[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/spec_decode/eagle_proposer.py
+++ b/vllm_ascend/spec_decode/eagle_proposer.py
@@ -41,7 +41,7 @@ from vllm_ascend.ops.rotary_embedding import update_cos_sin
 from vllm_ascend.ops.triton.spec_decode.utils import \
    prepare_inputs_padded_kernel
 from vllm_ascend.ops.triton.triton_utils import get_vectorcore_num
-from vllm_ascend.utils import enable_sp, shared_expert_dp_enabled
+from vllm_ascend.utils import enable_sp, shared_expert_dp_enabled, vllm_version_is

 # Currently we will fix block size to a small one since `num_reqs` can't be too large
 _PREPARE_INPUTS_BLOCK_SIZE = 4
@@ -455,7 +455,11 @@ class EagleProposer(VllmEagleProposer):
        self.input_ids[last_token_indices] = next_token_ids
        if self.use_cuda_graph and \
            num_tokens <= self.runner.cudagraph_batch_sizes[-1]:
-            num_input_tokens = self.vllm_config.pad_for_cudagraph(num_tokens)
+            if vllm_version_is('0.14.1'):
+                num_input_tokens = self.vllm_config.pad_for_cudagraph(num_tokens)
+            else:
+                num_input_tokens = self.runner.cudagraph_dispatcher._bs_to_padded_graph_size[
+                    num_tokens]
        else:
            num_input_tokens = num_tokens

--- a/vllm_ascend/spec_decode/mtp_proposer.py
+++ b/vllm_ascend/spec_decode/mtp_proposer.py
@@ -17,7 +17,7 @@ from vllm_ascend.attention.utils import AscendCommonAttentionMetadata
 from vllm_ascend.compilation.acl_graph import ACLGraphWrapper
 from vllm_ascend.ops.rotary_embedding import get_cos_and_sin_mla
 from vllm_ascend.spec_decode.eagle_proposer import EagleProposer
-from vllm_ascend.utils import ProfileExecuteDuration, lmhead_tp_enable
+from vllm_ascend.utils import ProfileExecuteDuration, lmhead_tp_enable, vllm_version_is


 class MtpProposer(EagleProposer):
@@ -97,7 +97,7 @@ class MtpProposer(EagleProposer):
            attn_metadata = None

        input_ids = self.input_ids[:num_tokens]
-        positions = self.positions[:num_tokens]
+        positions = self._get_positions(num_tokens)
        previous_hidden_states = self.hidden_states[:num_tokens]
        for i in range(self.num_speculative_tokens):
            if i > 0 and not in_graph_capturing and aclgraph_runtime_mode == CUDAGraphMode.FULL:
@@ -244,14 +244,18 @@ class MtpProposer(EagleProposer):
        # Note(qcs): We may need to refactor these check logics.
        if self.use_cuda_graph and num_scheduled_tokens <= self.runner.cudagraph_batch_sizes[
                -1]:
-            num_input_tokens = self.vllm_config.pad_for_cudagraph(
-                num_scheduled_tokens)
+            if vllm_version_is('0.14.1'):
+                num_input_tokens = self.vllm_config.pad_for_cudagraph(
+                    num_scheduled_tokens)
+            else:
+                num_input_tokens = self.runner.cudagraph_dispatcher._bs_to_padded_graph_size[
+                    num_scheduled_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._set_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,
@@ -311,7 +315,7 @@ class MtpProposer(EagleProposer):
                    model_kwargs = {}
                    model_kwargs["attn_metadata"] = attn_metadata
                    input_ids = self.input_ids[:num_input_tokens]
-                    positions = self.positions[:num_input_tokens]
+                    positions = self._get_positions(num_input_tokens)
                    hidden_states = self.hidden_states[:num_input_tokens]

                    hidden_states, positions = self.maybe_pad_and_reduce(
@@ -474,7 +478,7 @@ class MtpProposer(EagleProposer):

            # copy inputs to buffer for cudagraph
            self.input_ids[:batch_size] = input_ids
-            self.positions[:batch_size] = clamped_positions
+            self._set_positions(batch_size, clamped_positions)
            self.hidden_states[:hidden_states.shape[0]] = hidden_states
            if self.pcp_size * self.dcp_size > 1:
                # update local seq_len
@@ -495,7 +499,10 @@ class MtpProposer(EagleProposer):
            else:
                attn_metadata_i.slot_mapping[:batch_size] = slot_mapping
            if self.speculative_config.disable_padded_drafter_batch:
-                self.positions[batch_size:num_input_tokens] = 0
+                if self.uses_mrope:
+                   self.mrope_positions[:, batch_size:num_input_tokens] = 0
+                else:
+                    self.positions[batch_size:num_input_tokens] = 0
                self.input_ids[batch_size:num_input_tokens] = 0
                self.hidden_states[batch_size:num_input_tokens].fill_(0)

@@ -504,8 +511,8 @@ class MtpProposer(EagleProposer):
                prefill_metadata.seq_lens_list = prefill_metadata.seq_lens.tolist(
                )
                prefill_metadata.context_lens = attn_metadata_i.seq_lens
-                prefill_metadata.input_positions = self.positions[:
-                                                                  num_input_tokens]
+                prefill_metadata.input_positions = self._get_positions(
+                    num_input_tokens)
                prefill_metadata.max_seq_lens += 1
                prefill_metadata.max_seq_lens = min(
                    prefill_metadata.max_seq_lens,
@@ -520,8 +527,8 @@ class MtpProposer(EagleProposer):
                    decode_metadata.seq_lens_list = decode_seq_lens_list + [
                        0
                    ] * (graph_pad_size - len(decode_seq_lens_list))
-                decode_metadata.input_positions = self.positions[:
-                                                                 num_input_tokens]
+                decode_metadata.input_positions = self._get_positions(
+                    num_input_tokens)
                decode_metadata.max_seq_lens += 1
                decode_metadata.max_seq_lens = min(
                    decode_metadata.max_seq_lens,