Spaces:
Running
Running
Update Dockerfile
Browse files- Dockerfile +3 -5
Dockerfile
CHANGED
|
@@ -8,7 +8,8 @@ RUN apt update && apt install wget -y && rm -rf /var/lib/apt/lists/*
|
|
| 8 |
RUN wget "https://huggingface.co/LiquidAI/LFM2.5-2.6B-GGUF/resolve/main/LFM2.5-2.6B-Q4_K_M.gguf" -O /model.gguf
|
| 9 |
|
| 10 |
|
| 11 |
-
# Clear the entrypoint and use shell form CMD
|
|
|
|
| 12 |
ENTRYPOINT []
|
| 13 |
CMD /app/llama-server \
|
| 14 |
-m /model.gguf \
|
|
@@ -17,8 +18,5 @@ CMD /app/llama-server \
|
|
| 17 |
-c 4096 \
|
| 18 |
--parallel 1 \
|
| 19 |
--batch-size 256 --ubatch-size 256 \
|
| 20 |
-
--flash-attn \
|
| 21 |
-
--cache-type-k q8_0 --cache-type-v q8_0 \
|
| 22 |
--mlock --no-mmap \
|
| 23 |
-
--no-warmup
|
| 24 |
-
--log-disable > /dev/null 2>&1
|
|
|
|
| 8 |
RUN wget "https://huggingface.co/LiquidAI/LFM2.5-2.6B-GGUF/resolve/main/LFM2.5-2.6B-Q4_K_M.gguf" -O /model.gguf
|
| 9 |
|
| 10 |
|
| 11 |
+
# Clear the entrypoint and use shell form CMD
|
| 12 |
+
# NOTE: logs are ENABLED on purpose — do not add --log-disable until it runs fine
|
| 13 |
ENTRYPOINT []
|
| 14 |
CMD /app/llama-server \
|
| 15 |
-m /model.gguf \
|
|
|
|
| 18 |
-c 4096 \
|
| 19 |
--parallel 1 \
|
| 20 |
--batch-size 256 --ubatch-size 256 \
|
|
|
|
|
|
|
| 21 |
--mlock --no-mmap \
|
| 22 |
+
--no-warmup
|
|
|