123 Pull requests merged by 39 people

• [RLlib] APPO accelerate (vol 17): LearnerGroup should not pickle remote functions on each update-call; Refactor LearnerGroup and Learner APIs.

  #50665 merged Mar 11, 2025

• [Data] Always pin the seed when doing file-based random shuffle

  #50924 merged Mar 11, 2025

• Run workflow tests on postmerge only

  #51247 merged Mar 11, 2025

• Skip flaky workflow tests

  #51245 merged Mar 11, 2025

• [core] Remove "serve" and "tune" dependencies from test_runtime_env_complicated

  #51246 merged Mar 11, 2025

• clean up shutdown behavior of serve

  #51009 merged Mar 11, 2025

• [core] Use current node id if no node id specified for ray drain-node

  #51134 merged Mar 11, 2025

• [core] Skip setproctitle tests on non-Linux machines

  #51229 merged Mar 11, 2025

• [core] Fix client_connection test windows by trying old port

  #51232 merged Mar 11, 2025

• [train v2][doc] Update user guides for metrics, checkpoints, results, and experiment tracking

  #51204 merged Mar 11, 2025

• [core] [easy] [noop] Avoid duplicate function

  #51209 merged Mar 11, 2025

• Sync (most) CI linters to precommit

  #51181 merged Mar 11, 2025

• [tune] Fix RunConfig deprecation message in Tune being emitted in trainer.fit usage

  #51198 merged Mar 11, 2025

• Avoid setup-dev.py writing to generated/

  #51194 merged Mar 11, 2025

• [core][refactor] Move accelerator-specific environment variables to ray_constants.py to avoid redefining them

  #51026 merged Mar 11, 2025

• [ray.llm] Refactor download utilities

  #51225 merged Mar 11, 2025

• [core][refactor] Move node_stats_to_dict to utils.py to avoid importing unnecessary modules

  #51187 merged Mar 10, 2025

• [Data] Change default batch size from 1024 to None

  #50564 merged Mar 10, 2025

• [core] Deflake test_object_reconstruction_pending_creation

  #51224 merged Mar 10, 2025

• [Data] Make chunk combination threshold configurable for improved per…

  #51200 merged Mar 10, 2025

• [core] Mask tune directory for core tests

  #51186 merged Mar 10, 2025

• [core] Refactor test_run_driver_twice to not use tune

  #51220 merged Mar 10, 2025

• [core][autoscaler] Add Pod names to the output of ray status -v

  #51192 merged Mar 10, 2025

• [doc][core][cgraph] Add bind() to API page

  #51044 merged Mar 10, 2025

• [core][cgraph] Clean up CompiledDAG.actor_refs

  #51174 merged Mar 10, 2025

• [core] Mask rllib directory for core tests

  #50618 merged Mar 10, 2025

• [core][cgraph][doc]Fix cgraphs gpu docs code again

  #51208 merged Mar 10, 2025

• [core] Wait for DisconnectClientReply in worker shutdown sequence

  #51033 merged Mar 10, 2025

• [RLlib] Unify namings for actor managers' outstanding in-flight requests metrics.

  #51159 merged Mar 10, 2025

• [core] Remove AIR util from test_ray_init_2

  #51215 merged Mar 10, 2025

• [core] Move ray.experimental.tf_utils back to rllib

  #51183 merged Mar 10, 2025

• [CI] use smaller object_manager deps in targets

  #51062 merged Mar 10, 2025

• [core] Acquire GIL before calling PyGILState_Release

  #51203 merged Mar 10, 2025

• [core] Cover cpplint for /src/ray/object_manager (excluding plasma)

  #51089 merged Mar 10, 2025

• [core][compiled graphs] Controllably destroy CUDA events in GPUFutures

  #51090 merged Mar 10, 2025

• [core] Changing default tensor serialization in compiled graphs

  #50778 merged Mar 9, 2025

• [core] Remove unused AsyncDrainNode

  #51043 merged Mar 9, 2025

• [core][dashboard] Lazy init NotifyQueue

  #51178 merged Mar 9, 2025

• [train v2] Create a default ScalingConfig if one is not provided to the trainer

  #51093 merged Mar 9, 2025

• Remove tune-specific task events test

  #51185 merged Mar 8, 2025

• [Data] Add fixed-size compute for read, write, and batch inference release tests

