File size: 405 Bytes
5c1f654
 
ff24341
 
5c1f654
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
---
hub:
  repo_id: RoversX/tableBeluga-7B-instruct-pl-lora-Samantha-data-V1-ggml
  filename: ggml-model-q4_0.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