Skip to content

Zhiyaoj/solve conflict#7

Closed
JessicaJiang-123 wants to merge 144 commits intofeat/tb-eval-integrationfrom
zhiyaoj/solve-conflict
Closed

Zhiyaoj/solve conflict#7
JessicaJiang-123 wants to merge 144 commits intofeat/tb-eval-integrationfrom
zhiyaoj/solve-conflict

Conversation

@JessicaJiang-123
Copy link
Copy Markdown
Collaborator

@JessicaJiang-123 JessicaJiang-123 commented Jan 14, 2026

No description provided.

zhuzilin and others added 30 commits December 21, 2025 09:12
@zhuzilin
Remove config_mapping to use megatron-bridge (THUDM#1166)
9ac37f9
@qqwqqw689
Avoids repeated work. (THUDM#1163)
f3a96f8
@zhuzilin
Make tools/convert_torch_dist_to_hf.py not rely on megatron (THUDM#1167)
87cbed6
@zhuzilin
support converting dpsk mtp layer (THUDM#1169)
ece2624
@zijiexia @GuanxingLu
[FSDP] Add Masked importance sampling (THUDM#1122)
b74858a 
Co-authored-by: Guanxing Lu <747398423@qq.com>
@ChangyiYang
[TIS/MIS] fix and add better metric (THUDM#1174)
57b9c3d 
Co-authored-by: slime-local <slime-local@localhost>
@lr-tsinghua11
Fix optimizer schedule resume (THUDM#1152)
68c21dd
@zhuzilin
[docker] upgrade to megatron dev branch (THUDM#1153)
652674e
@lancerts
Minor fix (THUDM#1165)
8bc4b4d
@zhuzilin @Zhuohao-Li
Fix forward of Qwen3VLTextRotaryEmbedding in Megatron-Bridge (THUDM#1179
1766d0c 
)

Co-authored-by: Zhuohao-Li <garrick0508@gmail.com>
@zhuzilin
Reuse the text llm config for qwen3 vl models (THUDM#1180)
b749038
@zhuzilin
Don't save AutoBridge in args (THUDM#1181)
f94cd5f
@yitianlian
[Fix] Fix port error in PD disaggregation setting (THUDM#1175)
a52ddbd
@jiahe7ay
Fix prompt type bug in generate_with_search within examples/search-r1 (…
ab2d4d8 
…THUDM#1182)
@nanjiangwill @coding-famer
feat: support Qwen3 VL MoE (THUDM#1171)
9d92ce7 
Co-authored-by: Chenhe Gu <chenhegu0109@gmail.com>
@yitianlian
[Fix] Minor fix (THUDM#1183)
6dcd36e
@zhuzilin
Set parallel config for megatron bridge (THUDM#1184)
7dd5f63
@zhuzilin
Fix tools/convert_hf_to_torch_dist.py (THUDM#1186)
5a0b9d6
@zhuzilin
Don't calculate entropy grad when coef is 0 (THUDM#1185)
0d3ada3
@zhuzilin
Disable routing replay for critic (THUDM#1187)
03b11c9
@zhuzilin
Revert "Don't calculate entropy grad when coef is 0" (THUDM#1189)
f57d56f
@zhuzilin
Fix qwen3next for megatron dev branch (THUDM#1190)
4cfdacf
@nanjiangwill
fix: fix logging for rollout (THUDM#1188)
f638b1c
@zhuzilin
sync internal features (THUDM#1192)
b543a50
@zhuzilin
Fix check_weights api (THUDM#1194)
65f458c
@zhuzilin
Add --custom-rollout-log-function-path and --custom-eval-rollout-log-…
47a527d 
…function-path (THUDM#1196)
@yitianlian
[Feature] Add more logging for health monitor (THUDM#1195)
df5cf62
@maoquan-ms
fix: SFT tools support (THUDM#1198)
cc913c3
@yitianlian
[Featuren] Change default value of rollout health check (THUDM#1197)
9bf1e36
@Zhuohao-Li
Megatron VLM Support w/ SFT (2/N) (THUDM#1150)
80e0528
Hecate0821 and others added 29 commits January 12, 2026 10:40
@Hecate0821
[Doc] Add docs for R2/R3 (THUDM#1382)
3c2b630
@zhuzilin
[ci] borrow bot-slash-lint.yaml from miles (THUDM#1384)
7573b2c
@fy1214 @GeLee-Q
@Gao016 @yefei12
feat: add int4 reinforcement learning training support (Part2) (THUDM…
d2ec7ca 
…#1172)

