change llm model to qwen2 math

Dockerfile

@@ -97,8 +97,11 @@ EXPOSE 7860
 # CMD ["llama-server", "-hf", "TheBloke/deepseek-llm-7B-chat-GGUF:deepseek-llm-7b-chat.Q4_K_M.gguf",
 #      "--host", "0.0.0.0", "--port", "7860", "-c", "4096", "-t", "4", "-ngl", "0", "--cont-batching", "-b", "512"]
 #
-# Example Mixtral-8x7B-Instruct (very heavy for 10GB CPU spaces):
-CMD ["llama-server", "-hf", "TheBloke/Mixtral-8x7B-Instruct-v0.1-GGUF:mixtral-8x7b-instruct-v0.1.Q4_K_M.gguf", "--host", "0.0.0.0", "--port", "7860", "-c", "4096", "-t", "4", "-ngl", "0", "--cont-batching", "-b", "64", "--ubatch-size", "32"]
+# Example Mixtral-8x7B-Instruct (known loader incompatibilities on newer llama.cpp + too large for 10GB RAM):
+# CMD ["llama-server", "-hf", "TheBloke/Mixtral-8x7B-Instruct-v0.1-GGUF:mixtral-8x7b-instruct-v0.1.Q4_K_M.gguf", "--host", "0.0.0.0", "--port", "7860", "-c", "4096", "-t", "4", "-ngl", "0", "--cont-batching", "-b", "64", "--ubatch-size", "32"]
+#
+# Active: Qwen2-Math-7B Q4_K_M
+CMD ["llama-server", "-hf", "QuantFactory/Qwen2-Math-7B-GGUF:q4_k_m", "--host", "0.0.0.0", "--port", "7860", "-c", "8192", "-t", "4", "-ngl", "0", "--cont-batching", "-b", "128", "--ubatch-size", "64"]
 #
 # Example Qwen2.5-Coder 7B Instruct (32k context):
 # CMD ["llama-server", "-hf", "Qwen/Qwen2.5-Coder-7B-Instruct-GGUF:Qwen2.5-Coder-7B-Instruct.Q4_K_M.gguf",