Commits · f315374fdcdde7115587d3dc422e2ce7c1b43b19 · OpenDAS / dynamo

05 Dec, 2025 3 commits
- docs: Guide for Speculative Decoding in VLLM using Eagle3 and Meta-Llama-3.1-8B-Instruct (#3895) · f315374f
  Dilreet Raju authored Dec 05, 2025
```
Signed-off-by: DilreetRaju <dilreetraju@gmail.com>
```
  f315374f
- fix(sglang): fix IPv6 support for ZMQ endpoint (#4403) · 3dbab3f1
  dmitrygx authored Dec 05, 2025
```
Signed-off-by: Dmitrii Gabor <dmitryg1709@gmail.com>
```
  3dbab3f1
- feat: Add python-configurable engine routes for sglang profiling (#4617) · 3bc129dc
  William Arnold authored Dec 05, 2025
```
Signed-off-by: William Arnold <7565007+Aphoh@users.noreply.github.com>
```
  3bc129dc
03 Dec, 2025 6 commits
- feat: add more metrics in planner (#4710) · 732c1ed2
  Hongkuan Zhou authored Dec 03, 2025
```
Signed-off-by: hongkuanz <hongkuanz@nvidia.com>
```
  732c1ed2
- chore: bump vLLM to 0.11.2 (#4476) · 60feb955
  Karen Chung authored Dec 03, 2025
  
  60feb955
- fix: adjust usage of vLLM for deprecation of 'disable_log_requests' (#4659) · db5687f5
  Julien Debache authored Dec 03, 2025
```
Co-authored-by: Ryan McCormick <rmccormick@nvidia.com>
```
  db5687f5
- fix: enable LMCache metrics visibility with PROMETHEUS_MULTIPROC_DIR (#4654) · c6555852
  Keiven C authored Dec 02, 2025
```
Signed-off-by: Keiven Chang <keivenchang@users.noreply.github.com>
Co-authored-by: Keiven Chang <keivenchang@users.noreply.github.com>
```
  c6555852
- chore: clearer prometheus port handling in planner (#4707) · 135bce43
  Hongkuan Zhou authored Dec 02, 2025
```
Signed-off-by: hongkuanz <hongkuanz@nvidia.com>
```
  135bce43
- test: bring back the framework 1 gpu pre-merge tests + clean up pytest markers (#4698) · 7e499b5c
  Yan Ru Pei authored Dec 02, 2025
```
Signed-off-by: PeaBrane <yanrpei@gmail.com>
```
  7e499b5c
02 Dec, 2025 5 commits
- feat: Extend the KV Event Consolidator to support TensorRT-LLM (#4533) · 6067c59d
  Kris Hung authored Dec 02, 2025
```
Signed-off-by: krishung5 <krish@nvidia.com>
```
  6067c59d
- fix: ModelDeploymentCard obtains full set of eos_token_ids by taking union... · 4ace4c85
  GuanLuo authored Dec 02, 2025
```
fix: ModelDeploymentCard obtains full set of eos_token_ids by taking union from different files (#3192)
Signed-off-by: Guan Luo <gluo@nvidia.com>
Signed-off-by: Guan Luo <41310872+GuanLuo@users.noreply.github.com>
```
  4ace4c85
- feat: mocker <> planner profiler (#4651) · 2b5655fd
  Hongkuan Zhou authored Dec 02, 2025
```
Signed-off-by: hongkuanz <hongkuanz@nvidia.com>
```
  2b5655fd
- feat: Adding --dyn-endpoint-types flag (#4619) · 03070fd5
  KrishnanPrash authored Dec 02, 2025
```
Signed-off-by: Krishnan Prashanth <kprashanth@nvidia.com>
Signed-off-by: KrishnanPrash <140860868+KrishnanPrash@users.noreply.github.com>
```
  03070fd5
- feat: lora - centralize lora cache key, restructure folders, s3 resiliency (#4644) · 71f94eda
  Biswa Panda authored Dec 02, 2025
  
  71f94eda
01 Dec, 2025 1 commit

test: Add pytest markers (#4111) · 0f6dca6e

Pavithra Vijayakrishnan authored Dec 01, 2025


Signed-off-by: pvijayakrish <pvijayakrish@nvidia.com>
Signed-off-by: Pavithra Vijayakrishnan <160681768+pvijayakrish@users.noreply.github.com>

0f6dca6e

26 Nov, 2025 2 commits
- feat: add LoRA common APIs and implementation for lora management (#4464) · 6a0e67ed
  Biswa Panda authored Nov 26, 2025
  
  6a0e67ed
- feat: add trtllm env variables + gpu part numbers to config dump (#3912) · dfbcdbe0
  William Arnold authored Nov 25, 2025
```
Signed-off-by: William Arnold <7565007+Aphoh@users.noreply.github.com>
```
  dfbcdbe0
25 Nov, 2025 4 commits
- chore: make loading planner profile data one-stage for mockers (#4590) · 7a384793
  Yan Ru Pei authored Nov 25, 2025
```
Signed-off-by: PeaBrane <yanrpei@gmail.com>
```
  7a384793
- feat: Enable intel gaudi on dynamo (#4209) · d0e3b7b7
  Sihan Chen authored Nov 26, 2025
```
Signed-off-by: Spycsh <sihan.chen@intel.com>
```
  d0e3b7b7
- fix: Fix vllm MM aggregated mode Corner Case (#4583) · c90e3dff
  Indrajit Bhosale authored Nov 25, 2025
  
  c90e3dff
- feat: Add security flag to MM flow in vllm (#4556) · 550bf98c
  Indrajit Bhosale authored Nov 24, 2025
```
Co-authored-by: KrishnanPrash <140860868+KrishnanPrash@users.noreply.github.com>
```
  550bf98c
22 Nov, 2025 2 commits

chore: upgrade trtllm 1.2.0rc2 (#4405) · 9b8b9988

Dmitry Tokarev authored Nov 21, 2025


Signed-off-by: Dmitry Tokarev <dtokarev@nvidia.com>
Co-authored-by: tanmayv25 <tanmay2592@gmail.com>
Co-authored-by: Kyle McGill <kmcgill@nvidia.com>
Co-authored-by: Ryan McCormick <rmccormick@nvidia.com>
Co-authored-by: Tanmay Verma <tanmayv@nvidia.com>