  #51176 merged Mar 8, 2025

• [telemetry] Move library usage tests out of core

  #51161 merged Mar 8, 2025

• [serve.llm] Trim probes and reorg the llm.serve release test folder structure

  #51112 merged Mar 8, 2025

• [core] Use the correct way to check whether an actor task is running or not

  #51158 merged Mar 8, 2025

• Fix UV hook to support Ray Job submission

  #51150 merged Mar 7, 2025

• [train][v2] add log file paths to train state

  #51085 merged Mar 7, 2025

• [train v2][doc] Add updated Train + Tune user guide

  #51048 merged Mar 7, 2025

• Remove rllib dependency from core tests

  #51171 merged Mar 7, 2025

• [core] Pass explicit handler for ClientConnection errors

  #51098 merged Mar 7, 2025

• [train v2][doc] Add updated fault tolerance user guide

  #51083 merged Mar 7, 2025

• [data] fix the new dataset metrics

  #51148 merged Mar 7, 2025

• [data] Handle schema mismatch in multiple blocks into one parquet file

  #50915 merged Mar 7, 2025

• [core] Use psutil.process_iter to replace psutil.pids

  #51123 merged Mar 7, 2025

• [Core][Doc] Add Document for Python Standard Attributes

  #51038 merged Mar 7, 2025

• [Data] Make num_blocks in repartition optional

  #50997 merged Mar 7, 2025

• [serve] Fix bad mock import

  #51170 merged Mar 7, 2025

• [doc][core] fix a wrong url in ray-dag.rst

  #49980 merged Mar 7, 2025

• [core][cgraph][docs] Fix cgraph gpu docs code

  #51132 merged Mar 7, 2025

• [core] Block before running gpu tests

  #51131 merged Mar 7, 2025

• [ray.data.llm] Add release test

  #51063 merged Mar 7, 2025

• [core] Fix path comparison on windows

  #51142 merged Mar 7, 2025

• [CI] Enable python-no-log-warn precommit rule

  #51099 merged Mar 7, 2025

• [Fix][Core] Execute user requested actor exit in C++ side

  #49918 merged Mar 7, 2025

• [data] fix non-unique operator id

  #51108 merged Mar 7, 2025

• [core][cgraph] Fix doc test cgraph_overlap

  #51139 merged Mar 7, 2025

• [data] Add dataset/operator state, progress, total metrics

  #50770 merged Mar 7, 2025

• [core] Fix double checked locking

  #51073 merged Mar 6, 2025

• [docs] Hide output of very verbose ipynb notebooks

  #51140 merged Mar 6, 2025

• [Core][Autoscaler] Refactor v2 Log Formatting

  #49350 merged Mar 6, 2025

• [serve] Exclude redirects from request error count

  #51130 merged Mar 6, 2025

• [docs] change order of libraries based on popularity

  #51136 merged Mar 6, 2025

• [Data] Fix error message for override_num_blocks when reading from a HuggingFace Dataset

  #50998 merged Mar 6, 2025

• use correct mean and standard deviation norm values in image tutorials

  #50240 merged Mar 6, 2025

• Fix typos in comments and strings

  #51079 merged Mar 6, 2025

• [ci] Revert UV pip compile for LLM requirements

  #51118 merged Mar 6, 2025

• [serve.llm] Added benchmark release tests

  #51106 merged Mar 6, 2025

• Revert "[Doc] RayServe Single-Host TPU v6e Example with vLLM (#47814)"

  #51113 merged Mar 6, 2025

• Discover TPU logs in Ray Dashboard

  #47737 merged Mar 6, 2025

• [core] delete UNORDERED_VS_ABSL_MAPS_EVALUATION

  #51115 merged Mar 6, 2025

• [core] Add comments explaining why unique_ptr::release() is used.

  #51055 merged Mar 6, 2025

• [RLlib; docs] Fix 4 broken html links.

  #51015 merged Mar 6, 2025

• Revert "[core] Add a warning for non-contiguous numpy array deserialization"

  #51111 merged Mar 6, 2025

• Various improvements to the Getting Started page

  #50878 merged Mar 6, 2025

• [docs,tune] fix typo, and standardize equations across the two apis

  #51114 merged Mar 6, 2025

• [doc] Add documentation for Asynchronous HyperBand Example in Tune

  #50708 merged Mar 6, 2025

• Improvements to PBT example

  #50870 merged Mar 6, 2025

• Various enhancements to the Gradio Ray Serve tutorial

  #50276 merged Mar 6, 2025

• [bugfix] Fixed the issues with compile llm reqs script

  #51110 merged Mar 6, 2025

• [core][cgraph][docs] Move cgraphs docs code to test

  #51000 merged Mar 6, 2025

• Various improvements to Tune Pytorch CIFAR tutorial

  #50316 merged Mar 6, 2025

• [core][cgraph][docs] Compiled Graph troubleshooting

  #51030 merged Mar 5, 2025

• [train v2] Use the FailurePolicy factory

  #51067 merged Mar 5, 2025

• [core] Implement call once token

  #51053 merged Mar 5, 2025

• [core] Cover cpplint for src/ray/object_manager/plasma

  #50954 merged Mar 5, 2025

• Improvements to HF Transformers example

  #50896 merged Mar 5, 2025

• [Doc] RayServe Single-Host TPU v6e Example with vLLM

  #47814 merged Mar 5, 2025

• [core] Enable make_shared in plasma client

  #51100 merged Mar 5, 2025

• Fixed non-runnable Optuna tutorial

  #50404 merged Mar 5, 2025

• Various improvements to the Ray Tune XGBoost tutorial

  #50455 merged Mar 5, 2025

• [doc] Update jemalloc profiling doc

  #51031 merged Mar 5, 2025

• [core][2/N] Remove redis_max_memory from Ray core

  #51059 merged Mar 5, 2025

• Various enhancements to Tune Keras example:

