mirror of
https://github.com/hanzoai/visor.git
synced 2026-09-24 17:04:14 +00:00
Ayush RanjanandgVisor bot
16d95b1544
This image is being used for some GPU tests. Also clean up /bin/nvidia/bin from PATH from all GPU images. It's not used. PiperOrigin-RevId: 720649135
32 lines
840 B
Docker
32 lines
840 B
Docker
# https://hub.docker.com/r/ollama/ollama
|
|
FROM ollama/ollama:0.5.1
|
|
|
|
ENV PATH=$PATH:/usr/local/nvidia/bin
|
|
ENV OLLAMA_ORIGINS=*
|
|
ENV OLLAMA_HOST=0.0.0.0:11434
|
|
|
|
COPY pull.sh /tmp
|
|
|
|
# Pre-install models useful for benchmarking.
|
|
# These are huge (total ~96 GiB), but necessary to benchmark
|
|
# models of various sizes. They are in their own image file to
|
|
# keep the test-only image lighter by comparison.
|
|
|
|
# Useful as embedding model.
|
|
RUN /tmp/pull.sh snowflake-arctic-embed2:568m-l-fp16
|
|
|
|
# Useful as small model.
|
|
RUN /tmp/pull.sh gemma2:2b-instruct-fp16
|
|
|
|
# Useful as mid-size model.
|
|
RUN /tmp/pull.sh sailor2:8b-chat-fp16
|
|
|
|
# Useful as coding-specific model.
|
|
RUN /tmp/pull.sh qwen2.5-coder:7b-instruct-q8_0
|
|
|
|
# Useful as large model.
|
|
RUN /tmp/pull.sh llama2:70b-chat-q4_K_S
|
|
|
|
# Useful as vision model.
|
|
RUN /tmp/pull.sh llama3.2-vision:11b-instruct-fp16
|