_config, DeepSpeedFPConfig)
    self.reduce_results = reduce_results
    # Some other parameters
    if params_dtype is None:
        params_dtype = torch.get_default_dtype()
    self.params_dtype = params_dtype

    if not self.is_moe_layer:
        self.mlp = ArcticMLP(config,
                             quant_config=quant_config,
                             reduce_results=reduce_results,
                             prefix=f"{prefix}.mlp")
    else:
        self.gate = ReplicatedLinear(self.hidden_size,
                                     self.num_experts,
                                     bias=False,
                                     params_dtype=self.params_dtype,
                                     quant_config=quant_config,
                                     prefix=f"{prefix}.gate")
        if self.is_quant:
            self.ws = DeepSpeedFPParameter(
                torch.Size((self.num_experts, 2 * self.intermediate_size,
                            self.hidden_size)),
                params_dtype=params_dtype,
                quant_config=quant_config,
            )
            self.w2s = DeepSpeedFPParameter(
                torch.Size((self.num_experts, self.hidden_size,
                            self.intermediate_size)),
                params_dtype=params_dtype,
                quant_config=quant_config,
            )
        else:
            self.ws = nn.Parameter(
                torch.empty(self.num_experts,
                            2 * self.intermediate_size,
                            self.hidden_size,
                            device=current_platform.device_type,
                            dtype=self.params_dtype))
            self.w2s = nn.Parameter(
                torch.empty(self.num_experts,
                            self.hidden_size,
                            self.intermediate_size,
                            device=current_platform.device_type,
                            dtype=self.params_dtype))
        set_weight_attrs(self.ws, {
            "weight_loader": self.weight_loader,
        })
        set_weight_attrs(self.w2s, {
            "weight_loader": self.weight_loader,
        })

forward

forward(hidden_states: Tensor)

Source code in vllm/model_executor/models/arctic.py

def forward(self, hidden_states: torch.Tensor):
    if self.is_moe_layer:
        final_hidden_states = self.local_moe_fused(hidden_states)
    else:
        final_hidden_states = self.mlp(hidden_states)
    return final_hidden_states

local_moe_fused

local_moe_fused(hidden_states: Tensor) -> Tensor

Source code in vllm/model_executor/models/arctic.py

def local_moe_fused(self, hidden_states: torch.Tensor) -> torch.Tensor:
    num_tokens, hidden_size = hidden_states.shape
    hidden_states = hidden_states.view(-1, self.hidden_size)
    # router_logits: (num_tokens, n_experts)
    router_logits, _ = self.gate(hidden_states)
    do_normalize = self.top_k > 1
    topk_weights, topk_ids, token_expert_indices = fused_topk(
        hidden_states, router_logits, self.top_k, renormalize=do_normalize)
    # topk_ids: (num_tokens, k)
    if self.is_quant:
        if 2 * num_tokens <= self.num_experts:
            # If much fewer tokens than experts, use selective dequantize.
            ws_dequantized = self.ws.ds_selective_dequantize(
                topk_ids.flatten())
            w2s_dequantized = self.w2s.ds_selective_dequantize(
                topk_ids.flatten())
            # We gathered the experts to the tokens so update the mapping.
            topk_ids = torch.arange(
                0,
                topk_ids.numel(),
                device=topk_ids.device,
            ).reshape(topk_ids.shape)
        else:
            ws_dequantized = self.ws.ds_dequantize()
            w2s_dequantized = self.w2s.ds_dequantize()

    final_hidden_states = fused_experts(
        hidden_states,
        ws_dequantized if self.is_quant else self.ws,
        w2s_dequantized if self.is_quant else self.w2s,
        topk_weights,
        topk_ids,
        inplace=True)
    if self.reduce_results and self.tp_size > 1:
        final_hidden_states = tensor_model_parallel_all_reduce(
            final_hidden_states)
    return final_hidden_states.view(num_tokens, hidden_size)

weight_loader

weight_loader(
    param: Parameter,
    loaded_weight: Tensor,
    weight_name: str,
    expert_id: int,
)

Source code in vllm/model_executor/models/arctic.py

def weight_loader(self, param: nn.Parameter, loaded_weight: torch.Tensor,
                  weight_name: str, expert_id: int):
    tp_rank = get_tensor_model_parallel_rank()
    param_data = param.ds_dequantize() if self.is_quant else param.data
    shard_size = self.intermediate_size
    shard = slice(tp_rank * shard_size, (tp_rank + 1) * shard_size)
    if weight_name.endswith("w1.weight"):
        param_data[expert_id, 0:shard_size, :] = loaded_weight[shard, :]
    if weight_name.endswith("w3.weight"):
        param_data[expert_id,
                   shard_size:2 * shard_size, :] = loaded_weight[shard, :]
    if weight_name.endswith("w2.weight"):
        param_data[expert_id, :, :] = loaded_weight[:, shard]
    if self.is_quant:
        param.ds_quantize_(param_data)