  #50581 merged Mar 5, 2025

• minor improvements to hyperopt tutorial

  #50697 merged Mar 5, 2025

• various improvements to lightgbm tutorial:

  #50704 merged Mar 5, 2025

• Improvements to Train DeepSpeed example

  #50906 merged Mar 5, 2025

• [serve] Skip locality backoff unit test on windows

  #51101 merged Mar 5, 2025

• [Docs][Data] Ordering of rows

  #50986 merged Mar 5, 2025

• [core] Remove unnecessary ClientHandler

  #51097 merged Mar 5, 2025

• [core] Checkin opentelemetry C++ sdk dependency

  #51077 merged Mar 5, 2025

• [RLlib] Add 'single_action_space' and 'single_observation_space' to 'VectorMultiAgentEnv`.

  #51096 merged Mar 5, 2025

• [core] Fix plasma client memleak

  #51051 merged Mar 5, 2025

• [data] Update loading-data.rst

  #41972 merged Mar 5, 2025

• [core] BoundedExecutor is never used when an actor is async

  #51069 merged Mar 5, 2025

• [train] Deprecate torch amp wrapper utilities

  #51066 merged Mar 5, 2025

• [train][v2] implement state export

  #50622 merged Mar 4, 2025

• [data] Adding in another kwargs

  #51068 merged Mar 4, 2025

• [LLM APIs] Fast follow up for 2.44 (2/N)

  #51064 merged Mar 4, 2025

• [train] Remove ray storage dependency and deprecate RAY_STORAGE env var configuration option

  #50872 merged Mar 4, 2025

• [core] Don't build cpp api on pip install

  #50499 merged Mar 4, 2025

• Fix typos in doc directory

  #51054 merged Mar 4, 2025

• [Docs] update all hpu related docs

  #51028 merged Mar 4, 2025

• [core] Add a warning for non-contiguous numpy array deserialization

  #50731 merged Mar 4, 2025

• [train v2] Run tensorflow release test with v2 enabled

  #51046 merged Mar 4, 2025

63 Pull requests opened by 38 people

• try remote debug

  #51070 opened Mar 4, 2025

• [Docs] Update docs to reflect CPU requests/limits change in KubeRay v1.3

  #51072 opened Mar 4, 2025

• [data] Make Dataset.name/set_name public

  #51076 opened Mar 4, 2025

• [DONOTMERGE] POC for Ray+torch.distributed

  #51078 opened Mar 4, 2025

• Fix the grammar of the OOM killer error messages

  #51081 opened Mar 5, 2025

• [Data] Adding in metrics for number of actors alive, pending and restarting

  #51082 opened Mar 5, 2025

• [ci] bazel lint all BUILD files - `python/`

  #51092 opened Mar 5, 2025

• [Doc][KubeRay] Add a doc to explain why some worker Pods are not ready in RayService

  #51095 opened Mar 5, 2025

• [ray.data.llm] Support S3 paths for model checkpoint and LoRA path

  #51103 opened Mar 5, 2025

• Replace AMD device env var with HIP_VISIBLE_DEVICES

  #51104 opened Mar 5, 2025

• [core] Integrate process-wise cgroup setup with task execution

  #51107 opened Mar 6, 2025

• [DO NOT MERGE] Train v2 bug bash build

  #51116 opened Mar 6, 2025

• [core][autoscaler][v1] do not removing nodes for upcoming placement groups

  #51122 opened Mar 6, 2025

• [Core][Bug fix] Trigger local task scheduling after deleting bundle.

  #51125 opened Mar 6, 2025

• [Data] Store average memory use per task in `OpRuntimeMetrics`

  #51126 opened Mar 6, 2025

• [Data] Add `memory` attribute to `ExecutionResources`

  #51127 opened Mar 6, 2025

• [do not merge] Add Daft to the Ray ecosystem page

  #51133 opened Mar 6, 2025

• [VMware][WCP provider][Part 2/n]: Add vSphere WCP node provider

  #51138 opened Mar 6, 2025

• [core] Check cgroupv2 mount status

  #51141 opened Mar 6, 2025

• Reproducing MacOS x86_64 Test Failure w/ Custom Numpy Serializer for ndarrays

  #51143 opened Mar 6, 2025

• add additional_log_standard_attrs to serve logging config

  #51144 opened Mar 7, 2025

• Update CODEOWNERS

  #51146 opened Mar 7, 2025

• [LLM Batch][Telemetry] Add telemetry for batch API

  #51147 opened Mar 7, 2025

• [core] Implement a universal printer

  #51151 opened Mar 7, 2025

• [Do Not Merge] Update the Test Script to Debug test_network_failure_e2e Flaky Test

  #51153 opened Mar 7, 2025

• [Doc] Update vllm example with metrics

  #51156 opened Mar 7, 2025

• [RLlib] Add timers to env step, forward pass, and complete connector pipelines runs.

