Commit history

1. Commits on Apr 9, 2026

   • fix: use torch.topk instead of calling top_k parameter as function in sample_from_logits

     Show description for fde8f60
     kuishou68

     committed
     fde8f60View commit details

     Copy full SHA for fde8f60

     View code at this pointBrowse repository at this point
2. Commits on Dec 20, 2025

   • Auto-detect device for easier getting started

     alexliao

     committed
     369bc0aView commit details

     Copy full SHA for 369bc0a

     View code at this pointBrowse repository at this point
3. Commits on Oct 29, 2025

   • Refactor auto_regressive_inference to reduce memory allocations and cpu-gpy syncs.

     AnMakc

     committed
     b62f780View commit details

     Copy full SHA for b62f780

     View code at this pointBrowse repository at this point

   • Remove unnecessary CUDA cache clearing - no memory leak reproducible now.

     AnMakc

     committed
     6456913View commit details

     Copy full SHA for 6456913

     View code at this pointBrowse repository at this point

   • Do not collect tokenizer metrics during inference

     AnMakc

     committed
     21eb36aView commit details

     Copy full SHA for 21eb36a

     View code at this pointBrowse repository at this point

   • Use pytorch SDPA implementation

     AnMakc

     committed
     dccfa76View commit details

     Copy full SHA for dccfa76

     View code at this pointBrowse repository at this point
4. Commits on Oct 25, 2025

   • fix: define missing split_token in HierarchicalEmbedding

     SoYuCry

     committed
     a7e294cView commit details

     Copy full SHA for a7e294c

     View code at this pointBrowse repository at this point
5. Commits on Sep 16, 2025

   • Bug fix

     shiyu-coder

     authored
     8715716View commit details

     Copy full SHA for 8715716

     View code at this pointBrowse repository at this point

   • Bug fix

     shiyu-coder

     authored
     ac69e16View commit details

     Copy full SHA for ac69e16

     View code at this pointBrowse repository at this point
6. Commits on Sep 2, 2025

   • fix: add torch.cuda.empty_cache() during autoregressive inference

     Show description for e027051
     pengxiao-song

     authored
     e027051View commit details

     Copy full SHA for e027051

     View code at this pointBrowse repository at this point
7. Commits on Sep 1, 2025

   • update kronos model code

     

     quant

     committed
     38a643bView commit details

     Copy full SHA for 38a643b

     View code at this pointBrowse repository at this point
8. Commits on Jul 1, 2025

   • initial

     shiyu-coder

     committed
     9f946deView commit details

     Copy full SHA for 9f946de

     View code at this pointBrowse repository at this point
9. Loading