Update Dockerfile

Dockerfile

@@ -2,21 +2,25 @@ FROM ghcr.io/ggml-org/llama.cpp:full
 
 WORKDIR /app
 
-RUN apt update && apt install -y python3 python3-pip python3-venv
+RUN apt-get update && apt-get install -y --no-install-recommends python3 python3-pip python3-venv && \
+    rm -rf /var/lib/apt/lists/*
+
 RUN python3 -m venv /opt/venv
 ENV PATH="/opt/venv/bin:$PATH"
 
-RUN pip install -U pip huggingface_hub
+RUN pip install --no-cache-dir -U pip huggingface_hub
 
 RUN python3 -c 'from huggingface_hub import hf_hub_download; \
     repo="LiquidAI/LFM2.5-8B-A1B-GGUF"; \
-    hf_hub_download(repo_id=repo, filename="LFM2.5-8B-A1B-Q4_K_M.gguf", local_dir="/app")'
+    hf_hub_download(repo_id=repo, filename="LFM2.5-8B-A1B-Q4_0.gguf", local_dir="/app")'
 
 CMD ["--server", \
-     "-m", "/app/LFM2.5-8B-A1B-Q4_K_M.gguf", \
+     "-m", "/app/LFM2.5-8B-A1B-Q4_0.gguf", \
      "--host", "0.0.0.0", \
      "--port", "7860", \
      "-t", "2", \
+     "-b", "512", \
+     "-ub", "512", \
      "--cache-type-k", "q8_0", \
      "--cache-type-v", "iq4_nl", \
      "-c", "128000", \