* [Feature] support logprobs in chat/completions and completions endpoints
* Temporarily comment out text_offset due to incorrect logic
* Clean up temporary debug prints
* [Feature] support logprobs in offline mode via SamplingParams
* fix: serialize Logprob as dict before zmq send to fix msgpack error
* refactor: remove redundant methods to simplify codebase
* Fix missing fields in CompletionOutput.to_dict affecting msgpack serialization
* refactor: centralize param validation in engine_client to reduce duplication
* revert: rollback changes in offline_demo.py
* revert: rollback changes in offline_demo.py
* [bugfix] fix parameter validation for logprobs
* [bugfix] fix parameter validation for logprobs
* [bugfix] fix parameter validation for logprobs
* [bugfix] fix parameter validation for logprobs
---------
Co-authored-by: Jiang-Jia-Jun <163579578+Jiang-Jia-Jun@users.noreply.github.com>
* support bad_words
* support online infer bad_words
* update
* add CI test
* update
* update
* update
---------
Co-authored-by: Yuanle Liu <yuanlehome@163.com>
* support repetition early stop and support user to set the parameter
* remove log
* fix codestyle
* add the early_stop_config to rollout_config
* update config and EarlyStopper class
* fix the bug for triton
* modify the stop method
* update description
* modify the usage for stop_flags
---------
Co-authored-by: Yuanle Liu <yuanlehome@163.com>
* Support FD block scheduler v1
* Support FD block scheduler v1
* Support FD block scheduler v1
* Fix according to copilot review
* Fix according to review
* Remove is_dummy
* Fix bug when real_bsz=1
* Fix infer first token cost time
---------
Co-authored-by: Jiang-Jia-Jun <163579578+Jiang-Jia-Jun@users.noreply.github.com>
* fix the bug in cudagraph+prefix-caching but still have some bug with profile
Change-Id: Ibf2ba3f2e3b08641d03f4b1391d7c862c3efa397
* add the signal to make sure cache manager launched
* fix judge condition
* reomove useless control
* update control stream
* update
* fix xpu
* change the do_profile flag
* update
* add new threads to init cache_manager
---------
Co-authored-by: RAM <gstian5555@outlook.com>
* [SOT] Mark dynamic dims by type annotations
* fix conflict of forward_meta
* mark more attn backend
* fix missing annotated and add env SOT_SPECIALIZED_DIM_NUMBERS
* auto infer implicit 0 dim dynamic dim
* revert manual marked dims
* revert missing update
* auto infer can use unsafe code in warmup stage
* check -> type_match
* fix codestyle
* restore blank line
* empty commit
* add need_warmup nonlocal;
* add doc for resolver
* add missing type hints
* unquote "ForwardMeta"
* Support enabling cuda graph and custom all reduce at the same time, and fix the overwritten custom all reduce flag
* rename communication_op to communication
* cuda graph support padding batch
* Integrate the startup parameters for the graph optimization backend and provide support for user - defined capture sizes.
* Do not insert max_num_seqs when the user specifies a capture list
* Support set graph optimization config from YAML file
* update cuda graph ci
* fix ci bug
* fix ci bug
* Check if the controller port is available
* Global scheduler supports configuring hot updates
* add interface: /controller/scheduler
* add interface: /controller/scheduler
* [LLM] support multi node deploy
* Update engine.py
* fix bugs
* fix
* [LLM] support multi node deploy
* [LLM] support multi node deploy
---------
Co-authored-by: Jiang-Jia-Jun <163579578+Jiang-Jia-Jun@users.noreply.github.com>