Co-authored-by: GeLee-Q <leege233@gmail.com>
Co-authored-by: Gao016 <yngao016@163.com>
Co-authored-by: yefei12 <xjtu_yefeichen@163.com>
@Gao016 @fy1214
@GeLee-Q @yefei12 @yzlnew
feat: add int4 reinforcement learning training support (Part3) (THUDM…
931b050 
…#1368)

Co-authored-by: gaochang <gaochang@U-19PX2WQ1-0350.local>
Co-authored-by: fy1214 <282598660@qq.com>
Co-authored-by: GeLee-Q <865038696@qq.com>
Co-authored-by: yefei12 <xjtu_yefeichen@163.com>
Co-authored-by: yzlnew <yzlnew@gmail.com>
@zhuzilin
fix lint (THUDM#1385)
78a7bdd
@gxlvera
VLM Multi-turn, add Megatron in README (THUDM#1387)
9457895
@zhaochenyang20
Fix grammar and formatting in README.md (THUDM#1388)
f83559f
@zhuzilin
[refactor] refactor int4 qat code (THUDM#1390)
39a8c2f
@yueming-yuan
[1/X] Refactor: unify training backends by general utils, tested Mega…
9c909a4 
…tron & FSDP alignment (THUDM#1373)
@zhuzilin
bugfix (THUDM#1394)
997d620
@zhuzilin
bugfix (THUDM#1395)
f7d7663
@yueming-yuan
fix ppo utils & mis (THUDM#1396)
e47cecb
@Lawhy
feat(examples): add strands-sglang integration for agentic RL with TI…
52507f5 
…TO support (THUDM#1359)
@zhuzilin
remove swe-agent example (THUDM#1397)
7e56357
@Gao016 @GeLee-Q
@yefei12 @yzlnew
fix to suppot dpsk-v3.2 bf16 weight convert to fp8 (THUDM#1392)
e6151d8 
Co-authored-by: gaochang <gaochang@U-19PX2WQ1-0350.local>
Co-authored-by: GeLee-Q <865038696@qq.com>
Co-authored-by: yefei12 <xjtu_yefeichen@163.com>
Co-authored-by: yzlnew <yzlnew@gmail.com>
@zhuzilin
Only rank0 should call post_process_weights (THUDM#1398)
dd70beb
@zhuzilin
bugfix (THUDM#1400)
5f7cd38
@zhuzilin
[docker] support mtp for dpsk v3.2
9b05248
@zhuzilin
[docker] support mtp in dpsk v3.2 (THUDM#1401)
3464ff6
@lilei199908
bug gix (THUDM#1403)
c4ce76a
@yueming-yuan
[style] minor: remove subclass (THUDM#1402)
d29c394
@zhuzilin
[doc] suggest pip install -e . --no-deps (THUDM#1405)
bb3960f
@zhuzilin
[docs] add note for cudnn (THUDM#1406)
13e5dab
@yueming-yuan
[minor] Delete unused util file (THUDM#1408)
ca3ed90
@zhuzilin
[docker] ignore slime in sglang SafeUnpickler (THUDM#1409)
c3fbe10
@lilei199908
add r3 ci (THUDM#1407)
8c41473
@fzyzcjy
Fix rollout-all-samples (THUDM#1410)
f80bc39
@zijiexia
[docs] fix ai generate response (THUDM#1412)
2826b8f
@JessicaJiang-123
Merge remote-tracking branch 'upstream/main' into zhiyaoj/solve-conflict
c0034a1
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.