YanoljaNEXT-Rosetta-4B-2511-GGUF

# Example: use a Q5_K_M quantized file (adjust the path/model to your choice)
MODEL="path/to/YanoljaNEXT-Rosetta-4B-2511-q5_k_m.gguf"

# Build a formatted prompt using the included chat template roles
# (see release/YanoljaNEXT-Rosetta-4B-2511/chat_template.jinja)
read -r -d '' PROMPT <<'EOT'
<start_of_turn>instruction
Translate the user's text to Korean. Keep the JSON structure and keys.
Context: Simple introduction about a tech company.
Tone: Informative and helpful
Glossary:
- Yanolja NEXT -> 야놀자넥스트
- travel industry -> 여행 산업
Output format: JSON
Provide the final translation immediately without any other text.
<end_of_turn>
<start_of_turn>source
{"company_name": "Yanolja NEXT", "description": "Yanolja NEXT is a company that provides cutting-edge technology for the global travel industry."}
<end_of_turn>
<start_of_turn>translation\n
EOT

# Run llama.cpp (adjust -n/-c/--temp as needed)
llama-cli -m "$MODEL" -p "$PROMPT" -n 64 -c 4096 --temp 0.7 -no-cnv

MODEL="path/to/YanoljaNEXT-Rosetta-4B-2511-q5_k_m.gguf"
llama-server -m "$MODEL" -c 4096 --host 0.0.0.0 --port 8080