Spaces:
Runtime error
Runtime error
fice
Browse files- compose.yml +28 -0
- Dockerfile → fastapi/Dockerfile +1 -24
- app.py → fastapi/app.py +1 -0
- requirements.txt → fastapi/requirements.txt +0 -0
- ollama.service +0 -10
- ollama/Dockerfile +3 -0
- ollama/pull-llama3.sh +11 -0
compose.yml
ADDED
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
version: "3"
|
2 |
+
|
3 |
+
services:
|
4 |
+
web:
|
5 |
+
build: fastapi
|
6 |
+
ports:
|
7 |
+
-7860:8000
|
8 |
+
volumes:
|
9 |
+
-./fastapi:/app
|
10 |
+
networks:
|
11 |
+
- local-ai
|
12 |
+
ollama:
|
13 |
+
build: ollama
|
14 |
+
ports:
|
15 |
+
-11434:11434
|
16 |
+
volumes:
|
17 |
+
-tutorial-vol:/ollama
|
18 |
+
networks:
|
19 |
+
-local-ai
|
20 |
+
entrypoint: ["usr/bin/bash","/pull-llama3.sh"]
|
21 |
+
|
22 |
+
networks:
|
23 |
+
local-ai:
|
24 |
+
driver: bridge
|
25 |
+
|
26 |
+
volumes:
|
27 |
+
tutorial-vol:
|
28 |
+
driver: local
|
Dockerfile → fastapi/Dockerfile
RENAMED
@@ -1,35 +1,12 @@
|
|
1 |
# Utilise une image de base Python
|
2 |
FROM python:3.9-slim
|
3 |
|
4 |
-
# Installe les dépendances nécessaires
|
5 |
-
RUN apt-get update && \
|
6 |
-
apt-get install -y \
|
7 |
-
systemd \
|
8 |
-
sudo \
|
9 |
-
curl \
|
10 |
-
gnupg \
|
11 |
-
lsb-release
|
12 |
-
# Ajoute le fichier de service systemd
|
13 |
-
COPY ollama.service /etc/systemd/system/ollama.service
|
14 |
-
# Installe Ollama et configure les permissions
|
15 |
-
RUN curl -fsSL https://ollama.com/install.sh | sh
|
16 |
-
# Active le service et démarre systemd
|
17 |
-
RUN sudo systemctl enable ollama.service
|
18 |
-
RUN sudo systemctl daemon-reload
|
19 |
-
RUN sudo systemctl enable ollama
|
20 |
-
RUN sudo ollama run llama3
|
21 |
-
# Télécharge le modèle Llama3
|
22 |
-
# Ajoute le chemin d'Ollama au PATH
|
23 |
-
ENV PATH="/root/.ollama:${PATH}"
|
24 |
-
|
25 |
-
# Télécharge le modèle Llama3
|
26 |
-
RUN ollama pull llama3:latest
|
27 |
|
28 |
# Définit le répertoire de travail
|
29 |
WORKDIR /app
|
30 |
|
31 |
# Copie les fichiers de l'application dans le conteneur
|
32 |
-
COPY . /app/
|
33 |
|
34 |
# Installe les dépendances Python
|
35 |
RUN pip install --upgrade pip
|
|
|
1 |
# Utilise une image de base Python
|
2 |
FROM python:3.9-slim
|
3 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
4 |
|
5 |
# Définit le répertoire de travail
|
6 |
WORKDIR /app
|
7 |
|
8 |
# Copie les fichiers de l'application dans le conteneur
|
9 |
+
COPY .requirements.txt /app/requirements.txt
|
10 |
|
11 |
# Installe les dépendances Python
|
12 |
RUN pip install --upgrade pip
|
app.py → fastapi/app.py
RENAMED
@@ -46,6 +46,7 @@ async def generate_text(request: RequestModel):
|
|
46 |
model="llama3:latest",
|
47 |
messages=[{"role": "user", "content": prompt_first + request.text}],
|
48 |
)
|
|
|
49 |
generated_text = ""
|
50 |
for chunk in response:
|
51 |
if chunk.choices[0].delta.content is not None:
|
|
|
46 |
model="llama3:latest",
|
47 |
messages=[{"role": "user", "content": prompt_first + request.text}],
|
48 |
)
|
49 |
+
|
50 |
generated_text = ""
|
51 |
for chunk in response:
|
52 |
if chunk.choices[0].delta.content is not None:
|
requirements.txt → fastapi/requirements.txt
RENAMED
File without changes
|
ollama.service
DELETED
@@ -1,10 +0,0 @@
|
|
1 |
-
Description=Ollama Service
|
2 |
-
After=network-online.target
|
3 |
-
il me faut ajouter tout ca dans le dockerfile ya pas de commande qui permet de le gerer
|
4 |
-
[Service]
|
5 |
-
ExecStart=/usr/bin/ollama serve
|
6 |
-
User=ollama
|
7 |
-
Group=ollama
|
8 |
-
Restart=always
|
9 |
-
RestartSec=3
|
10 |
-
Environment="OLLAMA_MODELS=~/.ollama/models/"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ollama/Dockerfile
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
FROM ollama/ollama
|
2 |
+
|
3 |
+
COPY ./pull-llama3.sh /pull-llama3.sh
|
ollama/pull-llama3.sh
ADDED
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
./bin/ollama serve &
|
2 |
+
|
3 |
+
pif=$!
|
4 |
+
|
5 |
+
sleep 5
|
6 |
+
|
7 |
+
echo "Pulling llama3 model"
|
8 |
+
|
9 |
+
ollama pull llama3
|
10 |
+
|
11 |
+
wait $pid
|