yash184 commited on
Commit
9c1debc
·
verified ·
1 Parent(s): 97de08c

Update Dockerfile

Browse files
Files changed (1) hide show
  1. Dockerfile +12 -6
Dockerfile CHANGED
@@ -1,25 +1,31 @@
 
1
  FROM ghcr.io/ggml-org/llama.cpp:full
2
 
3
  WORKDIR /app
4
 
 
5
  RUN apt update && apt install -y python3 python3-pip python3-venv
 
 
6
  RUN python3 -m venv /opt/venv
7
  ENV PATH="/opt/venv/bin:$PATH"
8
 
 
9
  RUN pip install -U pip huggingface_hub
10
 
 
11
  RUN python3 -c 'from huggingface_hub import hf_hub_download; \
12
- repo="HauhauCS/Gemma-4-E2B-Uncensored-HauhauCS-Aggressive"; \
13
- hf_hub_download(repo_id=repo, filename="Gemma-4-E2B-Uncensored-HauhauCS-Aggressive-Q6_K_P.gguf", local_dir="/app"); \
14
- hf_hub_download(repo_id=repo, filename="mmproj-Gemma-4-E2B-Uncensored-HauhauCS-Aggressive-f16.gguf", local_dir="/app")'
15
 
 
 
16
  CMD ["--server", \
17
- "-m", "/app/Gemma-4-E2B-Uncensored-HauhauCS-Aggressive-Q6_K_P.gguf", \
18
- "--mmproj", "/app/mmproj-Gemma-4-E2B-Uncensored-HauhauCS-Aggressive-f16.gguf", \
19
  "--host", "0.0.0.0", \
20
  "--port", "7860", \
21
  "-t", "2", \
22
  "--cache-type-k", "q8_0", \
23
  "--cache-type-v", "iq4_nl", \
24
- "-c", "128000", \
25
  "-n", "38912"]
 
1
+ # Base image llama.cpp se hi le rahe hain
2
  FROM ghcr.io/ggml-org/llama.cpp:full
3
 
4
  WORKDIR /app
5
 
6
+ # System dependencies install karna
7
  RUN apt update && apt install -y python3 python3-pip python3-venv
8
+
9
+ # Virtual environment setup
10
  RUN python3 -m venv /opt/venv
11
  ENV PATH="/opt/venv/bin:$PATH"
12
 
13
+ # Python tools install karna
14
  RUN pip install -U pip huggingface_hub
15
 
16
+ # Model download karna (Qwen2.5-Coder-1.5B)
17
  RUN python3 -c 'from huggingface_hub import hf_hub_download; \
18
+ repo="bartowski/Qwen2.5-Coder-1.5B-Instruct-GGUF"; \
19
+ hf_hub_download(repo_id=repo, filename="Qwen2.5-Coder-1.5B-Instruct-Q6_K.gguf", local_dir="/app")'
 
20
 
21
+ # Server start karne ka command
22
+ # Humne mmproj hata diya hai taaki error na aaye
23
  CMD ["--server", \
24
+ "-m", "/app/Qwen2.5-Coder-1.5B-Instruct-Q6_K.gguf", \
 
25
  "--host", "0.0.0.0", \
26
  "--port", "7860", \
27
  "-t", "2", \
28
  "--cache-type-k", "q8_0", \
29
  "--cache-type-v", "iq4_nl", \
30
+ "-c", "32768", \
31
  "-n", "38912"]