ftobler/llama
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

tuned prompt

Browse Source
This commit is contained in:
Florin Tobler
2025-01-13 23:33:51 +01:00
parent a697f49698
commit 0c022d4731
1 changed files with 4 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
llama.py
View File

@@ -13,7 +13,7 @@ register_dummy()
def initialize_config(inference: Inference) -> Terminal: def initialize_config(inference: Inference) -> Terminal:
# systemmessage at the very begin of the chat. Will be concatenated with the automatic tool usage descriptions # systemmessage at the very begin of the chat. Will be concatenated with the automatic tool usage descriptions
system_prompt = "Hold a casual conversation with the user. Keep responses short at max 5 sentences and on point. Answer using markdown to the user. When providing code examples, avoid comments which provide no additional information." system_prompt = "Hold a casual conversation with the user. Keep responses short at max 5 sentences and on point. Answer using markdown to the user. When providing code examples, avoid comments which provide no additional information. Do not summarize."
current_date_and_time = datetime.datetime.now().strftime("Current date is %Y-%m-%d and its %H:%M %p right now.") current_date_and_time = datetime.datetime.now().strftime("Current date is %Y-%m-%d and its %H:%M %p right now.")
append_toolcalls = False append_toolcalls = False
if append_toolcalls: if append_toolcalls: