forked from AI-Hypercomputer/inference-benchmark
-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathDockerfile
More file actions
21 lines (15 loc) · 688 Bytes
/
Dockerfile
File metadata and controls
21 lines (15 loc) · 688 Bytes
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
FROM python:3.9.20-slim-bookworm as dev
RUN apt-get update -y \
&& apt-get install -y python3-pip git vim curl wget
RUN pip3 install --upgrade pip
WORKDIR /workspace
# install build and runtime dependencies
COPY requirements.txt requirements.txt
RUN pip install -r requirements.txt
RUN pip install -U "huggingface_hub[cli]"
RUN wget https://huggingface.co/datasets/anon8231489123/ShareGPT_Vicuna_unfiltered/resolve/main/ShareGPT_V3_unfiltered_cleaned_split.json
COPY benchmark_serving.py benchmark_serving.py
COPY latency_throughput_curve.sh latency_throughput_curve.sh
COPY datasets/*.json /workspace/
RUN chmod +x latency_throughput_curve.sh
RUN chmod +x benchmark_serving.py