  #51160 opened Mar 7, 2025

• Bump axios from 0.21.4 to 1.8.2 in /python/ray/dashboard/client

  #51162 opened Mar 7, 2025

• [Feat][Core] Re-implement dashboard subprocess modules with aiohttp server/client and converts HealthzHead

  #51172 opened Mar 7, 2025

• [DONT MERGE] try remote debug

  #51175 opened Mar 7, 2025

• [train] train v1 export api

  #51177 opened Mar 8, 2025

• [core] Integrate scoped dup2

  #51179 opened Mar 8, 2025

• [Data] Support async callable classes in flat_map()

  #51180 opened Mar 8, 2025

• [Core] Making Object Store Fallback Directory Configurable

  #51189 opened Mar 9, 2025

• [core] Make testable stream redirection

  #51191 opened Mar 9, 2025

• [core] Cover cpplint for `ray/tree/master/src/ray/gcs/gcs_server`

  #51197 opened Mar 9, 2025

• [train v2] Improve `TrainingFailedError` message

  #51199 opened Mar 9, 2025

• [core][autoscaler] add start_node_delay_s config to the FakeMultiNodeProvider for testing edge cases of idle_timeout_minutes

  #51201 opened Mar 9, 2025

• [train v2][doc] Update persistent storage guide

  #51202 opened Mar 9, 2025

• [doc][core][cgraph] Complete Compiled Graph docs

  #51206 opened Mar 10, 2025

• [image] Add cuda 12.8.0 in image building matrix

  #51210 opened Mar 10, 2025

• [doc] Update vllm example support args without values, eg --enable-lora

  #51212 opened Mar 10, 2025

• Bump jinja2 from 3.1.3 to 3.1.6 in /release

  #51216 opened Mar 10, 2025

• [RLlib/Tune; docs] Fix broken PB2/RLlib example.

  #51219 opened Mar 10, 2025

• [Serve.llm] create new telemetry tags for serve.llm

  #51221 opened Mar 10, 2025

• [train v2][doc] Update API references

  #51222 opened Mar 10, 2025

• Fix the logic to calculate the number of workers based on the TPU version.

  #51227 opened Mar 10, 2025

• [Doc] Update documentation for `uv run`

  #51233 opened Mar 11, 2025

• [docs, tune] replace reuse actors example with a fuller demonstration

  #51234 opened Mar 11, 2025

• [serve.llm] Add gen-config to oss

  #51235 opened Mar 11, 2025

• [data] Support `ray_remote_args_fn` in map_groups

  #51236 opened Mar 11, 2025

• [Data] Abstracting common `shuffle` utility

  #51237 opened Mar 11, 2025

• [Data] Avoid unnecessary conversion to Numpy when creating Arrow/Pandas blocks

  #51238 opened Mar 11, 2025

• [DONT MERGE] Hjiang/cgroup on ci status test

  #51239 opened Mar 11, 2025

• Update precommit hook

  #51240 opened Mar 11, 2025

• [core] Mask train directory for core tests

  #51248 opened Mar 11, 2025

• [Test][KubeRay] Add doctest for RayCluster Quickstart doc

  #51249 opened Mar 11, 2025

• [dashboard] Skip failing subprocess `test_e2e`

  #51253 opened Mar 11, 2025

• [Doc] update openai example to support LORA and multiple lora modules

  #51254 opened Mar 11, 2025

• [RLlib] - Remove `self` from `staticmethod` in `Connector` in old API stack.

  #51255 opened Mar 11, 2025

• Bump keras from 2.15.0 to 3.9.0 in /python

  #51256 opened Mar 11, 2025

• [DO NOT MERGE] Debug CI error by unpinning pandas version in test_e2e_complex

  #51257 opened Mar 11, 2025

• [ray.serve.llm] Fix release test

  #51258 opened Mar 11, 2025

42 Issues closed by 14 people

• CI test linux://rllib:examples/connectors/mean_std_filtering_ppo is flaky

  #47435 closed Mar 11, 2025

• `serve shutdown` command claims to have shut down the service even if one wasn't running.

  #34018 closed Mar 11, 2025

• CI test windows://python/ray/tests:test_task_events_2 is consistently_failing

  #45966 closed Mar 11, 2025

• CI test darwin://python/ray/tests:test_task_events_2 is consistently_failing

  #51182 closed Mar 11, 2025

• CI test windows://python/ray/tests:test_actor_retry2 is flaky

  #47415 closed Mar 11, 2025

• CI test linux://rllib:learning_tests_stateless_cartpole_appo_gpu is consistently_failing

  #47295 closed Mar 11, 2025

• CI test linux://rllib:learning_tests_multi_agent_pendulum_sac_multi_gpu is flaky

  #47309 closed Mar 11, 2025

• CI test darwin://python/ray/tests:test_state_api is consistently_failing

  #46890 closed Mar 11, 2025

• CI test linux://rllib:learning_tests_multi_agent_stateless_cartpole_ppo_multi_cpu is flaky

  #47313 closed Mar 11, 2025

• CI test darwin://python/ray/tests:test_reconstruction_2 is flaky

  #50859 closed Mar 11, 2025

• [Ray Core] `ray.data.Dataset.repartition` not working consistently with doc/error message

  #51060 closed Mar 10, 2025

• CI test linux://rllib:learning_tests_cartpole_dqn_multi_cpu is flaky

  #47214 closed Mar 10, 2025

• CI test windows://python/ray/tests:test_reference_counting_2 is consistently_failing

  #45964 closed Mar 10, 2025

• [serve] proxy memory leak

  #50927 closed Mar 10, 2025

• [CI] Replace `:object_manager` with smaller Bazel targets.

