Spaces:
Running
on
T4
Running
on
T4
Support embedding
Browse files- Dockerfile +3 -1
Dockerfile
CHANGED
@@ -5,6 +5,7 @@ ARG WHISPERMODEL
|
|
5 |
ARG MMPROJ
|
6 |
ARG TTSMODEL
|
7 |
ARG WAVTOKMODEL
|
|
|
8 |
ARG MODEL_NAME
|
9 |
ARG ADDITIONAL
|
10 |
RUN mkdir /opt/koboldcpp
|
@@ -19,4 +20,5 @@ RUN curl -fLo mmproj.gguf $MMPROJ || true
|
|
19 |
RUN curl -fLo tts.gguf $TTSMODEL || true
|
20 |
RUN curl -fLo wavtok.gguf $WAVTOKMODEL || true
|
21 |
RUN curl -fLo whispermodel.gguf $WHISPERMODEL || true
|
22 |
-
|
|
|
|
5 |
ARG MMPROJ
|
6 |
ARG TTSMODEL
|
7 |
ARG WAVTOKMODEL
|
8 |
+
ARG EMBEDMODEL
|
9 |
ARG MODEL_NAME
|
10 |
ARG ADDITIONAL
|
11 |
RUN mkdir /opt/koboldcpp
|
|
|
20 |
RUN curl -fLo tts.gguf $TTSMODEL || true
|
21 |
RUN curl -fLo wavtok.gguf $WAVTOKMODEL || true
|
22 |
RUN curl -fLo whispermodel.gguf $WHISPERMODEL || true
|
23 |
+
RUN curl -fLo embedmodel.gguf $EMBEDMODEL || true
|
24 |
+
CMD ./koboldcpp --model model.gguf --whispermodel whispermodel.gguf --sdmodel imgmodel.gguf --sdthreads 4 --sdquant --sdclamped --mmproj mmproj.gguf --ttsmodel tts.gguf --ttswavtokenizer wavtok.gguf --embeddingsmodel embedmodel.gguf $ADDITIONAL --port 7860 --hordemodelname $MODEL_NAME --hordemaxctx 1 --hordegenlen 1 --quiet --preloadstory default.json --chatcompletionsadapter adapter.json --ignoremissing $SECRET
|