SpecForge and TorchSpec
SpecForge and TorchSpec are competitors for training speculative decoding models, with SpecForge offering additional integration into the SGLang serving ecosystem while TorchSpec provides a PyTorch-native alternative.
About SpecForge
sgl-project/SpecForge
Train speculative decoding models effortlessly and port them smoothly to SGLang serving.
About TorchSpec
torchspec-project/TorchSpec
A PyTorch native library for training speculative decoding models
Decouples inference and training via a disaggregated pipeline that streams hidden states from vLLM or SGLang inference engines to distributed training workers through Mooncake's in-memory store, enabling independent scaling of each component. Integrates directly with PyTorch FSDP for distributed training, uses vLLM's Worker Extension API to avoid RPC serialization overhead, and supports vocabulary pruning with HuggingFace checkpoint conversion. Includes production examples for Qwen3, Kimi-K2.5, and MiniMax-M2.5 models with configurable training modes for resuming interrupted runs or continual training from existing weights.
Scores updated daily from GitHub, PyPI, and npm data. How scores work