Apppublic

programmersd/llama-gemma

sourceHugging Faceupdated 5mo agoView on Hugging Face

1likes

Dockerfile19 linesDownload Raw Back to root

1FROM ghcr.io/ggml-org/llama.cpp:server2 3ENV LLAMA_CACHE=/tmp/llama-cache4ENV HF_HUB_CACHE=/tmp/hf-cache5 6EXPOSE 78607 8CMD ["-hf", "unsloth/gemma-4-E2B-it-GGUF:Q4_0", \9"-c", "2048", \10"-t", "8", \11"-tb", "16", \12"-b", "1024", \13"-ub", "512", \14"-np", "1", \15"--flash-attn", "on", \16"--no-warmup", \17"--no-context-shift", \18"--host", "0.0.0.0", \19"--port", "7860"]