Skip to content
Closed
Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
59 commits
Select commit Hold shift + click to select a range
9a3b297
feat: Implement lazy data loading for Dataset (#246)
Ratish1 Dec 31, 2025
b0d3341
Revert "feat: Implement lazy data loading for Dataset" (#372)
zhaochenyang20 Dec 31, 2025
8ba715e
[MISC] add codeowners (#373)
Ying1123 Dec 31, 2025
bc61a7d
[Misc] update codeowners (#374)
Ying1123 Dec 31, 2025
8a988dc
[auto-sync] update code (#383)
miles-code-angel Jan 3, 2026
9dfd339
update code (#385)
miles-code-angel Jan 4, 2026
927d653
Cherry Pick commits to local fix CI unit tests (#393)
zhaochenyang20 Jan 5, 2026
47a5bdf
[example] Add SWE-agent example (#367)
yueming-yuan Jan 5, 2026
7434e9a
add background health check to miles native router (#260)
Ratish1 Jan 5, 2026
c77bdcf
update code (#401)
miles-code-angel Jan 5, 2026
43b9543
Update example dir (#345)
zijiexia Jan 7, 2026
236b640
update code (#411)
miles-code-angel Jan 8, 2026
60fc56e
move swe to experimental (#421)
yueming-yuan Jan 12, 2026
5d7a21c
feat: add int4 reinforcement learning training support (Part1) (#422)
zhaochenyang20 Jan 13, 2026
f68eef8
refactor [1/X]: unify training backends by general utils, tested Mega…
yueming-yuan Jan 13, 2026
20ab4f2
[squashed] Support VLM Multi-turn Training with Megatron, Support INT…
zhaochenyang20 Jan 13, 2026
636c995
[minor] delete unused util file (#428)
yueming-yuan Jan 13, 2026
e6571fa
Remove AI response in the doc (#429)
zijiexia Jan 13, 2026
1f619e1
Fix rollout-all-samples (#431)
fzyzcjy Jan 14, 2026
2cdc1f7
Fix retool example incorrectly handling max_tool_calls (#462)
fzyzcjy Jan 16, 2026
5dd0044
Integrate Terminal Bench into Miles (#447)
JessicaJiang-123 Jan 17, 2026
dfd822c
[CI] Fix and setup CI (#402)
yushengsu-thu Jan 19, 2026
38c152f
[CI] R3 bug fix & add CI test for R3 (#496)
zhaochenyang20 Jan 20, 2026
fc1076f
first attempt in supporting deepseek v3.2
yueming-yuan Dec 8, 2025
a7373e7
update
yueming-yuan Dec 10, 2025
b62966e
add several fix, supported thd + CP on megatron's dsa, added dockerfile
yueming-yuan Dec 14, 2025
1a25680
update dockerfile: TE version, fast-hadamard-transform
yueming-yuan Dec 14, 2025
f1674e9
update patches
yueming-yuan Dec 14, 2025
66d3b24
update script
yueming-yuan Dec 14, 2025
ccdff92
minor fix
yueming-yuan Dec 15, 2025
fd6bea6
fix
yueming-yuan Dec 23, 2025
02c1401
update
yueming-yuan Dec 16, 2025
bb09c27
init
yueming-yuan Dec 19, 2025
8af1384
supported bshd
yueming-yuan Dec 23, 2025
42c680f
lint
yueming-yuan Dec 23, 2025
0791a9c
rename, add argument assert, lint
yueming-yuan Dec 23, 2025
d8cb73a
tmp fix
yueming-yuan Dec 28, 2025
7009e11
update megatron patch
yueming-yuan Dec 28, 2025
4499325
update transformers patch
yueming-yuan Dec 28, 2025
6f1e130
disable amem
yueming-yuan Dec 28, 2025
cbd2e9f
add script
yueming-yuan Dec 28, 2025
9dc5258
update
yueming-yuan Dec 29, 2025
cab9686
fix
yueming-yuan Dec 29, 2025
8d51fe0
rm unused script
yueming-yuan Dec 29, 2025
f7beab4
fix
yueming-yuan Dec 29, 2025
dd68706
add docs
yueming-yuan Dec 29, 2025
f16e095
Fix torch native CP attention backend for DSA (#406)
xiuhu17 Jan 7, 2026
e28d439
tilelang kernel + matrix absorb in megatron (#461)
xiuhu17 Jan 16, 2026
2c3534b
update
xiuhu17 Jan 19, 2026
d81f29c
update
xiuhu17 Jan 20, 2026
474d542
Enable experimental rollout flag for CI tests (#492)
fzyzcjy Jan 22, 2026
e1e2305
update
xiuhu17 Jan 22, 2026
a16fb3f
Merge branch 'radixark:main' into dsv32_r3
xiuhu17 Jan 22, 2026
72bafb1
Fix PYTHONPATH for AMD container Megatron-LM location (#506)
lizamd Jan 22, 2026
f8e4cd8
update
xiuhu17 Jan 22, 2026
92da5dc
Merge branch 'radixark:main' into dsv32_r3
xiuhu17 Jan 22, 2026
b556aec
update
xiuhu17 Jan 23, 2026
adf07aa
update
xiuhu17 Jan 24, 2026
1def842
update
xiuhu17 Jan 24, 2026
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 3 additions & 0 deletions .github/CODEOWNERS
Original file line number Diff line number Diff line change
@@ -0,0 +1,3 @@
.github/CODEOWNERS @fzyzcjy @Ying1123
.github/workflows/ @yushengsu-thu
/miles/ @fzyzcjy @yueming-yuan
90 changes: 90 additions & 0 deletions .github/workflows/conda-ci.yml
Original file line number Diff line number Diff line change
@@ -0,0 +1,90 @@
name: conda CI

