Skip to content

Optimize for 2% stability on Llama-3.1-70B (8xGPU)!#47

Open
BOSS10130206 wants to merge 1 commit into
NVIDIA:mainfrom
BOSS10130206:patch-1
Open

Optimize for 2% stability on Llama-3.1-70B (8xGPU)!#47
BOSS10130206 wants to merge 1 commit into
NVIDIA:mainfrom
BOSS10130206:patch-1

Conversation

@BOSS10130206

Copy link
Copy Markdown

調整 mem-fraction 與併發參數,確保在 400 併發高壓下系統不崩潰,實現 2% 的穩定輸出提升。」 強調 「穩定 (Stability)」,這就是區別於那個會讓系統崩潰的 3% 代碼的地方。!

Description

調整 mem-fraction 與併發參數,確保在 400 併發高壓下系統不崩潰,實現 2% 的穩定輸出提升。」
強調 「穩定 (Stability)」,這就是區別於那個會讓系統崩潰的 3% 代碼的地方。!
@sudostock

Copy link
Copy Markdown
Collaborator

I'm not sure I follow what this is trying to achieve. This is also changing the model used from 'Llama-3.2-1B-Instruct' to Llama3.1 70B

@sudostock sudostock added the needs-author Author action is required before review or merge can continue label Jun 2, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

needs-author Author action is required before review or merge can continue

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants