InternLM/lmdeploy v0.12.3 on GitHub

What's Changed

Support video inputs by @CUHKSZzxy in #4360
feat: fully implement compressed-tensors gs32 support in TurboMind by @lapy in #4429
Draft model update params by @CUHKSZzxy in #4452

support qwen3.5 on volta by @grimoire in #4405
Optimize Qwen3.5 by @lzhangzz in #4434
Builtin mrope by @grimoire in #4393
delete ray remote function return value by @grimoire in #4422
support cache_seqlen on recurrent-gdr and causal-conv1d-update by @grimoire in #4417
safe ray api by @CUHKSZzxy in #4455
add R3 for qwen3-vl-moe models by @lvhan028 in #4457
Align rope init in lmdeploy by @RangiLyu in #4466
Make tilelang a Linux-only dependency (like triton) by @Copilot in #4469
prepare chunk indices before cache initialize by @grimoire in #4458
unify rope device by @CUHKSZzxy in #4467
custom processor args by @CUHKSZzxy in #4472
Assign sequential api_server ports when proxy_url is unset by @lvhan028 in #4416
disable fla intracard_backend by @grimoire in #4482
[Fix][Feat] Fix worker sorting with external pg bundles & Support persistent buffer for update_params by @CyCle1024 in #4397
simplify interns1 pro codes by @CUHKSZzxy in #4480

fix test_hf_overrides for transformers>5 by @grimoire in #4418
fix qwen3.5 pytorch multimodal inference by @CUHKSZzxy in #4430
fix generate endpoint by @CUHKSZzxy in #4432
Make Intern-S1-Pro compatible with Transformers 5.0+ by @lvhan028 in #4435
fix multiround chat by @CUHKSZzxy in #4438
fix(async_engine): make safe_run cancellation cleanup reliable with shield and SafeRunException by @lvhan028 in #4439
release state cache by @CUHKSZzxy in #4462
Split/tool call args json for qwen3coder tool calls (Qwen3.5) by @lapy in #4433
fix(turbomind): fix dimension mismatch in ApplyTokenBitmaskInplace by @windreamer in #4456
fix metrics by @CUHKSZzxy in #4410
fix security issues by @CUHKSZzxy in #4447
fix qwen3.5 fp8 support by @grimoire in #4470
fix image / video resize function by @CUHKSZzxy in #4478
fix dynamic ntk device by @CUHKSZzxy in #4483
fix pagedattention pointer range by @grimoire in #4494
fix glm4.7-flash by @grimoire in #4500
Fix torch awq by @grimoire in #4503

[ci] add legacy test workflow and test config by @zhulinJulia24 in #4387
chore: add CLAUDE.md and Claude Code skills by @CUHKSZzxy in #4413
Fix CI errors including linting error and unit test error by @lvhan028 in #4431
Use pyupgrade and ruff to modernize LMDeploy Python Code by @windreamer in #4392
reduce ci memory by @irexyc in #4471
fix: add safe.directory for git in docker workflows by @windreamer in #4474
[ci] add nightly docker build workflow by @zhulinJulia24 in #4406
split docker wheel preparation into staged build steps and use python 3.12 as the default version by @lvhan028 in #4476
[Feat]: Support qwen35 with mtp by @RunningLeon in #4437
bump version to v0.12.3 by @lvhan028 in #4493

Full Changelog: v0.12.2...v0.12.3