[Main2Main] Upgrade vllm commit to 0123 (#6169)

### What this PR does / why we need it? 1. ✅ Upgrade vllm commit to: 0115 (8471b27df97c3eb79f891802fc0e858f8f7ac6a0) Modify import paths due to the refactors： https://github.com/vllm-project/vllm/pull/32245 https://github.com/vllm-project/vllm/pull/32060 Test result: https://github.com/vllm-project/vllm-ascend/actions/runs/21034239336/job/60490156965?pr=5913 2. ✅Upgrade vllm commit to: 0119 (9a1f16da1e423ede2c2f52a9850cbfbb39cefe96) Fix `WorkerProc.__init__() missing 1 required positional argument: 'is_driver_worker'` due to https://github.com/vllm-project/vllm/pull/28506 Test result: https://github.com/vllm-project/vllm-ascend/actions/runs/21156263050/job/60841668755?5569 3. ✅Upgrade vllm commit to: 0120(148117ea2e689cd43df4be6892671a17cdae5833) 1. Add `skip_compiled` param in `set_forward_context` due to https://github.com/vllm-project/vllm/pull/30385 2. Modify `tests/ut/spec_decode/test_eagle_proposer.py` due to https://github.com/vllm-project/vllm/pull/24322 change `self.max_num_tokens = vllm_config.scheduler_config.max_num_batched_tokens + max_batch_size` 3. Modify UT import paths due to the refactors：https://github.com/vllm-project/vllm/pull/32060 Test result: https://github.com/vllm-project/vllm-ascend/actions/runs/21204851770/job/60999046946 4. ✅Upgrade vllm commit to: 0121(f23fb5a7c1b61350c5c40ca1115d3bf8cf2b8cc9) 1. vLLM switched `uses_mrope` from target to draft model config, making `positions`/`mrope_positions` mutually exclusive, breaking vllm-ascend's direct self.positions access and tests missing `draft_model_config.uses_mrope`. https://github.com/vllm-project/vllm/pull/32048 2. Moved bs_to_padded_graph_size from CompilationConfig to CudagraphDispatcher due to the refactor https://github.com/vllm-project/vllm/pull/30143 3. Remove unused `maybe_setup_kv_connector` due to https://github.com/vllm-project/vllm/pull/32077 Test result: https://github.com/vllm-project/vllm-ascend/actions/runs/21217728738/job/61043738834 6. ✅Upgrade vllm commit to: 0122(8ebf271bb6d1e7e9b1a55be73d755ef1a57dbbe5) Updating FusedMoEParallelConfig (added enable_eplb) and FusedMoEConfig due to https://github.com/vllm-project/vllm/pull/32414 Test result: https://github.com/vllm-project/vllm-ascend/actions/runs/21249922546/job/61148613054 8. ✅Upgrade vllm commit to: 0123(dc917cceb877dfd13f98c538c4c96158047d98bd) Setting temperature=0.0 due to the removal of the default temperature value in https://github.com/vllm-project/vllm/pull/32723 Test result: https://github.com/vllm-project/vllm-ascend/actions/runs/21280796875 ### Does this PR introduce _any_ user-facing change? ### How was this patch tested? - vLLM version: v0.14.0 - vLLM main: d68209402d --------- Signed-off-by: wjunLu <wjunlu217@gmail.com> Signed-off-by: Meihan-chen <jcccx.cmh@gmail.com> Co-authored-by: wjunLu <wjunlu217@gmail.com>
2026-01-27 08:44:36 +08:00
parent 9780a995e1
commit fea197ad50
25 changed files with 173 additions and 83 deletions
--- a/vllm_ascend/worker/model_runner_v1.py
+++ b/vllm_ascend/worker/model_runner_v1.py
@@ -103,7 +103,7 @@ from vllm_ascend.utils import (AscendDeviceType, ProfileExecuteDuration,
                               enable_sp, get_ascend_device_type,
                               is_drafter_moe_model, is_moe_model,
                               lmhead_tp_enable, maybe_trans_nz,
-                               set_weight_prefetch_method)
+                               set_weight_prefetch_method, vllm_version_is)
 from vllm_ascend.worker.npu_input_batch import NPUInputBatch
 from vllm_ascend.worker.pcp_utils import PCPManager

@@ -587,8 +587,12 @@ class NPUModelRunner(GPUModelRunner):
        if (self.use_aclgraph and total_num_scheduled_tokens
                <= self.cudagraph_batch_sizes[-1]):
            # Add padding to the batch size.
-            num_input_tokens = self.vllm_config.pad_for_cudagraph(
-                total_num_scheduled_tokens)
+            if vllm_version_is('0.14.1'):
+                num_input_tokens = self.vllm_config.pad_for_cudagraph(
+                    total_num_scheduled_tokens)
+            else:
+                num_input_tokens = self.cudagraph_dispatcher._bs_to_padded_graph_size[
+                    total_num_scheduled_tokens]
        elif self.use_aclgraph and enable_sp(self.vllm_config):
            # When using aclgraph, if total_num_scheduled_tokens exceeds the maximum graph size,
            # the model will fall back to running its FX graph in eager mode.
@@ -1403,9 +1407,17 @@ class NPUModelRunner(GPUModelRunner):
                head_dim=self.model_config.get_vocab_size(),
                generators=self.input_batch.sampling_metadata.generators)

