⚡

Fastcheck

Public

Fast CI for PRs

p88h:serialize-tensors

2m
Apply suggestions from code review#21154
Staszek Pasko·p88h:serialize-tensorsGitHub Icon d5903c621·Created
—
Serialize tensors using int8 views#21153
Staszek Paśko·p88h:serialize-tensorsGitHub Icon 103974939·Created
21m
Update vllm/v1/serial_utils.py#21150
Staszek Pasko·p88h:serialize-tensorsGitHub Icon e45c80a4b·Created
2m
Update vllm/v1/serial_utils.py#21149
Staszek Paśko·p88h:serialize-tensorsGitHub Icon a57024220·Created
1h
style#21142
Staszek Pasko·p88h:serialize-tensorsGitHub Icon cd4a2d455·Created

reidliu41:refactor-examples-series-08

15m
[Misc] refactor examples series - Chat Completion Client With Tools#21152
Reid·reidliu41:refactor-examples-series-08GitHub Icon 88d0ac320·Created
43m
[Misc] refactor examples series - OpenAI Chat Completion Client With Tools#21062
Reid·reidliu41:refactor-examples-series-08GitHub Icon c6d0c46b7·Created

qthequartermasterman:feature/vllm/add-input-embedding

15m
preprocess tensors to handle batched/misshaped prompt embeds to avoid handling shape mismatches in the engine and model runner#21151
Andrew Sansom·qthequartermasterman:feature/vllm/add-input-embeddingGitHub Icon 9a57acaf5·Created
2h
Merge branch 'vllm-project:main' into feature/vllm/add-input-embedding#20773
Andrew Sansom·qthequartermasterman:feature/vllm/add-input-embeddingGitHub Icon 934ceae63·Created
10s
[Core] [Bugfix] Add Input Embeddings#20772
Andrew Sansom·qthequartermasterman:feature/vllm/add-input-embeddingGitHub Icon 70198f625·Created
2h
add scheduler test to ensure prompt embeds and prompt tokens are not co-mingled#20690
Andrew Sansom·qthequartermasterman:feature/vllm/add-input-embeddingGitHub Icon d949f1b0b·Created
2h
refactor cuda_prepare_decode test#20587
Andrew Sansom·qthequartermasterman:feature/vllm/add-input-embeddingGitHub Icon 193ad5cba·Created

cascade812:sp_pass

41m
Merge branch 'main' into sp_pass#21148
Tyler Michael Smith·cascade812:sp_passGitHub Icon f60a87120·Created
3h
pass in dtype and device#20855
cascade·cascade812:sp_passGitHub Icon 662e69882·Created
2h
update#20834
cascade·cascade812:sp_passGitHub Icon 469511075·Created
2h
revert some false checkin#20820
cascade·cascade812:sp_passGitHub Icon 4d1bac7ee·Created
21m
address comments#20817
cascade·cascade812:sp_passGitHub Icon 8460be827·Created

pd_scheduling

1h
Initial stubs for P/D scheduling changes#21147
Tyler Michael Smith·pd_schedulingGitHub Icon 161010c38·Created
59s
Initial stubs for P/D scheduling changes#21146
Tyler Michael Smith·pd_schedulingGitHub Icon beb05e859·Created
16m
whitespace#21143
Tyler Michael Smith·pd_schedulingGitHub Icon b589f50c2·Created
23m
Initial stubs for P/D scheduling changes#21137
Tyler Michael Smith·pd_schedulingGitHub Icon c98c76131·Created
16m
add comment#21132
Tyler Michael Smith·pd_schedulingGitHub Icon 1fc75bcc2·Created

