from typing import Optional import numpy as np import torch import torch.nn as nn import torch.nn.functional as F from vllm.config import (CUDAGraphMode, VllmConfig, get_layers_from_vllm_config, set_current_vllm_config) from vllm.forward_context import BatchDescriptor from vllm.logger import init_logger from vllm.model_executor.layers.attention_layer_base import AttentionLayerBase from vllm.model_executor.model_loader import get_model_loader from vllm.model_executor.model_loader.utils import \ process_weights_after_loading from vllm.model_executor.models.deepseek_mtp import DeepSeekMTP from vllm.model_executor.models.deepseek_v2 import DeepseekV32IndexerCache from vllm.model_executor.models.llama_eagle3 import Eagle3LlamaForCausalLM from vllm.utils import cdiv from vllm.v1.attention.backends.utils import (AttentionMetadataBuilder, CommonAttentionMetadata) from vllm.v1.core.sched.output import SchedulerOutput from vllm.v1.sample.metadata import SamplingMetadata from vllm.v1.spec_decode.metadata import SpecDecodeMetadata from vllm.v1.utils import CpuGpuBuffer from vllm.v1.worker.gpu_input_batch import CachedRequestState, InputBatch from vllm_ascend.ascend_forward_context import set_ascend_forward_context from vllm_ascend.attention.utils import (AscendCommonAttentionMetadata, AscendPrefillContextParallelMetadata) from vllm_ascend.spec_decode.interface import Proposer, SpecDcodeType from vllm_ascend.utils import (ProfileExecuteDuration, lmhead_tp_enable, vllm_version_is) if vllm_version_is("0.11.0"): from vllm.model_executor.model_loader.utils import set_default_torch_dtype from vllm.utils import is_pin_memory_available else: from vllm.utils.platform_utils import is_pin_memory_available from vllm.utils.torch_utils import set_default_torch_dtype logger = init_logger(__name__) PADDING_SLOT_ID = -1 class MtpProposer(Proposer): def __init__( self, vllm_config: VllmConfig, device, runner, ): self.name = SpecDcodeType.MTP self.vllm_config = vllm_config self.speculative_config = vllm_config.speculative_config assert self.speculative_config is not None self.draft_model_config = self.speculative_config.draft_model_config self.method = self.speculative_config.method self.runner = runner self.device = device self.dtype = vllm_config.model_config.dtype self.max_model_len = vllm_config.model_config.max_model_len self.block_size = vllm_config.cache_config.block_size self.num_speculative_tokens = self.speculative_config.num_speculative_tokens self.max_num_tokens = vllm_config.scheduler_config.max_num_batched_tokens self.token_arange_np = np.arange(self.max_num_tokens) # We need to get the hidden size from the draft model config because # the draft model's hidden size can be different from the target model's # hidden size (e.g., Llama 3.3 70B). self.hidden_size = self.draft_model_config.get_hidden_size() self.pcp_size = self.runner.pcp_size self.dcp_size = self.runner.dcp_size self.pcp_rank = self.runner.pcp_rank self.attn_metadata_builder: Optional[AttentionMetadataBuilder] = None self.draft_indexer_metadata_builder: Optional[ AttentionMetadataBuilder] = None self.attn_layer_names: list[str] = [] self.indexer_layer_names: list[str] = [] self.use_aclgraph = self.runner._use_aclgraph() self.cudagraph_batch_sizes = (list( reversed( self.vllm_config.compilation_config.cudagraph_capture_sizes)) if self.use_aclgraph else []) # persistent buffers for aclgraph graph self.input_ids = torch.zeros(self.max_num_tokens, dtype=torch.int32, device=device) self.uses_mrope = self.vllm_config.model_config.uses_mrope if self.uses_mrope: # M-RoPE need (3, max_num_tokens) self.mrope_positions = torch.zeros((3, self.max_num_tokens), dtype=torch.int64, device=device) else: # RoPE need (max_num_tokens,) self.positions = torch.zeros(self.max_num_tokens, dtype=torch.int64, device=device) self.hidden_states = torch.zeros( (self.max_num_tokens, self.hidden_size), dtype=self.dtype, device=device) self.full_indices = range( self.runner.max_num_tokens * self.pcp_size * self.dcp_size + self.pcp_size * self.dcp_size * self.runner.max_num_reqs) # We