8000 Commits · Infini-AI-Lab/astraflow · GitHub

Repository navigation

Commits

Commits on Jul 26, 2026
- update recipe
  haizhongzheng
  committed
  13efd72
Commits on Jul 25, 2026
- update recipe
  haizhongzheng
  committed
  aadf26f
Commits on Jul 24, 2026
- udpate training recipes
  Haizhong Zheng
  committed
  343e808
Commits on Jul 14, 2026
Commits on Jul 13, 2026
Commits on Jul 12, 2026
Commits on Jul 10, 2026
- recipes: set chunked_prefill_size=8192 for Qwen3.5-4B (H100 rollout fix)
  
  haizhongzheng
  and
  claude
  committed
  bb7fe50
Commits on Jun 25, 2026
- docs: note Qwen3.5-4B re-validation on pinned sglang 0.5.13.post1
  
  haizhongzheng
  and
  claude
  committed
  34b17d2
Commits on Jun 24, 2026
- deps: pin sglang to validated 0.5.13.post1 release (supersedes git build)
  
  haizhongzheng
  and
  claude
  committed
  1e44ee7
Commits on Jun 23, 2026
- recipes: default dense Qwen3 to kernels-hub FlashAttention-2
  
  haizhongzheng
  and
  claude
  committed
  3d5b0cc
Commits on Jun 22, 2026
- deps: match astraflow35 -- pin flash-linear-attention 0.5.0; networkx 3.6.1
  
  haizhongzheng
  and
  claude
  committed
  9c844d6
- style: align comments with repo conventions; refresh sglang install note
  
  haizhongzheng
  and
  claude
  committed
  705f954
- deps: pin sglang to the validated git build (qwen3_5 + transformers 5.8.1)
  
  haizhongzheng
  and
  claude
  committed
  1c111c2
- deps: pin transformers 5.8.1 (validated stack) + kernels 0.14
  
  haizhongzheng
  and
  claude
  committed
  6607b11
- docs: add Qwen3.5-4B math recipe README (validated stack + eval)
  
  haizhongzheng
  and
  claude
  committed
  56ea705
- hygiene: drop dead is_qwen3_5_model; document qwen3_5 model routing
  
  haizhongzheng
  and
  claude
  committed
  e11e188
- Make Qwen3 dense recipes run on transformers-5.8 / torch-2.11 stack
  
  haizhongzheng
  and
  claude
  committed
  f40a14d
- Add Qwen3.5-4B math RL recipes (full + delta) + Qwen3.5 enablement
  
  haizhongzheng
  and
  claude
  committed
  021a836
Commits on Jun 21, 2026
- Merge pull request #18 from Liquid4All/liquid-cross-platform-verify
  haizhongzheng
  authored
  e5467b1
Commits on Jun 19, 2026
- Drop H100 sbatch wrappers (cluster-specific, not upstream)
  
  ChangyiYang
  and
  claude
  committed
  cfd8b6a
- docs(amd): focus on using Docker, drop step-by-step build content
  
  ChangyiYang
  and
  claude
  committed
  23e9949
- docs(amd): mirror installation.md structure (Option A custom / Option B docker)
  
  ChangyiYang
  and
  claude
  committed
  b212be8
- Trim PR scope: inline prune, drop enroot/pyxis tooling
  
  ChangyiYang
  and
  claude
  committed
  dc95b77
- Drop verification report + dead torch_memory_saver shim from PR
  
  ChangyiYang
  and
  claude
  committed
  965194c
- docs: add Installation (AMD ROCm) page
  
  ChangyiYang
  and
  claude
  committed
  dc3aa43
Commits on Jun 17, 2026
- fix: enable chunked prefill to avoid SGLang OOM in 8B math recipes
  haizhongzheng
  committed
  679d992

0