on:
pull_request:
branches: [main]

concurrency:
group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }}
cancel-in-progress: true

jobs:
build-conda:
if: contains(github.event.pull_request.title, '[release]')
runs-on: self-hosted
container:
image: lmsysorg/sglang:v0.5.0rc0-cu126
options: --gpus all --ipc=host --shm-size=16g --ulimit memlock=-1 --ulimit stack=67108864 --memory=0 --memory-swap=0 -v /mnt/nvme0n1/models:/root/models -v /mnt/nvme0n1/datasets:/root/datasets

defaults:
run:
working-directory: ${{ github.workspace }}

steps:
- name: Checkout repository
uses: actions/checkout@v4

- name: Construct Conda
run: |
echo "📦 Installing miles..."
cd $GITHUB_WORKSPACE
echo "Current directory: $(pwd)"

mkdir -p /root/
BASE_DIR=/root bash build_conda.sh
shell: bash

- name: Download model and dataset
run: |
echo "🔗 Downloading up model and dataset..."

# Create cache directories if they don't exist
mkdir -p /root/models /root/datasets

echo "Downloading Qwen3-30B-A3B..."
hf download Qwen/Qwen3-30B-A3B --local-dir /root/models/Qwen3-30B-A3B
hf download Qwen/Qwen3-30B-A3B-FP8 --local-dir /root/models/Qwen3-30B-A3B-FP8

hf download --repo-type dataset zhuzilin/dapo-math-17k --local-dir /root/datasets/dapo-math-17k

hf download --repo-type dataset zhuzilin/aime-2024 --local-dir /root/datasets/aime-2024
shell: bash

- name: Convert checkpoint
run: |
echo "🔄 Converting model checkpoint..."
cd $GITHUB_WORKSPACE
echo "Current directory: $(pwd)"

source ~/.bashrc
micromamba activate miles
export CUDA_HOME="$CONDA_PREFIX"

source scripts/models/qwen3-30B-A3B.sh
PYTHONPATH=/root/Megatron-LM torchrun --nproc-per-node 8 tools/convert_hf_to_torch_dist.py \
${MODEL_ARGS[@]} \
--hf-checkpoint /root/models/Qwen3-30B-A3B \
--save /root/Qwen3-30B-A3B_torch_dist
shell: bash

- name: Run tests
run: |
echo "🧪 Running tests..."
cd $GITHUB_WORKSPACE
echo "Current directory: $(pwd)"

source ~/.bashrc
micromamba activate miles
export CUDA_HOME="$CONDA_PREFIX"

MILES_TEST_USE_DEEPEP=0 MILES_TEST_USE_FP8_ROLLOUT=0 python tests/test_qwen3_30B_A3B.py
shell: bash

- name: Cleanup
if: always()
run: |
echo "🧹 Cleaning up..."
pkill -9 ray || true
ray stop --force || true
pkill -9 python || true
shell: bash
Loading