need +1 here because the arange is used to set query_start_loc, # which has one more element than batch_size. max_batch_size = vllm_config.scheduler_config.max_num_seqs max_num_slots_for_arange = max(max_batch_size + 1, self.max_num_tokens) self.arange = torch.arange(max_num_slots_for_arange, device=device, dtype=torch.int32) self.inputs_embeds = torch.zeros( (self.max_num_tokens, self.hidden_size), dtype=self.dtype, device=device) self.backup_next_token_ids = CpuGpuBuffer( max_batch_size, dtype=torch.int32, pin_memory=is_pin_memory_available(), device=device, with_numpy=True, ) self.use_sparse = hasattr(vllm_config.model_config.hf_config, "index_topk") def load_model(self, model) -> None: loader = get_model_loader(self.vllm_config.load_config) target_attn_layer_names = set( get_layers_from_vllm_config(self.vllm_config, AttentionLayerBase).keys()) target_indexer_layer_names = set( get_layers_from_vllm_config(self.vllm_config, DeepseekV32IndexerCache).keys()) draft_model_config = \ self.vllm_config.speculative_config.draft_model_config target_device = self.vllm_config.device_config.device with set_default_torch_dtype( draft_model_config.dtype), set_current_vllm_config( self.vllm_config): self.model = DeepSeekMTP( vllm_config=self.vllm_config).to(target_device) draft_attn_layer_names = (get_layers_from_vllm_config( self.vllm_config, AttentionLayerBase).keys() - target_attn_layer_names) indexer_layers = get_layers_from_vllm_config(self.vllm_config, DeepseekV32IndexerCache) draft_indexer_layer_names = indexer_layers.keys( ) - target_indexer_layer_names # NOTE: Currently we don't have specific attention backend and attention metadata # for deepseek v3.2 indexer, so we just exclude the indexer layers here. draft_attn_layer_names = draft_attn_layer_names - draft_indexer_layer_names assert len(draft_attn_layer_names) == 1 self.attn_layer_name = list(draft_attn_layer_names) self.model.load_weights( loader.get_all_weights( self.vllm_config.speculative_config.draft_model_config, self.model)) process_weights_after_loading(self.model, draft_model_config, target_device) @torch.inference_mode() def dummy_run(self, num_tokens: int, with_prefill: bool = False, skip_attn: bool = False, num_reqs: int = 0, num_tokens_across_dp=None, aclgraph_runtime_mode: CUDAGraphMode = CUDAGraphMode.NONE, batch_descriptor=None) -> None: ( num_tokens, num_tokens_across_dp, with_prefill, ) = self.runner._sync_metadata_across_dp(num_tokens, with_prefill) moe_comm_type = self.runner._select_moe_comm_method( num_tokens, with_prefill) attn_metadata = None input_ids = self.input_ids[:num_tokens] positions = self.positions[:num_tokens] previous_hidden_states = self.hidden_states[:num_tokens] for _ in range(self.num_speculative_tokens): with set_ascend_forward_context( attn_metadata, self.vllm_config, num_tokens=num_tokens, with_prefill=with_prefill, num_tokens_across_dp=num_tokens_across_dp, reserved_mc2_mask=self.runner.reserved_mc2_mask, moe_comm_type=moe_comm_type, in_profile_run=self.runner.in_profile_run, num_actual_tokens=0, aclgraph_runtime_mode=aclgraph_runtime_mode, batch_descriptor=batch_descriptor): self.model(input_ids=input_ids, positions=positions, hidden_states=previous_hidden_states) if with_prefill: break def generate_token_ids(self, sampled_token_ids: list[list[int]], sampling_metadata: SamplingMetadata = None, scheduler_output: SchedulerOutput = None, spec_decode_metadata: SpecDecodeMetadata = None, positions: torch.Tensor = None, num_scheduled_tokens: int = 0, hidden_states: torch.Tensor = None, attn_metadata=None, aux_hidden_states: torch.Tensor = None): common_attn_metadata = self.runner.spec_decode_common_attn_metadata if attn_metadata is not None and isinstance(attn_metadata, dict): attn_metadata = attn_metadata['model.layers.0.self_attn.attn'] if self.speculative_config.disable_padded_drafter_batch: # When padded-batch is disabled, the sampled_token_ids should be # the cpu-side list[list[int]] of valid sampled tokens for each # request, with invalid requests having empty lists. assert