PaddlePaddle / PaddlePaddle/FastDeploy
[Deterministic Inference] Support Deterministic Inference
Open
@gongshaotian is already working on this.
Since Oct 29, 2025.
Deterministic
- Dominant language
- Python
- Stars
- 3.7k
- Forks
- 756
- Avg merge
- 19h 28m
- Merged PRs (30d)
- 4
Description
MileStone 1: Support Dense Model Deterministic Inference
- Batch Invariant ops
- Refer to implement the Paddle version of
batch_invariant_ops - Other ops
- Refer to implement the Paddle version of
- Build batch invariance testing scripts for request level and operator level
- Cascade Append Attention Backend、FA3 Support Batch Invariant
- Support Dense Model
MileStone 2: Support CUDAGraph、ChunkPrefill、PrefixCache、Moe Model
- Support CUDAGraph
- Support ChunkPrefill
- Support PrefixCache
- Support Moe Model
MileStone 3: Support SpeculativeDecoding、 Parallelism、Quantization
- Support SpecDecoding
- MTP
- Parallelism
- TP
- EP
- Quantization
- BlockWise FP8
MileStone 4: Support RL trainning
- RL
Related resources
- Blog: https://thinkingmachines.ai/blog/defeating-nondeterminism-in-llm-inference/
- batch_invariant_ops: https://github.com/thinking-machines-lab/batch_invariant_ops
- SGLang : https://github.com/sgl-project/sglang/issues/10278
- vLLM：https://github.com/orgs/vllm-project/projects/29
- slime: https://github.com/THUDM/slime/pull/361
Contributor guide
No contributing guide indexed for this repository
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Assessment
This issue has not been assessed yet.