+        # Encoder-decoder models can only compile the pure decode steps where no
+        # encoder inputs are present. Use eager for the first pass.
+        num_encoder_reqs = len(scheduler_output.scheduled_encoder_inputs)
+        has_encoder_input = (
+            self.model_config.is_encoder_decoder and num_encoder_reqs > 0
+        )
+
        # Run forward pass
        with ProfileExecuteDuration().capture_async("forward"):
-            with set_ascend_forward_context(
+            with (
+                set_ascend_forward_context(
                    attn_metadata,
                    self.vllm_config,
                    num_tokens=num_input_tokens,
@@ -1414,26 +1426,18 @@ class NPUModelRunner(GPUModelRunner):
                    batch_descriptor=batch_descriptor,
                    num_actual_tokens=scheduler_output.
                    total_num_scheduled_tokens,
-                    model_instance=self.model):
-                self.maybe_setup_kv_connector(scheduler_output)
-
+                    model_instance=self.model,
+                    skip_compiled=has_encoder_input),
+                self.maybe_get_kv_connector_output(scheduler_output) as kv_connector_output,
+            ):
                hidden_states = self._generate_process_reqs_hidden_states(
                    num_input_tokens, input_ids, positions,
                    intermediate_tensors, inputs_embeds, model_kwargs)

-            self.maybe_wait_for_kv_save()
-            finished_sending, finished_recving = self.get_finished_kv_transfer(
-                scheduler_output)
-
            aux_hidden_states = None
            if self.use_aux_hidden_state_outputs:
                hidden_states, aux_hidden_states = hidden_states

-        kv_connector_output = KVConnectorOutput(
-            finished_sending=finished_sending,
-            finished_recving=finished_recving)
-        finished_sending = None
-        finished_recving = None
        with ProfileExecuteDuration().capture_async("post process"):
            # Broadcast PP output for external_launcher (torchrun)
            # to make sure we are synced across pp ranks
--- a/vllm_ascend/worker/v2/aclgraph_utils.py
+++ b/vllm_ascend/worker/v2/aclgraph_utils.py
@@ -22,7 +22,6 @@ from typing import Any
 import torch
 import torch.nn as nn
 from vllm.config import VllmConfig
-from vllm.v1.attention.backends.utils import AttentionMetadataBuilder
 from vllm.v1.kv_cache_interface import KVCacheConfig
 from vllm.v1.worker.gpu.block_table import BlockTables
 from vllm.v1.worker.gpu.cudagraph_utils import CudaGraphManager
@@ -31,6 +30,12 @@ from vllm.v1.worker.gpu.cudagraph_utils import \
 from vllm.v1.worker.gpu.input_batch import InputBuffers

 from vllm_ascend.worker.v2.utils import torch_cuda_wrapper
+from vllm_ascend.utils import vllm_version_is
+
+if vllm_version_is('0.14.1'):
+    from vllm.v1.attention.backends.utils import AttentionMetadataBuilder
+else:
+    from vllm.v1.attention.backend import AttentionMetadataBuilder


 class AclGraphManager(CudaGraphManager):
--- a/vllm_ascend/worker/v2/attn_utils.py
+++ b/vllm_ascend/worker/v2/attn_utils.py
@@ -23,13 +23,18 @@ from typing import Any, Tuple
 import numpy as np
 import torch
 from vllm.config import VllmConfig
-from vllm.v1.attention.backends.utils import AttentionMetadataBuilder
 from vllm.v1.kv_cache_interface import EncoderOnlyAttentionSpec, KVCacheConfig

 from vllm_ascend.attention.attention_mask import AttentionMaskBuilder
 from vllm_ascend.attention.attention_v1 import AscendAttentionState
 from vllm_ascend.attention.utils import (AscendCommonAttentionMetadata,
                                         AscendPrefillContextParallelMetadata)
+from vllm_ascend.utils import vllm_version_is
+
+if vllm_version_is('0.14.1'):
+    from vllm.v1.attention.backends.utils import AttentionMetadataBuilder
+else:
+    from vllm.v1.attention.backend import AttentionMetadataBuilder

 _ATTENTION_MASK_BUILDER = None

--- a/vllm_ascend/worker/v2/sample/penalties.py
+++ b/vllm_ascend/worker/v2/sample/penalties.py
@@ -20,7 +20,7 @@

 import torch
 from vllm.triton_utils import tl, triton
-from vllm.v1.worker.gpu.sample.metadata import SamplingMetadata
+from vllm.v1.sample.metadata import SamplingMetadata


@triton.jit
--- a/vllm_ascend/worker/v2/sample/sampler.py
+++ b/vllm_ascend/worker/v2/sample/sampler.py
@@ -17,7 +17,7 @@

 import torch
 from vllm.v1.sample.ops.topk_topp_sampler import apply_top_k_top_p
-from vllm.v1.worker.gpu.sample.metadata import SamplingMetadata
+from vllm.v1.sample.metadata import SamplingMetadata
 from vllm.v1.worker.gpu.sample.min_p import apply_min_p
 from vllm.v1.worker.gpu.sample.sampler import Sampler