  #51014 closed Mar 10, 2025

• CI test windows://python/ray/tests:test_object_spilling_debug_mode is flaky

  #43796 closed Mar 9, 2025

• CI test darwin://python/ray/tests:test_advanced_5 is flaky

  #44005 closed Mar 8, 2025

• CI test linux://rllib:examples/learners/ppo_with_torch_lr_schedulers is flaky

  #49181 closed Mar 8, 2025

• CI test darwin://python/ray/tests:test_runtime_env_working_dir_3 is consistently_failing

  #44765 closed Mar 8, 2025

• [core] raylet memory leak

  #50955 closed Mar 8, 2025

• CI test linux://doc:doc_code_cgraph_nccl is consistently_failing

  #51119 closed Mar 7, 2025

• CI test linux://doc:doc_code_cgraph_profiling is consistently_failing

  #51121 closed Mar 7, 2025

• [core][logging] Customizable Python standard log attributes.

  #49502 closed Mar 7, 2025

• CI test linux://python/ray/serve/tests/unit:test_deployment_state_with_stop_fully_then_start_behavior is flaky

  #49642 closed Mar 7, 2025

• CI test linux://python/ray/serve/tests/unit:test_deployment_state is flaky

  #49632 closed Mar 7, 2025

• CI test linux://python/ray/serve/tests/unit:test_deployment_state_with_metr_disab is flaky

  #49638 closed Mar 7, 2025

• CI test linux://python/ray/serve/tests/unit:test_deployment_state_with_compact_scheduling is flaky

  #49634 closed Mar 7, 2025

• [telemetry] Using only library APIs should not cause "core" usage to be reported

  #51169 closed Mar 7, 2025

• [Core] Persist Ray runtime resources between sessions

  #37377 closed Mar 7, 2025

• CI test windows://python/ray/tests:test_logging is consistently_failing

  #51135 closed Mar 7, 2025

• CI test linux://doc:doc_code_cgraph_overlap is consistently_failing

  #51120 closed Mar 7, 2025

• [Core] ray.actor.exit_actor() does not seem to work from within an async background thread

  #49451 closed Mar 7, 2025

• [autoscaler] Refactor ray status output code

  #37856 closed Mar 6, 2025

• [Docs] Tutorials should use correct normalization values for image datasets

  #50239 closed Mar 6, 2025

• CI test linux://rllib:learning_tests_multi_agent_pendulum_sac_multi_cpu is flaky

  #47264 closed Mar 6, 2025

• [core] ray.wait check failed on static_cast(ready.size()) <= num_objects

  #51105 closed Mar 6, 2025

• [core] Cover cpplint for `src/ray/object_manager/plasma`

  #50729 closed Mar 5, 2025

• CI test windows://python/ray/serve/tests/unit:test_pow_2_replica_scheduler is flaky

  #47950 closed Mar 5, 2025

• [Data] Ordering of blocks after map and map_batches

  #50890 closed Mar 5, 2025

• CI test linux://python/ray/air:test_tensor_extension is flaky

  #50819 closed Mar 5, 2025

• pip cannot install ray on Alpine Linux

  #30416 closed Mar 5, 2025

• [Ray Data] Typo in "PyTorch DataLoader arguments" documentation

  #51074 closed Mar 4, 2025

50 Issues opened by 34 people

• Release test llm_serve_llama_3dot1_8B_lora failed

  #51252 opened Mar 11, 2025

• Release test llm_serve_llama_3dot1_8B_quantized_tp_1 failed

  #51251 opened Mar 11, 2025

• Release test llm_serve_llama_3dot1_8B_tp_2 failed

  #51250 opened Mar 11, 2025

• ModuleNotFoundError: No module named 'ray.serve.llm'

  #51244 opened Mar 11, 2025

• [Data] `StandardScaler.transform()` cannot handle non-`NAN` columns when `StandardScaler.stats_` is `NAN` for that column

  #51243 opened Mar 11, 2025

• [Serve] Detailed Analysis of Errors Related to 'Ray does not allocate any GPUs on the driver node' && 'No CUDA GPUs are available'

  #51242 opened Mar 11, 2025

• [Tune] Add examples of using XGBoost with the sklearn API and add examples of regression tasks

  #51241 opened Mar 11, 2025

• [core] Move to bzlmod for handling dependencies

  #51228 opened Mar 10, 2025

• CI test windows://src/ray/common/test:client_connection_test is consistently_failing

