Gahow Wang
445e491123
third_party/vllm/ now tracked in git for direct patch management.
Based on vLLM v0.18.1 release with one patch applied:
vllm/v1/core/sched/scheduler.py:
Replace fatal assert with graceful skip when KV transfer callback
arrives for an already-aborted request during PD disaggregated serving.
Future vLLM modifications should be made directly in third_party/vllm/
and committed normally. The patches/ directory is kept as documentation
of what changed from upstream.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
1.3 KiB
1.3 KiB
MTP (Multi-Token Prediction)
MTP is a speculative decoding method where the target model includes native multi-token prediction capability. Unlike draft-model-based methods, you do not need to provide a separate draft model.
MTP is useful when:
- Your model natively supports MTP.
- You want model-based speculative decoding with minimal extra configuration.
Offline Example
from vllm import LLM, SamplingParams
prompts = ["The future of AI is"]
sampling_params = SamplingParams(temperature=0.8, top_p=0.95)
llm = LLM(
model="XiaomiMiMo/MiMo-7B-Base",
tensor_parallel_size=1,
speculative_config={
"method": "mtp",
"num_speculative_tokens": 1,
},
)
outputs = llm.generate(prompts, sampling_params)
for output in outputs:
prompt = output.prompt
generated_text = output.outputs[0].text
print(f"Prompt: {prompt!r}, Generated text: {generated_text!r}")
Online Example
vllm serve XiaomiMiMo/MiMo-7B-Base \
--tensor-parallel-size 1 \
--speculative_config '{"method":"mtp","num_speculative_tokens":1}'
Notes
- MTP only works for model families that support MTP in vLLM.
num_speculative_tokenscontrols speculative depth. A small value like1is a good default to start with.- If your model does not support MTP, use another method such as EAGLE or draft model speculation.