ShadowHunter222 commited on
Commit
93ee293
·
verified ·
1 Parent(s): 8de1c3c

Update Dockerfile

Browse files
Files changed (1) hide show
  1. Dockerfile +3 -3
Dockerfile CHANGED
@@ -4,8 +4,8 @@ FROM ghcr.io/ggml-org/llama.cpp:server
4
  RUN apt update && apt install wget -y && rm -rf /var/lib/apt/lists/*
5
 
6
 
7
- # 1. THE MAIN MODEL: LFM2.5-8B-A1B in Q4_K_M — MoE with only ~1.5B active params per token
8
- RUN wget "https://huggingface.co/LiquidAI/LFM2.5-8B-A1B-GGUF/resolve/main/LFM2.5-8B-A1B-Q4_K_M.gguf" -O /model.gguf
9
 
10
 
11
  # Clear the entrypoint and use shell form CMD
@@ -19,4 +19,4 @@ CMD /app/llama-server \
19
  --parallel 1 \
20
  --batch-size 256 --ubatch-size 256 \
21
  --mlock --no-mmap \
22
- --no-warmup
 
4
  RUN apt update && apt install wget -y && rm -rf /var/lib/apt/lists/*
5
 
6
 
7
+ # 1. THE MAIN MODEL: Granite-4.0-H-Tiny Q4_K_M — hybrid Mamba/Transformer MoE, 7B total / ~1B active
8
+ RUN wget "https://huggingface.co/ibm-granite/granite-4.0-h-tiny-GGUF/resolve/main/granite-4.0-h-tiny-Q4_K_M.gguf" -O /model.gguf
9
 
10
 
11
  # Clear the entrypoint and use shell form CMD
 
19
  --parallel 1 \
20
  --batch-size 256 --ubatch-size 256 \
21
  --mlock --no-mmap \
22
+ --no-warmup