[Speculative Decoding] Support suffix decoding

[Deleter-D]

Motivation

💡 If this PR is a Cherry Pick, the PR title needs to follow the format by adding the [Cherry-Pick] label at the very beginning and appending the original PR ID at the end. For example, [Cherry-Pick][CI] Add check trigger and logic(#5191)

💡 如若此PR是Cherry Pick，PR标题需遵循格式，在最开始加上[Cherry-Pick]标签，以及最后面加上原PR ID，例如[Cherry-Pick][CI] Add check trigger and logic(#5191)

Suffix Decoding is a model-free speculative decoding method that accelerates repetitive inference tasks (e.g., agent workflows, coding) using efficient CPU-based suffix trees for rapid draft token prediction, eliminating GPU overhead.

Modifications

• Support suffix decoding
• Refine CUDA Graph replay selection in speculative decoding

Usage or Command

python -m fastdeploy.entrypoints.openai.api_server \
    --model ${path_to_main_model} \
    --tensor-parallel-size 4 \
    --config ${path_to_FastDeploy}benchmarks/yaml/eb45t-32k-wint4-mtp-h100-tp4.yaml \
    --speculative-config '{"method": "mtp", "num_speculative_tokens": 4, "suffix_decoding_max_tree_depth": 64, "suffix_decoding_max_cached_requests": 10000, "suffix_decoding_max_spec_factor": 1.0, "suffix_decoding_min_token_prob": 0.1}'

Parameter Descriptions

# The maximum length of token sequences cached in suffix trees.
self.suffix_decoding_max_tree_depth: int = 64

# The limits of requests that can be stored in the cache.
self.suffix_decoding_max_cached_requests: int = -1

# The factor of matched length, calculated as num_draft_tokens = suffix_max_spec_factor * matched_length
self.suffix_decoding_max_spec_factor: float = 1.0

# The probability threshold for speculated tokens.
self.suffix_decoding_min_token_prob: float = 0.1

Accuracy Tests

N/A

Checklist

• Add at least a tag in the PR title.
  • Tag list: [[FDConfig],[APIServer],[Engine], [Scheduler], [PD Disaggregation], [Executor], [Graph Optimization], [Speculative Decoding], [RL], [Models], [Quantization], [Loader], [OP], [KVCache], [DataProcessor], [BugFix], [Docs], [CI], [Optimization], [Feature], [Benchmark], [Others], [XPU], [HPU], [GCU], [DCU], [Iluvatar], [Metax]]
  • You can add new tags based on the PR content, but the semantics must be clear.
• Format your code, run pre-commit before commit.
• Add unit tests. Please write the reason in this PR if no unit tests.
• Provide accuracy results.
• If the current PR is submitting to the release branch, make sure the PR has been submitted to the develop branch, then cherry-pick it to the release branch with the [Cherry-Pick] PR tag.

[paddle-bot]

Thanks for your contribution!