9b8b9988

chore: unify LMCache UX w/ and w/o Dynamo (#4509) · 6d69e8c7
Ziqi Fan authored Nov 21, 2025
```
Signed-off-by: Ziqi Fan <ziqif@nvidia.com>
```
6d69e8c7

21 Nov, 2025 2 commits
- chore: merge KvIndexer and ApproxKvIndexer (#4500) · c61e0dd3
  Yan Ru Pei authored Nov 21, 2025
```
Signed-off-by: PeaBrane <yanrpei@gmail.com>
```
  c61e0dd3
- fix: set `SGLANG_BLOCK_NONZERO_RANK_CHILDREN` to be no-op (#4527) · 6b432625
  ishandhanani authored Nov 20, 2025
  
  6b432625
20 Nov, 2025 1 commit
- fix: expose LMCache metrics via Dynamo metrics endpoint (#4461) · 75bf7c9b
  Keiven C authored Nov 19, 2025
```
Signed-off-by: Keiven Chang <keivenchang@users.noreply.github.com>
```
  75bf7c9b
19 Nov, 2025 3 commits
- feat: Enable autodeploy as a backend for TRT-LLM (#4347) · fe21d3dd
  William Zhang authored Nov 19, 2025
```
Signed-off-by: William Zhang <133824995+2ez4bz@users.noreply.github.com>
```
  fe21d3dd
- feat: Only monitor NATS metrics if using NATS request plane (#4442) · 69797b5a
  Graham King authored Nov 19, 2025
```
Signed-off-by: Graham King <grahamk@nvidia.com>
```
  69797b5a
- fix: Allow users to set `--kv-transfer-config` for vLLM (#4317) · 45529d06
  Joe Chandler authored Nov 18, 2025
```
Signed-off-by: Joe Chandler <jchandler@glia-ai.com>
```
  45529d06
18 Nov, 2025 7 commits
- feat: Support a dynamic default max_tokens for VLLM backend (#4156) · 58405177
  bin.pan authored Nov 19, 2025
```
Signed-off-by: bin <bin.pan@daocloud.io>
```
  58405177
- fix: respect router env vars in frontend configuration (#4431) · c65f1ac0
  Vladislav Nosivskoy authored Nov 19, 2025
```
Signed-off-by: Vladislav Nosivskoy <vladnosiv@gmail.com>
```
  c65f1ac0
- feat: support prompt_tokens_details in usage (#4239) · be9d6b2b
  Vladislav Nosivskoy authored Nov 19, 2025
```
Signed-off-by: Vladislav Nosivskoy <vladnosiv@gmail.com>
```
  be9d6b2b
- fix: Mark enable_kv_events=True for user-specified vLLM KVEventsConfig in scripts (#4418) · 7b967c77
  Karen Chung authored Nov 18, 2025
```
Signed-off-by: Karen Chung <karenc@nvidia.com>
```
  7b967c77
- feat: mocker can use planner profile data (#4422) · fc2ad4eb
  Yan Ru Pei authored Nov 18, 2025
```
Signed-off-by: PeaBrane <yanrpei@gmail.com>
Signed-off-by: Yan Ru Pei <yanrpei@gmail.com>
Co-authored-by: Hongkuan Zhou <tedzhouhk@gmail.com>
```
  fc2ad4eb
- fix: Move trtllm multimodal worker to use ModelInput.Token (#4373) · c3984bc2
  Indrajit Bhosale authored Nov 17, 2025
```
Signed-off-by: Indrajit Bhosale <iamindrajitb@gmail.com>
```
  c3984bc2
- fix: Fix port collision and pass prefill param in kvbm connector (#4411) · e3cee95f
  Kris Hung authored Nov 17, 2025
```
Signed-off-by: krishung5 <krish@nvidia.com>
```
  e3cee95f
17 Nov, 2025 3 commits
- fix: workaround vllm prometheus cleanup error on shutdown (#4323) · 6afb4d5b
  Keiven C authored Nov 17, 2025
```
Signed-off-by: Keiven Chang <keivenchang@users.noreply.github.com>
Co-authored-by: Keiven Chang <keivenchang@users.noreply.github.com>
```
  6afb4d5b
- feat: convert profiler perf data for mocker to use (#4370) · 24af5a33
  Hongkuan Zhou authored Nov 17, 2025
```
Signed-off-by: hongkuanz <hongkuanz@nvidia.com>
```
  24af5a33
- feat: Command line flag to set request plane mode: tcp, http or nats (#4365) · 886506c1
  Graham King authored Nov 17, 2025
```
Signed-off-by: Graham King <grahamk@nvidia.com>
```
  886506c1
14 Nov, 2025 1 commit
- feat: Use try-catch block for import to avoid error (#4339) · 9a765d6d
  Tzu-Ling Kan authored Nov 14, 2025
```
Signed-off-by: tzulingk@nvidia.com <tzulingk@nvidia.com>
```
  9a765d6d