File size: 1,270 Bytes
e97665c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
MODEL_PATH = ""
# if MODEL_PATH is "", default llama.cpp/gptq models 
# will be downloaded to: ./models

# Example ggml path:
# MODEL_PATH = "./models/llama-2-7b-chat.ggmlv3.q4_0.bin"
# MODEL_PATH = "./models/Llama-2-7b-Chat-GPTQ"

# options: llama.cpp, gptq, transformers
BACKEND_TYPE = "llama.cpp"

# only for transformers bitsandbytes 8 bit
LOAD_IN_8BIT = False

MAX_MAX_NEW_TOKENS = 2048
DEFAULT_MAX_NEW_TOKENS = 1024
MAX_INPUT_TOKEN_LENGTH = 4000

DEFAULT_SYSTEM_PROMPT = "
You are a movie recommender chatbot. You give movie recommendations to users based on their profile. Your job now is to fully understand the user profile based on the given context and give them recommendations based on their input. Here are some rules for you to follow while generating a response:
1: Give an explanation for why each of the recommendations is a good fit for the user
2: Give a maximum of 5 recommendations, unless specified otherwise by the user
3: Give a predicted rating for the movie on a scale of 1 to 5: this is a rating the user would give to the movie if they watched it
4: Mention how popular the movie is. Choose from among High, Medium, Low: High being most popular, Low being least
5: Avoid recommending movies already rated by the user

''' User Context '''
"