Commit
·
24caaad
1
Parent(s):
646c37a
feat: added clip model and chat_format
Browse files- Dockerfile +2 -1
- main.py +3 -1
Dockerfile
CHANGED
|
@@ -15,7 +15,8 @@ RUN pip install -U pip setuptools wheel && \
|
|
| 15 |
|
| 16 |
# Download model
|
| 17 |
RUN mkdir model && \
|
| 18 |
-
curl -L https://huggingface.co/mys/ggml_llava-v1.5-7b/resolve/main/ggml-model-q5_k.gguf -o model/gguf-model.bin
|
|
|
|
| 19 |
|
| 20 |
COPY ./start_server.sh ./
|
| 21 |
COPY ./main.py ./
|
|
|
|
| 15 |
|
| 16 |
# Download model
|
| 17 |
RUN mkdir model && \
|
| 18 |
+
curl -L https://huggingface.co/mys/ggml_llava-v1.5-7b/resolve/main/ggml-model-q5_k.gguf -o model/gguf-model.bin && \
|
| 19 |
+
curl -L https://huggingface.co/mys/ggml_llava-v1.5-7b/resolve/main/mmproj-model-f16.gguf -o model/clip-model.bin
|
| 20 |
|
| 21 |
COPY ./start_server.sh ./
|
| 22 |
COPY ./main.py ./
|
main.py
CHANGED
|
@@ -6,7 +6,9 @@ app = create_app(
|
|
| 6 |
Settings(
|
| 7 |
n_threads=2, # set to number of cpu cores
|
| 8 |
model="model/gguf-model.bin",
|
| 9 |
-
|
|
|
|
|
|
|
| 10 |
)
|
| 11 |
)
|
| 12 |
|
|
|
|
| 6 |
Settings(
|
| 7 |
n_threads=2, # set to number of cpu cores
|
| 8 |
model="model/gguf-model.bin",
|
| 9 |
+
clip_model_path="model/clip-model.bin",
|
| 10 |
+
embedding=True,
|
| 11 |
+
chat_format="llava-1-5"
|
| 12 |
)
|
| 13 |
)
|
| 14 |
|