Llava15ChatHandler - Cache Image Encoding

未關閉
#2,222 0 則留言 0 個 reaction 已指派 0 人 在 GitHub 檢視

維護者通常 1 天內回覆

還沒有人認領這個 Issue。

評估

難度
3/5
預估耗時
1-2 天
新手友好度
45/100
Issue 類型
功能
描述清晰度
基本清楚
活躍度
停滯
技術堆疊
python
領域
performance

研究方向

查看 Llava15ChatHandler 原始碼,了解目前每一輪如何執行影像編碼。目標是快取編碼後的結果。先檢查影像處理流程,並找出應在哪裡儲存及重複使用快取。檢查程式碼庫中是否已有快取機制。成功的實作將避免在多輪對話中重新編碼影像,從而提升推論速度,尤其是在 CPU 上。

由索引模型根據 Issue 內容生成。

描述

Is your feature request related to a problem? Please describe.
In multi-turn conversations that include images, the image encoding is performed on each turn. This slows down inference considerably, especially when running with pure CPU.

Describe the solution you'd like
Once an image has been encoded, the result can be reused, rather than re-encoding every turn.

Describe alternatives you've considered
Storing a detailed description of the image in the conversation history, and popping out the image_url part. This is not as flexible however, if wanting to query about something specfic later on that wasn't captured in the saved description.

Additional context
I have implemented a solution locally editing the source code (with some heavy AI assistance, as I am not an expert on low level LLM coding) and seem to have this feature working, by storing chunk signatures instead of immediately processing and manipulating the kv_cache_seq_rm, but not really sure if its a safe/sustainable approach.

I could share the source code or raise a draft PR if it'd be useful.

Thanks!

主要語言
Python
星號
10.7k
分支
1.5k
平均合併
1 小時 40 分鐘
30 天內合併 PR
5

環境準備

  • 沒有 Dockerfile 或 Docker Compose 檔案
  • 沒有 Pull Request 範本
  • 閱讀貢獻指南

從這裡開始

  1. 先讀完整個 Issue,再讀專案的貢獻指南。
  2. 在 Issue 下留言說明你要接手 —— 這能避免兩個人做同樣的事。
  3. Fork 儲存庫,在一個分支上完成修改。
  4. 送出 Pull Request,並在描述裡引用這個 Issue 編號。

abetlen/llama-cpp-python 的其他 Issue

查看 abetlen/llama-cpp-python 的全部 Issue

相似的 Issue

更多 Python Issue

把新 issue 寄到你的電子郵件信箱

精選適合新手參與的 GitHub issue 摘要。