Spaces:
Paused
Paused
Create Dockerfile
Browse files- Dockerfile +29 -0
Dockerfile
ADDED
|
@@ -0,0 +1,29 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
FROM ghcr.io/ggml-org/llama.cpp:server
|
| 2 |
+
|
| 3 |
+
EXPOSE 7860
|
| 4 |
+
|
| 5 |
+
ENTRYPOINT ["/app/llama-server"]
|
| 6 |
+
|
| 7 |
+
CMD [
|
| 8 |
+
"--host", "0.0.0.0",
|
| 9 |
+
"--port", "7860",
|
| 10 |
+
|
| 11 |
+
"--hf-repo", "XHToken/Spark-X2.5-1.7B-GGUF:Q8_0",
|
| 12 |
+
"--alias", "spark-x2.5-1.7b",
|
| 13 |
+
|
| 14 |
+
"--ctx-size", "32768",
|
| 15 |
+
"--parallel", "1",
|
| 16 |
+
|
| 17 |
+
"--threads", "2",
|
| 18 |
+
"--threads-batch", "2",
|
| 19 |
+
"--batch-size", "512",
|
| 20 |
+
"--ubatch-size", "256",
|
| 21 |
+
|
| 22 |
+
"--cache-type-k", "q8_0",
|
| 23 |
+
"--cache-type-v", "q8_0",
|
| 24 |
+
|
| 25 |
+
"--flash-attn", "auto",
|
| 26 |
+
"--cont-batching",
|
| 27 |
+
"--metrics",
|
| 28 |
+
"--no-webui"
|
| 29 |
+
]
|