matteo.mazzola
commited on
Commit
·
7fd5900
1
Parent(s):
dc53903
feat: first commit2
Browse files- Dockerfile +22 -0
Dockerfile
ADDED
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# Base image: Hugging Face TGI
|
2 |
+
FROM ghcr.io/huggingface/text-generation-inference:3.0.2
|
3 |
+
|
4 |
+
# Set working directory
|
5 |
+
WORKDIR /app
|
6 |
+
|
7 |
+
# Minimal storage directories
|
8 |
+
RUN mkdir -p /data && chmod 777 /data
|
9 |
+
RUN mkdir -p /.cache && chmod 777 /.cache
|
10 |
+
RUN mkdir -p /.triton && chmod 777 /.triton
|
11 |
+
|
12 |
+
# Expose the API port
|
13 |
+
EXPOSE 8080
|
14 |
+
|
15 |
+
# Optional: Set Hugging Face token for private model access
|
16 |
+
ARG HF_TOKEN
|
17 |
+
ENV HF_TOKEN=${HF_TOKEN}
|
18 |
+
|
19 |
+
# Model: Replace with any small or quantized TGI-compatible model
|
20 |
+
# Example: TheBloke/Mistral-7B-Instruct-v0.1-GPTQ or TinyLlama/TinyLlama-1.1B-Chat
|
21 |
+
CMD ["--model-id", "TinyLlama/TinyLlama-1.1B-Chat-v1.0", "--port", "8080"]
|
22 |
+
|