[main][Bugfix] Fixed an problem related to embeddings sharing by drslark · Pull Request #5967 · vllm-project/vllm-ascend

drslark · 2026-01-17T09:25:33Z

What this PR does / why we need it?

Cancel the embeddings sharing when the embeddings of main model and the embeddings of eagle model are different.

Does this PR introduce any user-facing change?

N/A

How was this patch tested?

Cause i don't have Meta-Llama-3.1-8B-Instruct locally, i commented it and run:

pytest -s tests/e2e/singlecard/spec_decode/test_v1_spec_decode.py::test_llama_qwen_eagle_acceptance

The output is fine:

.

======================================================================================================================== warnings summary =========================================================================================================================
<frozen importlib._bootstrap>:241
  <frozen importlib._bootstrap>:241: DeprecationWarning: builtin type SwigPyPacked has no __module__ attribute

<frozen importlib._bootstrap>:241
  <frozen importlib._bootstrap>:241: DeprecationWarning: builtin type SwigPyObject has no __module__ attribute

-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
====================================================================================================== 3 passed, 1 skipped, 2 warnings in 196.19s (0:03:16) =======================================================================================================

vLLM version: v0.13.0
vLLM main: vllm-project/vllm@2c24bc6

github-actions · 2026-01-17T09:25:47Z

👋 Hi! Thank you for contributing to the vLLM Ascend project. The following points will speed up your PR merge:‌‌

A PR should do only one thing, smaller PRs enable faster reviews.
Every PR should include unit tests and end-to-end tests ‌to ensure it works and is not broken by other future PRs.
Write the commit message by fulfilling the PR description to help reviewer and future developers understand.

If CI fails, you can run linting and testing checks locally according Contributing and Testing.

gemini-code-assist

Code Review

This pull request addresses an issue with embedding sharing in speculative decoding. The change correctly prevents unconditional sharing for non-MTP methods. However, the new condition for sharing, self.vllm_config.model_config.is_deepseek_mla and torch.equal(...), appears overly restrictive. It may prevent a valid memory optimization by not sharing embeddings on non-DeepSeek models even when their weights are identical. I have suggested removing the is_deepseek_mla check to align with the principle of sharing embeddings whenever they are equal, which seems to be the core intent of this bugfix.

gemini-code-assist · 2026-01-17T09:26:58Z

+            if self.vllm_config.model_config.is_deepseek_mla and \
+                torch.equal(self.model.model.embed_tokens.weight,
+                            model.model.embed_tokens.weight):


The condition to share embeddings seems too restrictive. By including self.vllm_config.model_config.is_deepseek_mla, you are preventing embedding sharing for non-DeepSeek models even when their embedding weights are identical. This could be a missed memory optimization.

Based on the PR description, the goal is to share embeddings if and only if they are identical. To achieve that, you should probably remove the is_deepseek_mla check.

if torch.equal(self.model.model.embed_tokens.weight, model.model.embed_tokens.weight):

github-actions · 2026-01-19T01:02:25Z

This pull request has conflicts, please resolve those before we can evaluate the pull request.

