Spaces:
Running
Running
Dmitry Beresnev commited on
Commit ·
34775a7
1
Parent(s): dc883f9
reduce context and batch
Browse files- Dockerfile +1 -1
Dockerfile
CHANGED
|
@@ -89,7 +89,7 @@ EXPOSE 7860
|
|
| 89 |
#
|
| 90 |
# Example Qwen2.5 7B Instruct (8k context, safer on 10GB RAM):
|
| 91 |
# Verified llama.cpp args from model card.
|
| 92 |
-
CMD ["llama-server", "--hf-repo", "paultimothymooney/Qwen2.5-7B-Instruct-Q4_K_M-GGUF", "--hf-file", "qwen2.5-7b-instruct-q4_k_m.gguf", "--host", "0.0.0.0", "--port", "7860", "-c", "
|
| 93 |
#
|
| 94 |
# Example Qwen2.5-Coder 7B Instruct (32k context):
|
| 95 |
# CMD ["llama-server", "-hf", "Qwen/Qwen2.5-Coder-7B-Instruct-GGUF:Qwen2.5-Coder-7B-Instruct.Q4_K_M.gguf",
|
|
|
|
| 89 |
#
|
| 90 |
# Example Qwen2.5 7B Instruct (8k context, safer on 10GB RAM):
|
| 91 |
# Verified llama.cpp args from model card.
|
| 92 |
+
CMD ["llama-server", "--hf-repo", "paultimothymooney/Qwen2.5-7B-Instruct-Q4_K_M-GGUF", "--hf-file", "qwen2.5-7b-instruct-q4_k_m.gguf", "--host", "0.0.0.0", "--port", "7860", "-c", "4096", "-t", "4", "-ngl", "0", "--cont-batching", "-b", "128", "--ubatch-size", "64"]
|
| 93 |
#
|
| 94 |
# Example Qwen2.5-Coder 7B Instruct (32k context):
|
| 95 |
# CMD ["llama-server", "-hf", "Qwen/Qwen2.5-Coder-7B-Instruct-GGUF:Qwen2.5-Coder-7B-Instruct.Q4_K_M.gguf",
|