RoversX's picture
Update config.yml
faf1664
raw
history blame
397 Bytes
---
hub:
repo_id: TheBloke/Stable-Platypus2-13B-GGML
filename: stable-platypus2-13b.ggmlv3.q3_K_L.bin
llama_cpp:
n_ctx: 2048
n_gpu_layers: 40 # llama 13b has 40 layers
chat:
stop:
- "</s>"
- "<unk>"
- "### USER:"
- "USER:"
queue:
max_size: 16
concurrency_count: 1 # leave this at 1, llama-cpp-python doesn't handle concurrent requests and will crash the entire app