Signed-off-by: drslark <slarksblood@qq.com>

…to FIA_rebase * 'main' of https://github.com/vllm-project/vllm-ascend: (24 commits) add dispath_ffn_combine_bf16 (vllm-project#5866) [BugFix] Fix input parameter bug of dispatch_gmm_combine_decode[RFC: issue 5476] (vllm-project#5932) [1/N][Feat] Xlite Qwen3 MoE Support (vllm-project#5951) [Bugfix] Fix setting of `speculative_config.enforce_eager` for dsv32 (vllm-project#5945) [bugfix][mm] change get_num_encoder_tokens to get_num_encoder_embeds in recompute_schedule.py (vllm-project#5132) [Bugfix] fix pcp qwen full graph FIA bug (vllm-project#6037) [Bugfix]Fixed precision issues caused by pooled request pooling (vllm-project#6049) 【main】【bugfix】Resolved memory deallocation failure in the pooling layer under re-computation workloads. (vllm-project#6045) [main][Bugfix] Fixed an problem related to embeddings sharing (vllm-project#5967) [Feature]refactor the npugraph_ex config, support online-infer with static kernel (vllm-project#5775) [CI][Lint] Show lint diff on failure (vllm-project#5956) [CI] Add wait logic for each individual case (vllm-project#6036) [CI] Add DeepSeek-V3.2-W8A8 nightly ci test (vllm-project#4633) model runner v2 support triton of penalty (vllm-project#5854) [Docs][Model] Support Qwen3-VL-Embedding & Qwen3-VL-Reranker (vllm-project#6034) [Tests] move qwen3 performance test from nightly to e2e (vllm-project#5980) [Bugfix] fix bug of pcp+mtp+async scheduler (vllm-project#5994) [Main2Main] Upgrade vllm commit to releases/v0.14.0 (vllm-project#5988) [Ops] Add layernorm for qwen3Next (vllm-project#5765) [Doc] Add layer_sharding additional config for DeepSeek-V3.2-W8A8 (vllm-project#5921) ...

…3B-EAGLE3 (#6138) ### What this PR does / why we need it? Due to the long-term lack of synchronization with the upstream code, a problem that led to a decrease in the acceptance rate of the Qwen3-30B-A3B-EAGLE3 draft model was introduced when fixing the bug（#5967）. Now, synchronize with the upstream and fix this bug ### Does this PR introduce _any_ user-facing change? No ### How was this patch tested? ```python from vllm import LLM, SamplingParams def main(): prompts = [ "The future of AI is", ] # Create a sampling params object. sampling_params = SamplingParams(temperature=0.8, top_p=0.95) # Create an LLM. llm = LLM( model="Qwen/Qwen3-30B-A3B", tensor_parallel_size=4, gpu_memory_utilization=0.9, enforce_eager=True, speculative_config={ "method": "eagle3", "model": "AngelSlim/Qwen3-a3B_eagle3" "num_speculative_tokens": 3, }, ) # Generate texts from the prompts. outputs = llm.generate(prompts, sampling_params) print(f"Outputs: {outputs}") for output in outputs: prompt = output.prompt generated_text = output.outputs[0].text print(f"Prompt: {prompt!r}, Generated text: {generated_text!r}") ``` - vLLM version: v0.13.0 - vLLM main: vllm-project/vllm@d682094 Signed-off-by: zhaomingyu <zhaomingyu13@h-partners.com> Co-authored-by: drslark <slarkblood@qq.com>

…3B-EAGLE3 (#6139) ### What this PR does / why we need it? Due to the long-term lack of synchronization with the upstream code, a problem that led to a decrease in the acceptance rate of the Qwen3-30B-A3B-EAGLE3 draft model was introduced when fixing the bug（#5967）. Now, synchronize with the upstream and fix this bug ### Does this PR introduce _any_ user-facing change? No ### How was this patch tested? ```python from vllm import LLM, SamplingParams def main(): prompts = [ "The future of AI is", ] # Create a sampling params object. sampling_params = SamplingParams(temperature=0.8, top_p=0.95) # Create an LLM. llm = LLM( model="Qwen/Qwen3-30B-A3B", tensor_parallel_size=4, gpu_memory_utilization=0.9, enforce_eager=True, speculative_config={ "method": "eagle3", "model": "AngelSlim/Qwen3-a3B_eagle3" "num_speculative_tokens": 3, }, ) # Generate texts from the prompts. outputs = llm.generate(prompts, sampling_params) print(f"Outputs: {outputs}") for output in outputs: prompt = output.prompt generated_text = output.outputs[0].text print(f"Prompt: {prompt!r}, Generated text: {generated_text!r}") ``` Signed-off-by: zhaomingyu <zhaomingyu13@h-partners.com> Signed-off-by: zhaomingyu13 <zhaomingyu13@h-partners.com> Co-authored-by: drslark <slarkblood@qq.com>

