On-Policy Self-Distillation for Post-Training Few-Step Autoregressive Video Generators
-
Updated
Sep 28, 2026 - Python
#
posttraining
Here are 9 public repositories matching this topic...
Tools for OpenDataArena: Fair, Open, and Transparent Arena for Data
-
Updated
Mar 15, 2026 - Python
Singer is a complete training and evaluation pipeline for classical Chinese regulated verse (格律诗).
-
Updated
Jul 19, 2026 - Python
600ms-budget full-duplex voice agent — cascade ASR·RAG·LLM·TTS (458ms) + end-to-end Omni (250ms) | Voice · ASR · TTS · LLM · Infra · Posttraining
-
Updated
Aug 27, 2026 - Python
A Data-Centric RL and On-Policy Distillation Framework for LLM Post-Training
-
Updated
Sep 9, 2026 - Python
A general purpose rig for fine-tuning LLMs
-
Updated
Apr 4, 2026 - Python
TARDIS is a causal video-generation framework for temporal consistency that transports predictable content from previous frames via confidence-weighted motion transport, projects the residual into a transport-orbit tangent space (handled by a deterministic corrector) versus a normal quotient space (handled by diffusion)
-
Updated
Sep 2, 2026 - Python
Post-training code LLMs: SFT + DPO on preference pairs labelled by a sandbox and static analysis, with controlled reward ablations and reproducible HumanEval evaluation. v2 (in progress): a benchmark of PEFT, alignment and distillation methods for code.
python machine-learning deep-learning jetbrains static-analysis code-generation code-quality cyclomatic-complexity post-training peft dpo huggingface trl humaneval llm qlora qwen reward-modeling reward-hacking posttraining
-
Updated
Sep 27, 2026 - Python
Modern, improved reproduction of the prominent GPT 2 (124M)
-
Updated
Mar 30, 2026 - Python
Add this topic to your repo
To associate your repository with the posttraining topic, visit your repo's landing page and select "manage topics."