File size: 506 Bytes
444b0ba
 
be70a3c
 
 
 
444b0ba
 
2ceea3f
444b0ba
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
---
hub:
  repo_id: TheBloke/wizard-vicuna-13B-GGML
  filename: wizard-vicuna-13B.ggmlv3.q4_1.bin
  # repo_id: TheBloke/Wizard-Vicuna-13B-Uncensored-GGML
  # filename: Wizard-Vicuna-13B-Uncensored.ggmlv3.q4_1.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