…roject#5967) ### What this PR does / why we need it? Cancel the embeddings sharing when the embeddings of main model and the embeddings of eagle model are different. ### Does this PR introduce _any_ user-facing change? N/A ### How was this patch tested? Cause i don't have `Meta-Llama-3.1-8B-Instruc`t locally, i commented it and run: ```shell pytest -s tests/e2e/singlecard/spec_decode/test_v1_spec_decode.py::test_llama_qwen_eagle_acceptance ``` The output is fine: ```text . ======================================================================================================================== warnings summary ========================================================================================================================= <frozen importlib._bootstrap>:241 <frozen importlib._bootstrap>:241: DeprecationWarning: builtin type SwigPyPacked has no __module__ attribute <frozen importlib._bootstrap>:241 <frozen importlib._bootstrap>:241: DeprecationWarning: builtin type SwigPyObject has no __module__ attribute -- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html ====================================================================================================== 3 passed, 1 skipped, 2 warnings in 196.19s (0:03:16) ======================================================================================================= ``` - vLLM version: v0.13.0 - vLLM main: vllm-project/vllm@2c24bc6 Signed-off-by: drslark <slarksblood@qq.com>

…3B-EAGLE3 (vllm-project#6138) ### What this PR does / why we need it? Due to the long-term lack of synchronization with the upstream code, a problem that led to a decrease in the acceptance rate of the Qwen3-30B-A3B-EAGLE3 draft model was introduced when fixing the bug（vllm-project#5967）. Now, synchronize with the upstream and fix this bug ### Does this PR introduce _any_ user-facing change? No ### How was this patch tested? ```python from vllm import LLM, SamplingParams def main(): prompts = [ "The future of AI is", ] # Create a sampling params object. sampling_params = SamplingParams(temperature=0.8, top_p=0.95) # Create an LLM. llm = LLM( model="Qwen/Qwen3-30B-A3B", tensor_parallel_size=4, gpu_memory_utilization=0.9, enforce_eager=True, speculative_config={ "method": "eagle3", "model": "AngelSlim/Qwen3-a3B_eagle3" "num_speculative_tokens": 3, }, ) # Generate texts from the prompts. outputs = llm.generate(prompts, sampling_params) print(f"Outputs: {outputs}") for output in outputs: prompt = output.prompt generated_text = output.outputs[0].text print(f"Prompt: {prompt!r}, Generated text: {generated_text!r}") ``` - vLLM version: v0.13.0 - vLLM main: vllm-project/vllm@d682094 Signed-off-by: zhaomingyu <zhaomingyu13@h-partners.com> Co-authored-by: drslark <slarkblood@qq.com>

…3B-EAGLE3 (vllm-project#6139) ### What this PR does / why we need it? Due to the long-term lack of synchronization with the upstream code, a problem that led to a decrease in the acceptance rate of the Qwen3-30B-A3B-EAGLE3 draft model was introduced when fixing the bug（vllm-project#5967）. Now, synchronize with the upstream and fix this bug ### Does this PR introduce _any_ user-facing change? No ### How was this patch tested? ```python from vllm import LLM, SamplingParams def main(): prompts = [ "The future of AI is", ] # Create a sampling params object. sampling_params = SamplingParams(temperature=0.8, top_p=0.95) # Create an LLM. llm = LLM( model="Qwen/Qwen3-30B-A3B", tensor_parallel_size=4, gpu_memory_utilization=0.9, enforce_eager=True, speculative_config={ "method": "eagle3", "model": "AngelSlim/Qwen3-a3B_eagle3" "num_speculative_tokens": 3, }, ) # Generate texts from the prompts. outputs = llm.generate(prompts, sampling_params) print(f"Outputs: {outputs}") for output in outputs: prompt = output.prompt generated_text = output.outputs[0].text print(f"Prompt: {prompt!r}, Generated text: {generated_text!r}") ``` Signed-off-by: zhaomingyu <zhaomingyu13@h-partners.com> Signed-off-by: zhaomingyu13 <zhaomingyu13@h-partners.com> Co-authored-by: drslark <slarkblood@qq.com>

