15308555518/ MaxKB
1
0
Fork 0
mirror of https://github.com/1Panel-dev/MaxKB.git synced 2025-12-26 01:33:05 +00:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity

build: add new tokenizer.

This commit is contained in:
liqiang-fit2cloud 2025-07-08 19:22:16 +08:00
parent 573ec2b706
commit a98cde9912
2 changed files with 14 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
.github/workflows/build-and-push-vector-model.yml vendored
View File

@ -5,7 +5,7 @@ on:
inputs:
dockerImageTag:
description: 'Docker Image Tag'
default: 'v2.0.1'
default: 'v2.0.2'
required: true
architecture:
description: 'Architecture'

15
installer/Dockerfile-vector-model
View File

@ -10,12 +10,23 @@
# 不知道为什么用上面的脚本重新拉一遍向量模型比之前的大很多，所以还是用下面的脚本复用原来已经构建好的向量模型
FROM python:3.11-slim-bookworm AS tmp-stage1
COPY installer/install_model_token.py install_model_token.py
RUN pip3 install --upgrade pip setuptools && \
pip install pycrawlers && \
pip install transformers && \
python3 install_model_token.py && \
cp -r model/base/hub model/tokenizer
FROM ghcr.io/1panel-dev/maxkb-vector-model:v1.0.1 AS vector-model
FROM alpine AS tmp-stage
FROM alpine AS tmp-stage2
COPY --from=vector-model /opt/maxkb/app/model /opt/maxkb-app/model
COPY --from=vector-model /opt/maxkb/app/model/base/hub /opt/maxkb-app/model/tokenizer
COPY --from=tmp-stage1 model/tokenizer /opt/maxkb-app/model/tokenizer
RUN rm -rf /opt/maxkb-app/model/embedding/shibing624_text2vec-base-chinese/onnx
FROM scratch
COPY --from=tmp-stage /opt/maxkb-app/model /opt/maxkb-app/model
COPY --from=tmp-stage2 /opt/maxkb-app/model /opt/maxkb-app/model