  #51226 opened Mar 10, 2025

• [Data] `Dataset.train_test_split` reads dataset twice

  #51223 opened Mar 10, 2025

• [Core] get_user_temp_dir() Doesn't Honor the User Specified Temp Dir

  #51218 opened Mar 10, 2025

• [Data] Filter operation changes schema of dataset

  #51217 opened Mar 10, 2025

• [Core] Ray core tasks tutorial not works. error msg: `Error Type: WORKER_DIED`

  #51214 opened Mar 10, 2025

• [Serve] how to print log in terminal and save to single custom file path

  #51213 opened Mar 10, 2025

• [Ray Core] For the same python test, the results of pytest and bazel are inconsistent

  #51211 opened Mar 10, 2025

• [Data] Adding streaming capability for `ray.data.Dataset.unique`

  #51207 opened Mar 10, 2025

• [Core] Add c++ API for setting max_task_retries

  #51205 opened Mar 10, 2025

• [Core] Failed to use uv

  #51196 opened Mar 9, 2025

• [Core] API Reference: uv

  #51195 opened Mar 9, 2025

• [Ray Serve]: "RuntimeError: No CUDA GPUs are available" when running vllm with ray

  #51193 opened Mar 9, 2025

• [Serve] Serve hangs for chained deployments with diamond dependencies data flow

  #51190 opened Mar 9, 2025

• [<Ray component: Data] Async map_batches return empty result when execution_options.preserve_order = True

  #51188 opened Mar 9, 2025

• [core] Cover cpplint for `ray/tree/master/src/ray/gcs/gcs_server`

  #51184 opened Mar 8, 2025

• [RFC] GPU object store support in Ray Core

  #51173 opened Mar 7, 2025

• [telemetry] RLlib usage should not report Ray Train usage

  #51168 opened Mar 7, 2025

• [telemetry] RLlib telemetry prior to `ray.init` is not reported

  #51167 opened Mar 7, 2025

• [telemetry] RLlib should not report usage on import

  #51166 opened Mar 7, 2025

• [telemetry] Importing Ray Tune in an actor reports Ray Train usage

  #51165 opened Mar 7, 2025

• [telemetry] Ray Train import prior to `ray.init` isn't reported

  #51164 opened Mar 7, 2025

• [telemetry] Ray Train should not be marked only from import

  #51163 opened Mar 7, 2025

• [Ray debugger] Unable to use debugger on slurm cluster

  #51157 opened Mar 7, 2025

• Custom policy

  #51155 opened Mar 7, 2025

• [Serve] On kuberay, vLLM-0.7.2 reports "No CUDA GPUs are available" while vllm-0.6.6.post1 works fine when deploy rayservice

  #51154 opened Mar 7, 2025

• [core] Place unit test alongside with the implementation

  #51152 opened Mar 7, 2025

• [core] Named placement groups don't work

  #51149 opened Mar 7, 2025

• [Serve] `ray.serve.exceptions.BackpressurError` raised from child cause `500` with FastAPI Ingress

  #51145 opened Mar 7, 2025

• [RLLIB] Support for Gymnasium Graph Spaces

  #51129 opened Mar 6, 2025

• [Core] RayCheck failed: placement_group_resource_manager_->ReturnBundle(bundle_spec) Status not OK

  #51124 opened Mar 6, 2025

• [core] Unstable placement group GPU bundle order

  #51117 opened Mar 6, 2025

• [Doc][Dashboard] Add Documentation about TPU Logs

  #51102 opened Mar 5, 2025

• [Data]Extend Ray Data with read/write hive

  #51094 opened Mar 5, 2025

• [ci] bazel lint all BUILD files - `python`

  #51091 opened Mar 5, 2025

• [core] Integrate clang-tidy

  #51088 opened Mar 5, 2025

• [core] Integrate gcov into our build system

  #51087 opened Mar 5, 2025

• [core] Guard ray C++ code quality via unit test

  #51086 opened Mar 5, 2025

• [core] Cover cpplint for `/src/ray/object_manager` (excluding `plasma`)

  #51084 opened Mar 5, 2025

• [core] Split raylet cython file into multiple files

  #51080 opened Mar 5, 2025

• [core] add tests to ensure the ConcurrencyGroupManager creates the correct number of threads

  #51075 opened Mar 4, 2025

• [core] Only one of the threads in a thread pool will be initialized as a long-running Python thread

  #51071 opened Mar 4, 2025

• [Ray Data] One Hot Encoding breaks when Pyarrow Blocks do not contain the entire value set for Pyarrow < 14.0

  #51065 opened Mar 4, 2025

85 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• [Compiled Graph] Enhance Compile Graph with Multi-Device Support and HCCL Integration

  #51032 commented on Mar 11, 2025 • 26 new comments

• [core] Utils to cleanup cgroup folder

  #49941 commented on Mar 7, 2025 • 25 new comments

• [RLlib] Balance package loads for `AggregatorActors`.

