[pull] main from vllm-project:main #15
Closed
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
![pull[bot]](https://avatars.githubusercontent.com/in/12910?s=60&v=4){kind=small}
Co-authored-by: Lei Wen <[email protected]>
Signed-off-by: Prashant Gupta <[email protected]> Co-authored-by: Roger Wang <[email protected]>
Co-authored-by: Philipp Moritz <[email protected]> Co-authored-by: Woosuk Kwon <[email protected]> Co-authored-by: mgoin <[email protected]> Co-authored-by: Tyler Michael Smith <[email protected]> Co-authored-by: Cody Yu <[email protected]>
…int (#3467) Co-authored-by: Lily Liu <[email protected]> Co-authored-by: Cyrus Leung <[email protected]>
…#4494) Co-authored-by: Simon Mo <[email protected]>
… obtain the CUDA version. (#4173) Signed-off-by: AnyISalIn <[email protected]>
Signed-off-by: Travis Johnson <[email protected]>
Co-authored-by: Lei Wen <[email protected]>
Co-authored-by: Lei Wen <[email protected]> Co-authored-by: Sage Moore <[email protected]>
This PR updates the tuning script for the fused_moe kernel to support FP8 and also adds configurations for TP4. Note that for the configuration I removed num_warps and num_stages for small batch sizes since that improved performance and brought the benchmarks on par with the numbers before in that regime to make sure this is a strict improvement over the status quo. All the numbers below are for mistralai/Mixtral-8x7B-Instruct-v0.1, 1000 input and 50 output tokens. Before this PR (with static activation scaling): qps = 1: 9.8 ms ITL, 0.49s e2e latency qps = 2: 9.7 ms ITL, 0.49s e2e latency qps = 4: 10.1 ms ITL, 0.52s e2e latency qps = 6: 11.9 ms ITL, 0.59s e2e latency qps = 8: 14.0 ms ITL, 0.70s e2e latency qps = 10: 15.7 ms ITL, 0.79s e2e latency After this PR (with static activation scaling): qps = 1: 9.8 ms ITL, 0.49s e2e latency qps = 2: 9.7 ms ITL, 0.49s e2e latency qps = 4: 10.2 ms ITL, 0.53s e2e latency qps = 6: 11.9 ms ITL, 0.59s e2e latency qps = 8: 11.9 ms ITL, 0.59s e2e latency qps = 10: 12.1 ms ITL, 0.61s e2e latency
Remove the device="cuda" declarations in mixtral as promised in #4343
…n is not 1 and max_tokens is large & Add tests for preemption (#4451)
|
/ok-to-test |
|
PR needs rebase. Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. |
Co-authored-by: Cade Daniel <[email protected]>
|
@pull[bot]: The following tests failed, say
Full PR test history. Your PR dashboard. Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. I understand the commands that are listed here. |
|
PR needs rebase. Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository. |
Co-authored-by: Michael Goin <[email protected]>
Merged
|
Closed in favour of #18 |
Merged
dtrifiro
pushed a commit
that referenced
this pull request
Jul 26, 2024
This PR updates our grpc_server to add TGIS-style logs similar to https://github.com/IBM/text-generation-inference/blob/main/router/src/grpc_server.rs#L504-L512 This also disables the vllm per-request logging so that we don't double-log each request The timing info collected here is pretty rough, it doesn't plumb into the LLMEngine, it just times the generators to get the total time spent in the engine. We could do better, but this is a start. Example logs: ``` INFO 04-09 21:51:01 logs.py:43] generate_stream{input=[b'This is the story of Obama ridin...'] prefix_id= input_chars=[70] params=sampling { } stopping { max_new_tokens: 200 min_new_tokens: 16 } response { } decoding { } tokenization_time=0.45ms queue_and_inference_time=1096.67ms time_per_token=5.48ms total_time=1097.12ms input_toks=16}: Streaming response generated 200 tokens before NOT_FINISHED, output 848 chars: b' California. The story is told i...' INFO 04-09 21:51:08 logs.py:43] generate{input=[b'Lorem ipsum dolor sit amet, cons...', b'foooood man where is it'] prefix_id= input_chars=[469] params=sampling { } stopping { max_new_tokens: 20 min_new_tokens: 16 } response { } decoding { } tokenization_time=2.03ms queue_and_inference_time=122.23ms time_per_token=6.11ms total_time=124.26ms input_toks=124}: Sub-request 0 from batch of 2 generated 20 tokens before MAX_TOKENS, output 25 chars: b'?\\n\\n<!--\\n<!--\\n<!--\\n<!--\\n<!' INFO 04-09 21:51:08 logs.py:43] generate{input=[b'Lorem ipsum dolor sit amet, cons...', b'foooood man where is it'] prefix_id= input_chars=[469] params=sampling { } stopping { max_new_tokens: 20 min_new_tokens: 16 } response { } decoding { } tokenization_time=2.07ms queue_and_inference_time=122.22ms time_per_token=6.11ms total_time=124.29ms input_toks=7}: Sub-request 1 from batch of 2 generated 20 tokens before MAX_TOKENS, output 70 chars: b"?\\nI don't know.\\nI don't know.\\nI ..." ``` --------- Signed-off-by: Joe Runde <[email protected]> Signed-off-by: Joe Runde <[email protected]>
prarit
pushed a commit
to prarit/vllm
that referenced
this pull request
Oct 18, 2024
Correctly calculating the same value for the required cache blocks num for all torchrun processes
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
See Commits and Changes for more details.
Created by
pull[bot]
Can you help keep this open source service alive? 💖 Please sponsor : )