ArcticModel

Bases: Module

Source code in vllm/model_executor/models/arctic.py

@support_torch_compile
class ArcticModel(nn.Module):

    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.vocab_size = config.vocab_size
        self.embed_tokens = VocabParallelEmbedding(
            self.vocab_size,
            config.hidden_size,
            org_num_embeddings=self.vocab_size)
        self.start_layer, self.end_layer, self.layers = make_layers(
            config.num_hidden_layers,
            lambda prefix: ArcticDecoderLayer(
                config, cache_config, quant_config, prefix=prefix),
            prefix=f"{prefix}.layers")
        self._attn_implementation = config._attn_implementation
        self.norm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps)
        self.make_empty_intermediate_tensors = (
            make_empty_intermediate_tensors_factory(["hidden_states"],
                                                    config.hidden_size))

    def get_input_embeddings(self, input_ids: torch.Tensor) -> torch.Tensor:
        return self.embed_tokens(input_ids)

    def forward(
        self,
        input_ids: torch.Tensor,
        positions: torch.Tensor,
        intermediate_tensors: Optional[IntermediateTensors],
        inputs_embeds: Optional[torch.Tensor] = None,
    ) -> Union[torch.Tensor, IntermediateTensors]:
        if get_pp_group().is_first_rank:
            if inputs_embeds is not None:
                hidden_states = inputs_embeds
            else:
                hidden_states = self.get_input_embeddings(input_ids)
        else:
            assert intermediate_tensors is not None
            hidden_states = intermediate_tensors["hidden_states"]
        for layer in self.layers[self.start_layer:self.end_layer]:
            hidden_states = layer(positions, hidden_states)
        if not get_pp_group().is_last_rank:
            return IntermediateTensors({"hidden_states": hidden_states})
        hidden_states = self.norm(hidden_states)
        return hidden_states

_attn_implementation instance-attribute

_attn_implementation = _attn_implementation

embed_tokens instance-attribute

embed_tokens = VocabParallelEmbedding(
    vocab_size, hidden_size, org_num_embeddings=vocab_size
)

make_empty_intermediate_tensors instance-attribute

make_empty_intermediate_tensors = (
    make_empty_intermediate_tensors_factory(
        ["hidden_states"], hidden_size
    )
)

norm instance-attribute

norm = RMSNorm(hidden_size, eps=rms_norm_eps)

vocab_size instance-attribute

vocab_size = vocab_size

__init__

__init__(*, vllm_config: VllmConfig, prefix: str = '')

Source code in vllm/model_executor/models/arctic.py

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.vocab_size = config.vocab_size
    self.embed_tokens = VocabParallelEmbedding(
        self.vocab_size,
        config.hidden_size,
        org_num_embeddings=self.vocab_size)
    self.start_layer, self.end_layer, self.layers = make_layers(
        config.num_hidden_layers,
        lambda prefix: ArcticDecoderLayer(
            config, cache_config, quant_config, prefix=prefix),
        prefix=f"{prefix}.layers")
    self._attn_implementation = config._attn_implementation
    self.norm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps)
    self.make_empty_intermediate_tensors = (
        make_empty_intermediate_tensors_factory(["hidden_states"],
                                                config.hidden_size))

forward

forward(
    input_ids: Tensor,
    positions: Tensor,
    intermediate_tensors: Optional[IntermediateTensors],
    inputs_embeds: Optional[Tensor] = None,
) -> Union[Tensor, IntermediateTensors]

Source code in vllm/model_executor/models/arctic.py

def forward(
    self,
    input_ids: torch.Tensor,
    positions: torch.Tensor,
    intermediate_tensors: Optional[IntermediateTensors],
    inputs_embeds: Optional[torch.Tensor] = None,
) -> Union[torch.Tensor, IntermediateTensors]:
    if get_pp_group().is_first_rank:
        if inputs_embeds is not None:
            hidden_states = inputs_embeds
        else:
            hidden_states = self.get_input_embeddings(input_ids)
    else:
        assert intermediate_tensors is not None
        hidden_states = intermediate_tensors["hidden_states"]
    for layer in self.layers[self.start_layer:self.end_layer]:
        hidden_states = layer(positions, hidden_states)
    if not get_pp_group().is_last_rank:
        return IntermediateTensors({"hidden_states": hidden_states})
    hidden_states = self.norm(hidden_states)
    return hidden_states

get_input_embeddings

get_input_embeddings(input_ids: Tensor) -> Tensor

Source code in vllm/model_executor/models/arctic.py

def get_input_embeddings(self, input_ids: torch.Tensor) -> torch.Tensor:
    return self.embed_tokens(input_ids)