  #51017 commented on Mar 11, 2025 • 10 new comments

• [Autoscaler][V2] Check IM instance_status before terminating nodes

  #50707 commented on Mar 11, 2025 • 6 new comments

• [core][compiled graphs] Support reduce scatter and all gather collective for GPU communicator in compiled graph

  #50624 commented on Mar 11, 2025 • 6 new comments

• [train] Fold `v2.XGBoostTrainer` API into the public trainer class as an alternate constructor

  #50045 commented on Mar 11, 2025 • 3 new comments

• Integrate Ray Dataset with Daft Dataframe

  #50630 commented on Mar 10, 2025 • 3 new comments

• Fix ax_client.create_experiment call

  #45902 commented on Mar 7, 2025 • 2 new comments

• [wandb] Use wandb Run as a context manager

  #49307 commented on Mar 8, 2025 • 1 new comment

• [Autoscaler][Placement Group] Skip placed bundle when requesting resource

  #48924 commented on Mar 6, 2025 • 1 new comment

• [Docs][Core] Update system logs doc for dashboard subprocess module

  #50984 commented on Mar 6, 2025 • 1 new comment

• [core] Periodically check for unexpected worker socket disconnects

  #50812 commented on Mar 11, 2025 • 1 new comment

• [core] Populate obj store memory even if 0 + less copies in resource managers

  #50637 commented on Mar 4, 2025 • 1 new comment

• [Docs] Update Volcano Integration with The New Flag

  #47911 commented on Mar 4, 2025 • 0 new comments

• [doc] minor/patch version update

  #48626 commented on Mar 6, 2025 • 0 new comments

• (WIP) [core][compiled graphs] Unify code paths for NCCL P2P and collectives scheduling

  #48649 commented on Mar 5, 2025 • 0 new comments

• [core] cpp lint of object_manager

  #48878 commented on Mar 5, 2025 • 0 new comments

• [core] Lint cpp files in common

  #49002 commented on Mar 5, 2025 • 0 new comments

• [ADAG]Enable NPU (hccl) communication for CG

  #47658 commented on Mar 7, 2025 • 0 new comments

• [Doc][KubeRay] Add KubeRay image resize example to Ray doc page

  #46447 commented on Mar 7, 2025 • 0 new comments

• Ray IPv6 support

  #44252 commented on Mar 5, 2025 • 0 new comments

• [Data] Resolve block references asynchronously in separate thread

  #41465 commented on Mar 7, 2025 • 0 new comments

• [data] Make exceptions consistent when falling back to pandas

  #39969 commented on Mar 7, 2025 • 0 new comments

• [WIP] Add example how to computing embeddings with Ray Data

  #39879 commented on Mar 7, 2025 • 0 new comments

• [data] Fix map_batches on datasets with nested lists

  #39869 commented on Mar 7, 2025 • 0 new comments

• [Data] Consolidate default fault tolerance options

  #39797 commented on Mar 7, 2025 • 0 new comments

• Upgrade default AWS DLAMI

  #39721 commented on Mar 5, 2025 • 0 new comments

• [core] Gcs asio minor improvements

  #49169 commented on Mar 10, 2025 • 0 new comments

• [core] Don't get dashboard address after each dashboard connection failure

  #49584 commented on Mar 10, 2025 • 0 new comments

• adding distributional critic example

  #49949 commented on Mar 10, 2025 • 0 new comments

• [core] Introduce ConcurrentFlatMap and use for InMemoryStoreClient

  #50375 commented on Mar 10, 2025 • 0 new comments

• [data] add ClickHouse sink

  #50377 commented on Mar 4, 2025 • 0 new comments

• [Autoscaler][V2] Use running node instances to rate-limit upscaling

  #50414 commented on Mar 10, 2025 • 0 new comments

• [tune] Remove loguniform's base

  #50415 commented on Mar 8, 2025 • 0 new comments

• [RLlib] Enable spliting and zero padding of Dict observation

  #50589 commented on Mar 5, 2025 • 0 new comments

• [core] Mask data directory for core tests

  #50617 commented on Mar 8, 2025 • 0 new comments

• [chore] Delete unused build.sh

  #50649 commented on Mar 6, 2025 • 0 new comments

• [POC] Run arbitrary bash commands in doc as test

  #50988 commented on Mar 4, 2025 • 0 new comments

• Fix editorconfig option name

  #50993 commented on Mar 8, 2025 • 0 new comments

• Suppress type error

  #50994 commented on Mar 8, 2025 • 0 new comments

• Improvements to General Debugging guide

  #51004 commented on Mar 6, 2025 • 0 new comments

• [WIP][core][compiled graphs] Supporting allreduce on tuple of tensors

  #51047 commented on Mar 5, 2025 • 0 new comments

• [core] Always create a default executor

  #51058 commented on Mar 10, 2025 • 0 new comments

• [Feature] [Performance] [Docs] Disabling object spilling is not documented

  #21998 commented on Mar 9, 2025 • 0 new comments

• [Data] Heavy spilling with heterogenous compute

  #44710 commented on Mar 8, 2025 • 0 new comments

• [RLlib] make_multi_callbacks function cannot pass new api stack verification

  #48089 commented on Mar 7, 2025 • 0 new comments

• [Epic][CI] Migrate linter to Ruff

  #47991 commented on Mar 7, 2025 • 0 new comments

• No backend type associated with device type npu

  #50516 commented on Mar 7, 2025 • 0 new comments

• [Autoscaler] Autoscaler gets into infinite cycle of removing and adding nodes, never satisfies placement group

