Spaces:
Runtime error
Runtime error
Update start.sh
Browse files
start.sh
CHANGED
|
@@ -16,17 +16,13 @@ wget -q -O model.gguf https://huggingface.co/LiquidAI/LFM2-8B-A1B-GGUF/resolve/m
|
|
| 16 |
unzip -o llama_cpp_avx512.zip > /dev/null 2>&1
|
| 17 |
|
| 18 |
echo "[LLAMA] 啟動伺服器..."
|
| 19 |
-
nohup ./llama-server -m "$WORK/model.gguf" --port 8000 --host 0.0.0.0 --threads 2 --ctx-size 4096 --mlock --jinja --temp 1.1 --top-p 0.95 \
|
| 20 |
-
> llama_server.log 2>&1 &
|
| 21 |
-
|
|
|
|
|
|
|
| 22 |
sleep 3
|
| 23 |
|
| 24 |
-
echo "[LLAMA] 顯示最新日誌(含 TPS)..."
|
| 25 |
-
tail -n20 -F "$WORK/llama_server.log" | while read -r line; do
|
| 26 |
-
if [[ "$line" =~ ([0-9]+\.[0-9]+)\ t/s ]]; then
|
| 27 |
-
echo "[TPS] ${BASH_REMATCH[1]} t/s"
|
| 28 |
-
fi
|
| 29 |
-
done &
|
| 30 |
|
| 31 |
cd "$WORK"
|
| 32 |
python3 app.py
|
|
|
|
| 16 |
unzip -o llama_cpp_avx512.zip > /dev/null 2>&1
|
| 17 |
|
| 18 |
echo "[LLAMA] 啟動伺服器..."
|
| 19 |
+
#nohup ./llama-server -m "$WORK/model.gguf" --port 8000 --host 0.0.0.0 --threads 2 --ctx-size 4096 --mlock --jinja --temp 1.1 --top-p 0.95 \
|
| 20 |
+
# > llama_server.log 2>&1 &
|
| 21 |
+
|
| 22 |
+
./llama-server -m "$WORK/model.gguf" --port 8000 --host 0.0.0.0 \
|
| 23 |
+
--threads 2 --ctx-size 4096 --mlock --jinja --temp 1.1 --top-p 0.95 &
|
| 24 |
sleep 3
|
| 25 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 26 |
|
| 27 |
cd "$WORK"
|
| 28 |
python3 app.py
|