…roject#5967) ### What this PR does / why we need it? Cancel the embeddings sharing when the embeddings of main model and the embeddings of eagle model are different. ### Does this PR introduce _any_ user-facing change? N/A ### How was this patch tested? Cause i don't have `Meta-Llama-3.1-8B-Instruc`t locally, i commented it and run: ```shell pytest -s tests/e2e/singlecard/spec_decode/test_v1_spec_decode.py::test_llama_qwen_eagle_acceptance ``` The output is fine: ```text . ======================================================================================================================== warnings summary ========================================================================================================================= <frozen importlib._bootstrap>:241 <frozen importlib._bootstrap>:241: DeprecationWarning: builtin type SwigPyPacked has no __module__ attribute <frozen importlib._bootstrap>:241 <frozen importlib._bootstrap>:241: DeprecationWarning: builtin type SwigPyObject has no __module__ attribute -- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html ====================================================================================================== 3 passed, 1 skipped, 2 warnings in 196.19s (0:03:16) ======================================================================================================= ``` - vLLM version: v0.13.0 - vLLM main: vllm-project/vllm@2c24bc6 Signed-off-by: drslark <slarksblood@qq.com>

…3B-EAGLE3 (vllm-project#6138) ### What this PR does / why we need it? Due to the long-term lack of synchronization with the upstream code, a problem that led to a decrease in the acceptance rate of the Qwen3-30B-A3B-EAGLE3 draft model was introduced when fixing the bug（vllm-project#5967）. Now, synchronize with the upstream and fix this bug ### Does this PR introduce _any_ user-facing change? No ### How was this patch tested? ```python from vllm import LLM, SamplingParams def main(): prompts = [ "The future of AI is", ] # Create a sampling params object. sampling_params = SamplingParams(temperature=0.8, top_p=0.95) # Create an LLM. llm = LLM( model="Qwen/Qwen3-30B-A3B", tensor_parallel_size=4, gpu_memory_utilization=0.9, enforce_eager=True, speculative_config={ "method": "eagle3", "model": "AngelSlim/Qwen3-a3B_eagle3" "num_speculative_tokens": 3, }, ) # Generate texts from the prompts. outputs = llm.generate(prompts, sampling_params) print(f"Outputs: {outputs}") for output in outputs: prompt = output.prompt generated_text = output.outputs[0].text print(f"Prompt: {prompt!r}, Generated text: {generated_text!r}") ``` - vLLM version: v0.13.0 - vLLM main: vllm-project/vllm@d682094 Signed-off-by: zhaomingyu <zhaomingyu13@h-partners.com> Co-authored-by: drslark <slarkblood@qq.com>

…roject#5967) ### What this PR does / why we need it? Cancel the embeddings sharing when the embeddings of main model and the embeddings of eagle model are different. ### Does this PR introduce _any_ user-facing change? N/A ### How was this patch tested? Cause i don't have `Meta-Llama-3.1-8B-Instruc`t locally, i commented it and run: ```shell pytest -s tests/e2e/singlecard/spec_decode/test_v1_spec_decode.py::test_llama_qwen_eagle_acceptance ``` The output is fine: ```text . ======================================================================================================================== warnings summary ========================================================================================================================= <frozen importlib._bootstrap>:241 <frozen importlib._bootstrap>:241: DeprecationWarning: builtin type SwigPyPacked has no __module__ attribute <frozen importlib._bootstrap>:241 <frozen importlib._bootstrap>:241: DeprecationWarning: builtin type SwigPyObject has no __module__ attribute -- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html ====================================================================================================== 3 passed, 1 skipped, 2 warnings in 196.19s (0:03:16) ======================================================================================================= ``` - vLLM version: v0.13.0 - vLLM main: vllm-project/vllm@2c24bc6 Signed-off-by: drslark <slarksblood@qq.com> Signed-off-by: zrj026 <zhangrunjiang026@gmail.com>