huydhn:torch-2.7.0-final-rc

1h
More tweak#21145
Huy Do·huydhn:torch-2.7.0-final-rcGitHub Icon 174c2a15b·Created
17m
Use --index-strategy unsafe-best-match temporarily#21141
Huy Do·huydhn:torch-2.7.0-final-rcGitHub Icon 363303b07·Created
20m
Wrong installation order#21136
Huy Do·huydhn:torch-2.7.0-final-rcGitHub Icon c33698ec5·Created
12m
Install xformers from source#21133
Huy Do·huydhn:torch-2.7.0-final-rcGitHub Icon 52a7c4beb·Created
44m
Merge branch 'main' into torch-2.7.0-final-rc#21110
Huy Do·huydhn:torch-2.7.0-final-rcGitHub Icon a131606d4·Created

jeffrey-dot-li:upstream-main

1h
review: add tests, refactor#21144
Jeffrey Li·jeffrey-dot-li:upstream-mainGitHub Icon 74306de41·Created
2h
fix: fix issue with n>1 sampling on v1 requests overriding each other#21124
Jeffrey Li·jeffrey-dot-li:upstream-mainGitHub Icon 640db5107·Created
17s
fix: fix issue with n>1 sampling on v1 requests overriding each other#21123
Jeffrey Li·jeffrey-dot-li:upstream-mainGitHub Icon c9e7c0cb5·Created
3m
fix: fix issue with n>1 sampling on v1 requests overriding each other#21122
Jeffrey Li·jeffrey-dot-li:upstream-mainGitHub Icon 6d8180685·Created

neuralmagic:bdellabe/hotfix-lmeval-vllm-v1

1h
[New Model]: Snowflake Arctic Embed (Family) (#16649)#21138
Brian Dellabetta·neuralmagic:bdellabe/hotfix-lmeval-vllm-v1GitHub Icon 3d3ab3689·Created
3h
exclude erroring out when len(input_ids) == max_prompt_len#20215
Brian Dellabetta·neuralmagic:bdellabe/hotfix-lmeval-vllm-v1GitHub Icon be0915ec3·Created
2h
resolve bug preventing lm_eval working with vllm v1 engine#19369
Brian Dellabetta·neuralmagic:bdellabe/hotfix-lmeval-vllm-v1GitHub Icon 8cb394791·Created

neuralmagic:cutlass-moe-bf16-weights

52m
update benchmark_cutlass_moe#21134
Varun Sundar Rabindranath·neuralmagic:cutlass-moe-bf16-weightsGitHub Icon 2256ab492·Created
2h
Merge branch 'main' into cutlass-moe-bf16-weights#21087
Varun Sundar Rabindranath·neuralmagic:cutlass-moe-bf16-weightsGitHub Icon 5263d8d56·Created
2h
Add more moe benchmark shapes#20925
Varun Sundar Rabindranath·neuralmagic:cutlass-moe-bf16-weightsGitHub Icon e47180623·Created
2h
Comment out output prints in tests#20769
ElizaWszola·neuralmagic:cutlass-moe-bf16-weightsGitHub Icon 99edc71b8·Created
55s
add expert parallel to torch hash#20768
ElizaWszola·neuralmagic:cutlass-moe-bf16-weightsGitHub Icon 5b9ab4f72·Created

kartikx:dev/kartikx/data-parallel-error

2h
fix formatting#21129
Kartik Ramesh·kartikx:dev/kartikx/data-parallel-errorGitHub Icon 694e48a92·Created
1h
fix formatting#21111
Kartik Ramesh·kartikx:dev/kartikx/data-parallel-errorGitHub Icon e9240ba91·Created
2h
fix formatting#21027
Kartik Ramesh·kartikx:dev/kartikx/data-parallel-errorGitHub Icon 96c8576c2·Created
25m
Update benchmarks/benchmark_throughput.py#21023
Kartik Ramesh·kartikx:dev/kartikx/data-parallel-errorGitHub Icon 78ebf57f2·Created
2h
Raise error for data-parallel with benchmark_throughput#20781
Kartik Ramesh·kartikx:dev/kartikx/data-parallel-errorGitHub Icon 869c37d5b·Created