isinstance(sampled_token_ids, list), \ "sampled_token_ids should be a python list when" \ "padded-batch is disabled." next_token_ids = self.prepare_next_token_ids_cpu( sampled_token_ids, self.runner.requests, self.runner.input_batch, scheduler_output.num_scheduled_tokens) else: # When using padded-batch, the sampled_token_ids should be # the gpu tensor of sampled tokens for each request, of shape # (num_reqs, num_spec_tokens + 1) with rejected tokens having # value -1. assert isinstance(sampled_token_ids, torch.Tensor), \ "sampled_token_ids should be a torch.Tensor when" \ "padded-batch is enabled." next_token_ids, valid_sampled_tokens_count = \ self.prepare_next_token_ids_padded( common_attn_metadata, sampled_token_ids, self.runner.requests, self.runner.input_batch, self.runner.discard_request_indices.gpu, self.runner.num_discarded_requests ) is_prefill = len(scheduler_output.scheduled_new_reqs) > 0 req_scheduled_tokens = scheduler_output.num_scheduled_tokens long_seq_metadata: AscendPrefillContextParallelMetadata = \ self.runner.long_seq_metadata if self.pcp_size > 1 else None if spec_decode_metadata is None: # update pcp related params if self.pcp_size > 1 and is_prefill: token_indices_to_sample = None target_token_ids = self.runner.input_ids_pcp_full[: num_scheduled_tokens] target_positions = positions[:num_scheduled_tokens] target_hidden_states = hidden_states else: token_indices_to_sample = None # input_ids can be None for multimodal models. target_token_ids = self.runner.input_ids[:num_scheduled_tokens] target_positions = positions[:num_scheduled_tokens] target_hidden_states = hidden_states[:num_scheduled_tokens] else: if self.speculative_config.disable_padded_drafter_batch: token_indices_to_sample = None common_attn_metadata, token_indices =\ self._prepare_inputs( common_attn_metadata, sampled_token_ids, spec_decode_metadata.num_draft_tokens) else: common_attn_metadata, token_indices, \ token_indices_to_sample =\ self.prepare_inputs_padded( common_attn_metadata, spec_decode_metadata, valid_sampled_tokens_count) target_token_ids = self.runner.input_ids[token_indices] target_positions = positions[token_indices] target_hidden_states = hidden_states[token_indices] draft_token_ids = self._propose( target_token_ids=target_token_ids, target_positions=target_positions, target_hidden_states=target_hidden_states, next_token_ids=next_token_ids, last_token_indices=token_indices_to_sample, common_attn_metadata=common_attn_metadata, sampling_metadata=sampling_metadata, is_prefill=is_prefill, req_scheduled_tokens=req_scheduled_tokens, long_seq_metadata=long_seq_metadata, ) return draft_token_ids def _prepare_inputs( self, common_attn_metadata: CommonAttentionMetadata, sampled_token_ids: list[list[int]], num_draft_tokens: list[int], ) -> tuple[CommonAttentionMetadata, torch.Tensor]: """ This function is used to prepare the inputs for speculative decoding. It updates to the common_attn_metadata to account for the rejected tokens (and newly sampled tokens). It also returns the token indices of the tokens that should be fed to the speculator. """ # E.g. # common_attn_metadata.query_start_loc{_cpu}: # [0, q1, q1 + q2, q1 + q2 + q3] # common_attn_metadata.seq_lens{_cpu}: [s1, s2, s3] # num_rejected_tokens: [n1, n2, n3] # This function computes the intermediate values: # num_tokens_per_req: [q1 - n1, q2 - n2, q3 - n3] # And returns: # common_attn_metadata.query_start_loc{_cpu}: # [0, q1 - n1, q1 + q2 - n1 - n2, q1 + q2 + q3 - n1 - n2 - n3] # common_attn_metadata.seq_lens{_cpu}: # [s1 - n1 + 1, s2 - n2 + 1, s3 - n3 + 1] # token_indices: [0, 1, ..., q1 - n1 - 1, # q1, q1 + 1, ..., q1 + q2 - n2 - 1, # q1 + q2, q1 + q2 + 1, ..., q1 + q2 + q3 - n3 - 1] num_rejected_tokens = [ n + 1 - len(sampled_token_ids[i]) if n > 0 else 0 for i, n in enumerate(num_draft_tokens) ] num_rejected_tokens = torch.tensor(num_rejected_tokens, dtype=torch.int32) device = common_attn_metadata.query_start_loc.device query_start_loc_cpu = common_attn_metadata.query_start_loc_cpu new_seq_lens_cpu = common_attn_metadata.seq_lens_cpu - num_rejected_tokens # [0, q1, q1 + q2, q1 + q2 + q3] -> [q1, q2, q3] new_query_len_per_req = query_start_loc_cpu[ 1:] - query_start_loc_cpu[:-1] # [q1, q2, q3] -> [q1 - n1, q2 - n2, q3 - n3] new_num_tokens_per_req = new_query_len_per_req - num_rejected_tokens new_num_tokens_per_req_np = new_num_tokens_per_req.numpy() # [q1 - n1, q2 - n2, q3 - n3] -> # [0, q1 - n1, q1 + q2 - n1 - n2, q1 + q2 + q3 - n1 - n2 - n3] new_query_start_loc_cpu = torch.zeros( query_start_loc_cpu.shape, dtype=torch.int32, pin_memory=is_pin_memory_available(), ) new_query_start_loc_np = new_query_start_loc_cpu.numpy() np.cumsum(new_num_tokens_per_req_np, out=new_query_start_loc_np[1:]) total_num_tokens = new_query_start_loc_np[-1] # Example assuming num_tokens_per_req_np = [2, 4, 3] # this implies that `new_query_start_locs` is: # [0, 2, 6, 9] -> # [0, 0, 2, 2, 2, 2, 6, 6, 6] # _r1_ ____r2____ ___r3__ new_query_start_locs_expanded = np.repeat(new_query_start_loc_np[:-1], new_num_tokens_per_req_np) # [0, 1, 2, 3, 4, 5, 6, 7, 8] -> # [0, 1, 0, 1, 2, 3, 0, 1, 2] # _r1_ ____r2____ ___r3__ token_offests = (self.token_arange_np[:total_num_tokens] - new_query_start_locs_expanded) # Expand starting positions to match token pattern # [0, q1, q1 + q2] -> # [0, 0, q1, q1, q1, q1, q1 + q2, q1 + q2, q1 + q2] # _r1_ _____r2_______ ___________r3____________ old_query_start_locs_expanded = np.repeat( query_start_loc_cpu[:-1].numpy(), new_num_tokens_per_req_np) # Final token indices are: # [0, 1, // req 1 # q1 + 0, q1 + 1, q1 + 2, q1 + 3, // req 2 # q1 + q2 + 0, q1 + q2 + 1, q1 + q2 + 2] // req 3 token_indices_np = token_offests + old_query_start_locs_expanded token_indices = torch.from_numpy(token_indices_np).to( device, non_blocking=True) spec_common_attn_metadata = AscendCommonAttentionMetadata( query_start_loc=new_query_start_loc_cpu.to(device, non_blocking=True), query_start_loc_cpu=new_query_start_loc_cpu, seq_lens=new_seq_lens_cpu.to(device, non_blocking=True), seq_lens_cpu=new_seq_lens_cpu, num_computed_tokens_cpu=common_attn_metadata. num_computed_tokens_cpu, num_reqs=common_attn_metadata.num_reqs, num_actual_tokens=total_num_tokens, max_query_len=new_query_len_per_req.max().item(), block_table_tensor=common_attn_metadata.block_table_tensor, slot_mapping=common_attn_metadata.slot_mapping[token_indices], actual_seq_lengths_q=self.runner.actual_seq_lengths_q, positions=common_attn_metadata.positions[token_indices], attn_mask=self.runner.attn_mask, spec_attn_mask=self.runner.spec_attn_mask, attn_state=self.runner.attn_state, graph_pad_size=self.runner.graph_pad_size, decode_token_per_req=self.runner.decode_token_per_req, ) return spec_common_attn_metadata, token_indices def _propose( self, # [num_tokens] target_token_ids: torch.Tensor, # [num_tokens] or [3, num_tokens] when M-RoPE is enabled target_positions: torch.Tensor, # [num_tokens, hidden_size] target_hidden_states: torch.Tensor, # [batch_size] next_token_ids: torch.Tensor, last_token_indices: Optional[torch.Tensor], common_attn_metadata: CommonAttentionMetadata, sampling_metadata: SamplingMetadata, mm_embed_inputs: Optional[tuple[list[torch.Tensor], torch.Tensor]] = None, is_prefill=False, req_scheduled_tokens=None, long_seq_metadata=None, ) -> torch.Tensor: num_tokens = target_token_ids.shape[0] batch_size = next_token_ids.shape[0] if last_token_indices is None: last_token_indices = common_attn_metadata.query_start_loc[1:] - 1 if self.method == "eagle3": assert isinstance(self.model, Eagle3LlamaForCausalLM) target_hidden_states = self.model.combine_hidden_states( target_hidden_states) assert target_hidden_states.shape[-1] == self.hidden_size # Shift the input ids by one token. # E.g., [a1, b1, b2, c1, c2, c3] -> [b1, b2, c1, c2, c3, c3] self.input_ids[:num_tokens - 1] = target_token_ids[1:] # Replace the last token with the next token. # E.g., [b1, b2, c1, c2, c3, c3] -> [a2, b2, b3, c2, c3, c4] self.input_ids[last_token_indices] = next_token_ids # update pcp related params if self.pcp_size > 1 and is_prefill: num_tokens, input_ids, target_hidden_states, max_query_len, seq_lens, cu_num_tokens = \ self._split_pcp_input(req_scheduled_tokens, num_tokens, target_hidden_states) # graph mode padding not considered now num_input_tokens = num_tokens self.input_ids[:num_input_tokens].copy_(input_ids) common_attn_metadata.prefill_context_parallel_metadata = long_seq_metadata common_attn_metadata.num_actual_tokens = num_tokens common_attn_metadata.max_query_len = max_query_len common_attn_metadata.seq_lens_cpu = seq_lens.cpu() common_attn_metadata.query_start_loc = \ cu_num_tokens[:batch_size + 1] common_attn_metadata.query_start_loc_cpu = \ cu_num_tokens[:batch_size + 1].cpu() assert self.runner is not None builder = self.runner.attn_groups[0][0].get_metadata_builder() attn_metadata_mtp = builder.build(0, common_attn_metadata, self.runner.get_model()) attn_metadata = {} for layer_name in self.attn_layer_name: attn_metadata[layer_name] = attn_metadata_mtp if self.use_aclgraph and num_tokens <= self.cudagraph_batch_sizes[-1]: # Acl graph mode, add padding to the batch size num_input_tokens = self.vllm_config.pad_for_cudagraph(num_tokens) else: # Eager mode, no padding needed num_input_tokens = num_tokens # copy inputs to buffer for cudagraph self.positions[:num_tokens] = target_positions self.hidden_states[:num_tokens] = target_hidden_states # eager/acl piecewise mode need to update num_tokens_across_dp (num_input_tokens, num_tokens_across_dp, with_prefill) = self.runner._sync_metadata_across_dp( num_input_tokens, self.runner.with_prefill) moe_comm_type = self.runner._select_moe_comm_method( num_input_tokens, with_prefill) batch_descriptor = BatchDescriptor(num_tokens=num_input_tokens, uniform_decode=False) aclgraph_runtime_mode, batch_descriptor = \ self.runner.aclgraph_dispatcher.dispatch(batch_descriptor) if aclgraph_runtime_mode not in [ CUDAGraphMode.PIECEWISE, CUDAGraphMode.NONE ]: # Fallback to piecewise graph, when acl full graph is enabled logger.debug( "Currently the eagle proposer only supports cudagraph_mode " f"PIECEWISE, and is forced to set graph mode from {aclgraph_runtime_mode} " "to CUDAGraphMode.PIECEWISE") aclgraph_runtime_mode = CUDAGraphMode.PIECEWISE for step in range(self.num_speculative_tokens): with set_ascend_forward_context( attn_metadata, self.vllm_config, num_tokens=num_input_tokens, with_prefill=with_prefill, num_tokens_across_dp=num_tokens_across_dp, reserved_mc2_mask=self.runner.reserved_mc2_mask, moe_comm_type=moe_comm_type, aclgraph_runtime_mode=aclgraph_runtime_mode, batch_descriptor=batch_descriptor, in_profile_run=self.runner.in_profile_run, num_actual_tokens=num_tokens): with ProfileExecuteDuration().capture_async('mtp_forward'): model_kwargs = {} model_kwargs["attn_metadata"] = attn_metadata hidden_states = self.model( input_ids=self.input_ids[:num_input_tokens], positions=self.positions[:num_input_tokens], hidden_states=self.hidden_states[:num_input_tokens]) num_indices = last_token_indices.shape[0] if lmhead_tp_enable(): if not self.runner.with_prefill: max_num_reqs_across_dp = num_input_tokens else: max_num_reqs_across_dp = self.vllm_config.scheduler_config.max_num_seqs last_token_indices = nn.functional.pad( last_token_indices, (0, max_num_reqs_across_dp - num_indices)) sample_hidden_states = hidden_states[last_token_indices] logits = self.model.compute_logits(sample_hidden_states) if lmhead_tp_enable() and num_indices < logits.shape[0]: logits = logits[:num_indices] draft_token_ids = logits.argmax(dim=-1) if self.num_speculative_tokens == 1: # [batch_size, 1] return draft_token_ids.view(-1, 1) if step == 0: draft_token_ids_list = [draft_token_ids] else: draft_token_ids_list.append(draft_token_ids) # prepare next mtp inputs # mtp>1: prefill skip or decode skip last loop if with_prefill: for _ in range(self.num_speculative_tokens - 1): draft_token_ids_list.append(draft_token_ids) if step == self.num_speculative_tokens - 1 or with_prefill: break attn_metadata_i = attn_metadata[self.attn_layer_name[0]] if step == 0: positions = target_positions[last_token_indices] hidden_states = hidden_states[last_token_indices] slot_mapping = attn_metadata_i.slot_mapping[last_token_indices] attn_metadata_i.slot_mapping.fill_(-1) attn_metadata_i.query_start_loc = self.arange[:batch_size + 1] last_token_indices = self.arange[:batch_size] if attn_metadata_i.num_decode_tokens != 0: attn_metadata_i.num_decode_tokens = batch_size input_ids = draft_token_ids_list[-1].int() positions += 1 attn_metadata_i.decode.actual_seq_lengths_q = attn_metadata_i.query_start_loc[ 1:batch_size + 1].tolist() attn_metadata_i.decode.cos = builder.cos_cache[ positions].unsqueeze(1).unsqueeze(2) attn_metadata_i.decode.sin = builder.sin_cache[ positions].unsqueeze(1).unsqueeze(2) # NOTE(woosuk): We should handle the case where the draft model # generates tokens beyond the max model length. Since it is complex # to remove such requests from the batch, we keep them in the batch # but adjust the position ids and slot mappings to avoid the # out-of-range access during the model execution. The draft tokens # generated with this adjustment should be ignored. exceeds_max_model_len = positions >= self.runner.model_config.max_model_len # Mask out the position ids that exceed the max model length. # Otherwise, we may get out-of-range error in RoPE. clamped_positions = torch.where(exceeds_max_model_len, 0, positions) # Increment the sequence lengths. attn_metadata_i.seq_lens[:batch_size] += 1 # For the requests that exceed the max model length, we set the # sequence length to 1 to minimize their overheads in attention. exceeds_max_model_len_cpu = exceeds_max_model_len.to( attn_metadata_i.seq_lens.device, non_blocking=True) attn_metadata_i.seq_lens[:batch_size].masked_fill_( exceeds_max_model_len_cpu, 1) # Mask out the slot mappings that exceed the max model length. # Otherwise, the KV cache will be inadvertently updated with the # padding tokens. slot_mapping += 1 slot_mapping.masked_fill_(exceeds_max_model_len, PADDING_SLOT_ID) # copy inputs to buffer for cudagraph self.input_ids[:batch_size] = input_ids self.positions[:batch_size] = clamped_positions self.hidden_states[:hidden_states.shape[0]] = hidden_states attn_metadata_i.slot_mapping[:batch_size] = slot_mapping if attn_metadata_i.prefill is not None: attn_metadata_i.prefill.seq_lens = attn_metadata_i.seq_lens attn_metadata_i.prefill.seq_lens_list = attn_metadata_i.prefill.seq_lens.tolist( ) attn_metadata_i.prefill.context_lens = attn_metadata_i.seq_lens attn_metadata_i.prefill.input_positions = self.positions[: num_input_tokens] attn_metadata_i.prefill.max_seq_lens += 1 attn_metadata_i.prefill.max_seq_lens = min( attn_metadata_i.prefill.max_seq_lens, self.runner.model_config.max_model_len) if attn_metadata_i.decode is not None: attn_metadata_i.decode.seq_lens = attn_metadata_i.seq_lens attn_metadata_i.decode.seq_lens_list = attn_metadata_i.decode.seq_lens.tolist( ) attn_metadata_i.decode.input_positions = self.positions[: num_input_tokens] attn_metadata_i.decode.max_seq_lens += 1 attn_metadata_i.decode.max_seq_lens = min( attn_metadata_i.decode.max_seq_lens, self.runner.model_config.max_model_len) # mtp>1: [batch_size, k] draft_token_ids = torch.stack(draft_token_ids_list, dim=1) return draft_token_ids # TODO Using torch instead of triton may result in poor performance def _prepare_input_kernel(self, out_ptr: torch.Tensor, cu_query_lens: torch.Tensor, cu_num_tokens: torch.Tensor, block_size: int): device = cu_query_lens.device dtype = out_ptr.dtype offsets = torch.arange(block_size, device=device, dtype=dtype) start_pos = cu_num_tokens[:-1] end_pos = cu_num_tokens[1:] num_tokens = end_pos - start_pos global_indices = (start_pos.view(-1, 1) + offsets.view(1, -1)) values = (cu_query_lens[:-1].view(-1, 1) + offsets.view(1, -1)) mask = (offsets.view(1, -1) < num_tokens.view(-1, 1)) global_indices_flat = global_indices[mask] values_flat = values[mask] out_ptr[global_indices_flat] = values_flat def prepare_next_token_ids_cpu( self, sampled_token_ids: list[list[int]], requests: dict[str, CachedRequestState], gpu_input_batch: InputBatch, num_scheduled_tokens: dict[str, int], ) -> torch.Tensor: """ This function is used to prepare the inputs for speculative decoding. It calculates the next token ids for each request based on the sampled token ids from the CPU. If a request has no sampled token ids (e.g., during the initial decoding steps), it falls back to using the request state to get the next token id. """ req_ids = gpu_input_batch.req_ids next_token_ids: list[int] = [] for i, token_ids in enumerate(sampled_token_ids): if token_ids: # Common case. next_token_id = token_ids[-1] else: # Partial prefill (rare case). # Get the next token id from the request state. req_id = req_ids[i] req_state = requests[req_id] seq_len = req_state.num_computed_tokens + num_scheduled_tokens[ req_id] next_token_id = req_state.get_token_id(seq_len) next_token_ids.append(next_token_id) next_token_ids = torch.tensor(next_token_ids, dtype=torch.int32, device=self.input_ids.device) return next_token_ids def prepare_next_token_ids_padded( self, common_attn_metadata: CommonAttentionMetadata, sampled_token_ids: torch.Tensor, requests: dict[str, CachedRequestState], gpu_input_batch: InputBatch, discard_request_indices: torch.Tensor, num_discarded_requests: int, ) -> tuple[torch.Tensor, torch.Tensor]: """ This function is used to prepare the inputs for speculative decoding. It calculates the next token ids and the number of valid sampled tokens for each request, considering the "discarded" requests whose next token is not sampled and comes from `request.get_token_id()` instead. It also accounts for the rejected tokens in `sampled_token_ids`. This function must use device functions to operate on the inputs, and should not introduce any blocking CPU-GPU synchronization. """ # TODO(Ben): Combine this into a custom fused kernel # Precompute get_token_id for when there is no valid next token num_reqs = gpu_input_batch.num_reqs self.backup_next_token_ids.np[:num_reqs] = np.array([ requests[gpu_input_batch.req_ids[i]].get_token_id( common_attn_metadata.seq_lens_cpu[i].item()) for i in range(num_reqs) ]) self.backup_next_token_ids.copy_to_gpu(num_reqs) # Mask out the sampled tokens indices that should not be sampled. discard_sampled_tokens_req_indices = discard_request_indices[: num_discarded_requests] valid_sampled_token_ids_gpu = sampled_token_ids.clone() valid_sampled_token_ids_gpu.index_fill_( 0, discard_sampled_tokens_req_indices, -1) # Generate a mask for all valid tokens within those requests valid_mask = (valid_sampled_token_ids_gpu != -1) & ( valid_sampled_token_ids_gpu < gpu_input_batch.vocab_size) # Count the number of valid tokens in each request valid_sampled_tokens_count = valid_mask.sum(dim=1) # Get the rightmost valid index per row last_valid_indices = valid_sampled_tokens_count - 1 last_valid_indices_safe = torch.clamp(last_valid_indices, min=0) # Get last valid token from each row # (assume undefined state where there is no valid token) selected_tokens = torch.gather( valid_sampled_token_ids_gpu, 1, last_valid_indices_safe.unsqueeze(1)).squeeze(1) # Use last token if valid, pre-computed backup if not batch_size = valid_sampled_token_ids_gpu.shape[0] next_token_ids = torch.where( last_valid_indices != -1, selected_tokens, self.backup_next_token_ids.gpu[:batch_size], ) return next_token_ids, valid_sampled_tokens_count def prepare_inputs_padded( self, common_attn_metadata: CommonAttentionMetadata, spec_decode_metadata: SpecDecodeMetadata, valid_sampled_tokens_count: torch.Tensor, ) -> tuple[CommonAttentionMetadata, torch.Tensor, torch.Tensor]: """ This function is used to prepare the inputs for speculative decoding It updates the common_attn_metadata for speculative decoding, but does not consider the rejected tokens. Instead, all tokens are included as inputs to the speculator, with the rejected tokens used as padding and filtered out later by `token_indices_to_sample`. No blocking CPU operations should be introduced in this function. """ num_draft_tokens_gpu = torch.cat([ spec_decode_metadata.cu_num_draft_tokens[0:1], spec_decode_metadata.cu_num_draft_tokens[1:] - spec_decode_metadata.cu_num_draft_tokens[:-1], ]) num_rejected_tokens_gpu = torch.where( num_draft_tokens_gpu > 0, num_draft_tokens_gpu + 1 - valid_sampled_tokens_count, torch.zeros_like(num_draft_tokens_gpu), ) query_start_loc_cpu = common_attn_metadata.query_start_loc_cpu new_query_len_per_req = query_start_loc_cpu[ 1:] - query_start_loc_cpu[:-1] total_num_tokens = query_start_loc_cpu[-1].item() token_indices = self.arange[:total_num_tokens] spec_common_attn_metadata = AscendCommonAttentionMetadata( query_start_loc=common_attn_metadata.query_start_loc, query_start_loc_cpu=query_start_loc_cpu, seq_lens_cpu=common_attn_metadata.seq_lens, num_reqs=common_attn_metadata.num_reqs, num_actual_tokens=total_num_tokens, max_query_len=new_query_len_per_req.max().item(), actual_seq_lengths_q=self.runner.actual_seq_lengths_q, block_table_tensor=common_attn_metadata.block_table_tensor, slot_mapping=common_attn_metadata.slot_mapping, positions=common_attn_metadata.positions, attn_mask=self.runner.attn_mask, spec_attn_mask=self.runner.spec_attn_mask, attn_state=self.runner.attn_state, graph_pad_size=self.runner.graph_pad_size, decode_token_per_req=self.runner.decode_token_per_req, num_computed_tokens_cpu=common_attn_metadata. num_computed_tokens_cpu, seq_lens=common_attn_metadata.seq_lens) token_indices_to_sample = (common_attn_metadata.query_start_loc[1:] - 1 - num_rejected_tokens_gpu) return spec_common_attn_metadata, token_indices, token_indices_to_sample def _split_pcp_input(self, req_scheduled_tokens, num_tokens, target_hidden_states): """ Split input_ids and target_hidden_states in pcp group. 1. input_ids padding: [t0, t1, t2, t3, t4, t5] -> [t0, t1, t2, t3, t4, t5, pad, pad] 2. split input_ids: pcp0 [t0, t1, pad, pad], pcp1 [t2, t3, t4, t5] 3. split target_hidden_states (already include cp padding): [h0, h1, h2, h3, h4, h5, pad, pad] -> pcp0 [h0, h1, pad, pad], pcp1 [h2, h3, h4, h5] 4. also update max_query_len, seq_lens, cu_num_tokens according to pcp split. """ def _pcp_pad_and_split(num_tokens): num_pcp_padded_scheduled_tokens = cdiv( num_tokens, 2 * self.pcp_size) * 2 * self.pcp_size pcp_pad = num_pcp_padded_scheduled_tokens - num_tokens chunk_size = num_pcp_padded_scheduled_tokens // (2 * self.pcp_size) # split position_ids (and use split position_ids to split input_ids afterwards) req_position_cp: list[int] = [] req_position_cp.extend( self.full_indices[self.pcp_rank * chunk_size:(self.pcp_rank + 1) * chunk_size]) req_position_cp.extend( self.full_indices[num_pcp_padded_scheduled_tokens - (self.pcp_rank + 1) * chunk_size:num_pcp_padded_scheduled_tokens - self.pcp_rank * chunk_size]) return req_position_cp, num_pcp_padded_scheduled_tokens, pcp_pad num_pcp_scheduled_tokens = [] input_ids_list = self.input_ids[:num_tokens] ori_start_index = 0 pad_start_index = 0 pcp_split_input_ids_list = [] pcp_split_hidden_states_list = [] for ori_num_tokens in req_scheduled_tokens.values(): req_position_pcp, num_pcp_padded_scheduled_tokens, num_pcp_pad = \ _pcp_pad_and_split(ori_num_tokens) actual_num_tokens = len(req_position_pcp) num_pcp_scheduled_tokens.append(actual_num_tokens) pad_input_ids = F.pad( input_ids_list[ori_start_index:ori_start_index + ori_num_tokens], (0, num_pcp_pad)) ori_start_index += ori_num_tokens pcp_chunk_indices = [ pad_start_index + pos for pos in req_position_pcp ] pcp_split_input_ids = pad_input_ids[req_position_pcp] pcp_split_hidden_states = target_hidden_states[pcp_chunk_indices] pcp_split_input_ids_list.append(pcp_split_input_ids) pcp_split_hidden_states_list.append(pcp_split_hidden_states) pad_start_index += num_pcp_padded_scheduled_tokens num_tokens = sum(num_pcp_scheduled_tokens) input_ids = torch.cat(pcp_split_input_ids_list) target_hidden_states = torch.cat(pcp_split_hidden_states_list, dim=0) max_query_len = max(num_pcp_scheduled_tokens) seq_lens = torch.tensor(num_pcp_scheduled_tokens, dtype=torch.int32) cu_num_tokens = torch.tensor( np.insert(np.cumsum(np.array(num_pcp_scheduled_tokens)), 0, 0)) return num_tokens, input_ids, target_hidden_states, max_query_len, seq_lens, cu_num_tokens