…3B-EAGLE3 (vllm-project#6138) ### What this PR does / why we need it? Due to the long-term lack of synchronization with the upstream code, a problem that led to a decrease in the acceptance rate of the Qwen3-30B-A3B-EAGLE3 draft model was introduced when fixing the bug（vllm-project#5967）. Now, synchronize with the upstream and fix this bug ### Does this PR introduce _any_ user-facing change? No ### How was this patch tested? ```python from vllm import LLM, SamplingParams def main(): prompts = [ "The future of AI is", ] # Create a sampling params object. sampling_params = SamplingParams(temperature=0.8, top_p=0.95) # Create an LLM. llm = LLM( model="Qwen/Qwen3-30B-A3B", tensor_parallel_size=4, gpu_memory_utilization=0.9, enforce_eager=True, speculative_config={ "method": "eagle3", "model": "AngelSlim/Qwen3-a3B_eagle3" "num_speculative_tokens": 3, }, ) # Generate texts from the prompts. outputs = llm.generate(prompts, sampling_params) print(f"Outputs: {outputs}") for output in outputs: prompt = output.prompt generated_text = output.outputs[0].text print(f"Prompt: {prompt!r}, Generated text: {generated_text!r}") ``` - vLLM version: v0.13.0 - vLLM main: vllm-project/vllm@d682094 Signed-off-by: zhaomingyu <zhaomingyu13@h-partners.com> Co-authored-by: drslark <slarkblood@qq.com> Signed-off-by: zrj026 <zhangrunjiang026@gmail.com>

…roject#5967) ### What this PR does / why we need it? Cancel the embeddings sharing when the embeddings of main model and the embeddings of eagle model are different. ### Does this PR introduce _any_ user-facing change? N/A ### How was this patch tested? Cause i don't have `Meta-Llama-3.1-8B-Instruc`t locally, i commented it and run: ```shell pytest -s tests/e2e/singlecard/spec_decode/test_v1_spec_decode.py::test_llama_qwen_eagle_acceptance ``` The output is fine: ```text . ======================================================================================================================== warnings summary ========================================================================================================================= <frozen importlib._bootstrap>:241 <frozen importlib._bootstrap>:241: DeprecationWarning: builtin type SwigPyPacked has no __module__ attribute <frozen importlib._bootstrap>:241 <frozen importlib._bootstrap>:241: DeprecationWarning: builtin type SwigPyObject has no __module__ attribute -- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html ====================================================================================================== 3 passed, 1 skipped, 2 warnings in 196.19s (0:03:16) ======================================================================================================= ``` - vLLM version: v0.13.0 - vLLM main: vllm-project/vllm@2c24bc6 Signed-off-by: drslark <slarksblood@qq.com>

…3B-EAGLE3 (vllm-project#6138) ### What this PR does / why we need it? Due to the long-term lack of synchronization with the upstream code, a problem that led to a decrease in the acceptance rate of the Qwen3-30B-A3B-EAGLE3 draft model was introduced when fixing the bug（vllm-project#5967）. Now, synchronize with the upstream and fix this bug ### Does this PR introduce _any_ user-facing change? No ### How was this patch tested? ```python from vllm import LLM, SamplingParams def main(): prompts = [ "The future of AI is", ] # Create a sampling params object. sampling_params = SamplingParams(temperature=0.8, top_p=0.95) # Create an LLM. llm = LLM( model="Qwen/Qwen3-30B-A3B", tensor_parallel_size=4, gpu_memory_utilization=0.9, enforce_eager=True, speculative_config={ "method": "eagle3", "model": "AngelSlim/Qwen3-a3B_eagle3" "num_speculative_tokens": 3, }, ) # Generate texts from the prompts. outputs = llm.generate(prompts, sampling_params) print(f"Outputs: {outputs}") for output in outputs: prompt = output.prompt generated_text = output.outputs[0].text print(f"Prompt: {prompt!r}, Generated text: {generated_text!r}") ``` - vLLM version: v0.13.0 - vLLM main: vllm-project/vllm@d682094 Signed-off-by: zhaomingyu <zhaomingyu13@h-partners.com> Co-authored-by: drslark <slarkblood@qq.com>