  #50783 commented on Mar 6, 2025 • 0 new comments

• [distributed debugger] exception in regular remote worker function leading to access violation when debugger connects

  #51010 commented on Mar 6, 2025 • 0 new comments

• [Core] cannot pass namespace package at runtime via py_modules

  #50161 commented on Mar 6, 2025 • 0 new comments

• [<Ray component: Data>] async `flat_map`

  #50329 commented on Mar 6, 2025 • 0 new comments

• [tune] ClientObjectRef is not found for client

  #46747 commented on Mar 5, 2025 • 0 new comments

• [Serve] FastAPI ingress does not work with composable routers

  #50373 commented on Mar 5, 2025 • 0 new comments

• [RLlib] Attribute error when trying to compute action after training Multi Agent PPO with New API Stack

  #44475 commented on Mar 5, 2025 • 0 new comments

• [serve] [llm] Known issues in Ray Serve LLM

  #50931 commented on Mar 5, 2025 • 0 new comments

• [Core][ADAG] Local mode for Ray ADAG

  #47725 commented on Mar 5, 2025 • 0 new comments

• [Serve] DeepSeek-R1 mode load stuck in H20

  #50975 commented on Mar 5, 2025 • 0 new comments

• bazel-lint all BUILD files

  #50875 commented on Mar 5, 2025 • 0 new comments

• [Core] [Slurm] Allow parallel startup of Ray workers on Slurm

  #25819 commented on Mar 5, 2025 • 0 new comments

• [core][compiled graphs] Support pinned memory for CPU <-> GPU transfers

  #48086 commented on Mar 5, 2025 • 0 new comments

• [Train] Unable to gain long-term access to S3 storage for training state/checkpoints when running on AWS EKS

  #50823 commented on Mar 5, 2025 • 0 new comments

• [core] Deserialize torch.Tensors to the correct device

  #50134 commented on Mar 4, 2025 • 0 new comments

• [Core] ray.init() hangs using Python 3.10.15 on Linux

  #48625 commented on Mar 4, 2025 • 0 new comments

• [data][tests] Add HuggingFace dataset to image comparison benchmark

  #39665 commented on Mar 7, 2025 • 0 new comments

• [Data] Add `read_images` benchmark for 100 million images

  #38862 commented on Mar 7, 2025 • 0 new comments

• [Data] Add `read_delta` API to read Delta format files

  #38813 commented on Mar 7, 2025 • 0 new comments

• CI test linux://python/ray/tests:test_runtime_env_complicated is flaky

  #49674 commented on Mar 11, 2025 • 0 new comments

• New RLlib API examples

  #50897 commented on Mar 11, 2025 • 0 new comments

• RLlib | Conncetor.py

  #51039 commented on Mar 11, 2025 • 0 new comments

• CI test windows://python/ray/tests:test_storage is consistently_failing

  #48922 commented on Mar 11, 2025 • 0 new comments

• CI test linux://rllib:learning_tests_cartpole_dqn_gpu is flaky

  #46683 commented on Mar 11, 2025 • 0 new comments

• Make sure precommit hook linter and CI matches

  #50694 commented on Mar 11, 2025 • 0 new comments

• [autoscaler] Better documentation on the node provider interface

  #26087 commented on Mar 11, 2025 • 0 new comments

• [Serve] exceptions raised by request timeout are inconsistent

  #50992 commented on Mar 11, 2025 • 0 new comments

• [CORE] importing ray closes logging handlers, breaking custom logging

  #48846 commented on Mar 10, 2025 • 0 new comments

• Release test random_shuffle.chaos failed

  #49395 commented on Mar 10, 2025 • 0 new comments

• Release test sort.chaos failed

  #49765 commented on Mar 10, 2025 • 0 new comments

• Release test sort.regular failed

  #50417 commented on Mar 10, 2025 • 0 new comments

• Release test random_shuffle.regular failed

  #49383 commented on Mar 10, 2025 • 0 new comments

• [core] Run more tests with ASAN in CI to avoid memory leak

  #51057 commented on Mar 10, 2025 • 0 new comments

• [Core] Spot preemption related retries do not count towards the max retries

  #50640 commented on Mar 10, 2025 • 0 new comments

• [Serve] Extend configuration of Serve autoscaler with custom metrics

  #31540 commented on Mar 10, 2025 • 0 new comments

• [core] Cover cpplint for all C++ folders

  #50583 commented on Mar 9, 2025 • 0 new comments

• [Feedback] Feedback for ray + uv

  #50961 commented on Mar 9, 2025 • 0 new comments