The solution is to give max_length
:
llama = pipeline(
"conversational",
model="meta-llama/Llama-2-7b-chat-hf",
max_length=2000
)
The solution is to give max_length
:
llama = pipeline(
"conversational",
model="meta-llama/Llama-2-7b-chat-hf",
max_length=2000
)