…roject#5967) ### What this PR does / why we need it? Cancel the embeddings sharing when the embeddings of main model and the embeddings of eagle model are different. ### Does this PR introduce _any_ user-facing change? N/A ### How was this patch tested? Cause i don't have `Meta-Llama-3.1-8B-Instruc`t locally, i commented it and run: ```shell pytest -s tests/e2e/singlecard/spec_decode/test_v1_spec_decode.py::test_llama_qwen_eagle_acceptance ``` The output is fine: ```text . ======================================================================================================================== warnings summary ========================================================================================================================= <frozen importlib._bootstrap>:241 <frozen importlib._bootstrap>:241: DeprecationWarning: builtin type SwigPyPacked has no __module__ attribute <frozen importlib._bootstrap>:241 <frozen importlib._bootstrap>:241: DeprecationWarning: builtin type SwigPyObject has no __module__ attribute -- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html ====================================================================================================== 3 passed, 1 skipped, 2 warnings in 196.19s (0:03:16) ======================================================================================================= ``` - vLLM version: v0.13.0 - vLLM main: vllm-project/vllm@2c24bc6 Signed-off-by: drslark <slarksblood@qq.com> Signed-off-by: zrj026 <zhangrunjiang026@gmail.com>

…3B-EAGLE3 (vllm-project#6138) ### What this PR does / why we need it? Due to the long-term lack of synchronization with the upstream code, a problem that led to a decrease in the acceptance rate of the Qwen3-30B-A3B-EAGLE3 draft model was introduced when fixing the bug（vllm-project#5967）. Now, synchronize with the upstream and fix this bug ### Does this PR introduce _any_ user-facing change? No ### How was this patch tested? ```python from vllm import LLM, SamplingParams def main(): prompts = [ "The future of AI is", ] # Create a sampling params object. sampling_params = SamplingParams(temperature=0.8, top_p=0.95) # Create an LLM. llm = LLM( model="Qwen/Qwen3-30B-A3B", tensor_parallel_size=4, gpu_memory_utilization=0.9, enforce_eager=True, speculative_config={ "method": "eagle3", "model": "AngelSlim/Qwen3-a3B_eagle3" "num_speculative_tokens": 3, }, ) # Generate texts from the prompts. outputs = llm.generate(prompts, sampling_params) print(f"Outputs: {outputs}") for output in outputs: prompt = output.prompt generated_text = output.outputs[0].text print(f"Prompt: {prompt!r}, Generated text: {generated_text!r}") ``` - vLLM version: v0.13.0 - vLLM main: vllm-project/vllm@d682094 Signed-off-by: zhaomingyu <zhaomingyu13@h-partners.com> Co-authored-by: drslark <slarkblood@qq.com> Signed-off-by: zrj026 <zhangrunjiang026@gmail.com>

…roject#5967) ### What this PR does / why we need it? Cancel the embeddings sharing when the embeddings of main model and the embeddings of eagle model are different. ### Does this PR introduce _any_ user-facing change? N/A ### How was this patch tested? Cause i don't have `Meta-Llama-3.1-8B-Instruc`t locally, i commented it and run: ```shell pytest -s tests/e2e/singlecard/spec_decode/test_v1_spec_decode.py::test_llama_qwen_eagle_acceptance ``` The output is fine: ```text . ======================================================================================================================== warnings summary ========================================================================================================================= <frozen importlib._bootstrap>:241 <frozen importlib._bootstrap>:241: DeprecationWarning: builtin type SwigPyPacked has no __module__ attribute <frozen importlib._bootstrap>:241 <frozen importlib._bootstrap>:241: DeprecationWarning: builtin type SwigPyObject has no __module__ attribute -- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html ====================================================================================================== 3 passed, 1 skipped, 2 warnings in 196.19s (0:03:16) ======================================================================================================= ``` - vLLM version: v0.13.0 - vLLM main: vllm-project/vllm@2c24bc6 Signed-off-by: drslark <slarksblood@qq.com>

