On demand #1
Open
On demand #1
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
This commit enables MoE on demand. On our GPUs, this cuts the memory requirements by about half. Inference is still much slower. You can instantiate MoE on demand on this branch using Mixtral with command: ''' LLM(model="/home/user/Workspace_4tb/grant/Mixtral-8x7B-v0.1", max_model_len=2048, gpu_memory_utilization=0.98, enforce_eager=True, num_gpu_blocks_override=1000) ''' Fixes MoE On Demand
This commit enables MoE on demand for DeepSeek V2 You can instantiate MoE on demand on this branch using DeepSeek Lite with command: ''' LLM(model="deepseek-ai/DeepSeek-V2-Lite", max_model_len=2048, gpu_memory_utilization=0.98, enforce_eager=True, num_gpu_blocks_override=1000) ''' Fixes MoE On Demand
|
👋 Hi! Thank you for contributing to the vLLM project. Once the PR is approved and ready to go, your PR reviewer(s) can run CI to test the changes comprehensively before merging. To run CI, PR reviewers can do one of these:
🚀 |
Prefill stage is not yet complete
gnpinkert
force-pushed
the
on-demand
branch
from
75fe2d1
to
07787f3
Compare
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Outdated suggestions cannot be applied.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.
Commit 75fe2d1 is the commit that uses our custom DRL model to predict experts for mixtral. ppo_moe must be installed