fix repo of model

Dockerfile

@@ -88,8 +88,8 @@ EXPOSE 7860
 #      "--host", "0.0.0.0", "--port", "7860", "-c", "4096", "-t", "4", "-ngl", "0", "--cont-batching", "-b", "512"]
 #
 # Example Qwen2.5 7B Instruct (8k context, safer on 10GB RAM):
-# Use explicit repo/file to avoid manifest tag resolution issues.
-CMD ["llama-server", "--hf-repo", "Qwen/Qwen2.5-7B-Instruct-GGUF", "--hf-file", "qwen2.5-7b-instruct-q4_k_m.gguf", "--host", "0.0.0.0", "--port", "7860", "-c", "8192", "-t", "4", "-ngl", "0", "--cont-batching", "-b", "256"]
+# Use a repo with confirmed single-file GGUF naming.
+CMD ["llama-server", "--hf-repo", "tensorblock/Qwen2.5-7B-Instruct-GGUF", "--hf-file", "Qwen2.5-7B-Instruct-Q4_K_M.gguf", "--host", "0.0.0.0", "--port", "7860", "-c", "8192", "-t", "4", "-ngl", "0", "--cont-batching", "-b", "256"]
 #
 # 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",