…3B-EAGLE3 (vllm-project#6138) ### What this PR does / why we need it? Due to the long-term lack of synchronization with the upstream code, a problem that led to a decrease in the acceptance rate of the Qwen3-30B-A3B-EAGLE3 draft model was introduced when fixing the bug（vllm-project#5967）. Now, synchronize with the upstream and fix this bug ### Does this PR introduce _any_ user-facing change? No ### How was this patch tested? ```python from vllm import LLM, SamplingParams def main(): prompts = [ "The future of AI is", ] # Create a sampling params object. sampling_params = SamplingParams(temperature=0.8, top_p=0.95) # Create an LLM. llm = LLM( model="Qwen/Qwen3-30B-A3B", tensor_parallel_size=4, gpu_memory_utilization=0.9, enforce_eager=True, speculative_config={ "method": "eagle3", "model": "AngelSlim/Qwen3-a3B_eagle3" "num_speculative_tokens": 3, }, ) # Generate texts from the prompts. outputs = llm.generate(prompts, sampling_params) print(f"Outputs: {outputs}") for output in outputs: prompt = output.prompt generated_text = output.outputs[0].text print(f"Prompt: {prompt!r}, Generated text: {generated_text!r}") ``` - vLLM version: v0.13.0 - vLLM main: vllm-project/vllm@d682094 Signed-off-by: zhaomingyu <zhaomingyu13@h-partners.com> Co-authored-by: drslark <slarkblood@qq.com>

drslark requested a review from wangxiyuan as a code owner January 17, 2026 09:25

gemini-code-assist bot reviewed Jan 17, 2026

View reviewed changes

drslark force-pushed the main branch 2 times, most recently from 584e8db to f3518c0 Compare January 17, 2026 12:44

drslark mentioned this pull request Jan 17, 2026

[Bug]: [main] when appling eagle3 to QuaRot model, the acceptence rates drop significantly #5974

Closed

github-actions bot added the merge-conflicts label Jan 19, 2026

drslark force-pushed the main branch from f3518c0 to bdafb77 Compare January 19, 2026 01:10

github-actions bot removed the merge-conflicts label Jan 19, 2026

drslark force-pushed the main branch 2 times, most recently from fa7a46a to 1e59244 Compare January 19, 2026 07:43

wangxiyuan approved these changes Jan 19, 2026

View reviewed changes

wangxiyuan added ready read for review ready-for-test start test by label for PR labels Jan 19, 2026

drslark force-pushed the main branch 5 times, most recently from f0a1094 to af0ebd5 Compare January 20, 2026 06:18

[main][Bugfix] Fixed an problem related to embeddings sharing

5dafe71

Signed-off-by: drslark <slarksblood@qq.com>

drslark force-pushed the main branch from af0ebd5 to 5dafe71 Compare January 20, 2026 06:46

yiz-liu approved these changes Jan 20, 2026

View reviewed changes

yiz-liu merged commit b247509 into vllm-project:main Jan 20, 2026
20 checks passed

This was referenced Jan 22, 2026

[Bugfix] Fix the issue of the acceptance rate decline for Qwen3-30B-A3B-EAGLE3 #6138

Merged

[Bugfix] Fix the issue of the acceptance rate decline for Qwen3-30B-A3B-EAGLE3 #6139

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[main][Bugfix] Fixed an problem related to embeddings sharing#5967

[main][Bugfix] Fixed an problem related to embeddings sharing#5967
yiz-liu merged 1 commit intovllm-project:mainfrom
drslark:main

drslark commented Jan 17, 2026 •

edited

Loading

Uh oh!

github-actions bot commented Jan 17, 2026

Uh oh!

gemini-code-assist bot left a comment

Uh oh!

gemini-code-assist bot Jan 17, 2026

Uh oh!

github-actions bot commented Jan 19, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Conversation

drslark commented Jan 17, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

What this PR does / why we need it?

Does this PR introduce any user-facing change?

How was this patch tested?

Uh oh!

github-actions bot commented Jan 17, 2026

Uh oh!

gemini-code-assist bot left a comment

Choose a reason for hiding this comment

Code Review

Uh oh!

gemini-code-assist bot Jan 17, 2026

Choose a reason for hiding this comment

Uh oh!

github-actions bot commented Jan 19, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

drslark commented Jan 17, 2026 •

edited

Loading