Generates text from predefined prompts using `Llama 2`.
Example::
# returns string completion
"What is Python?" | kapi.complete()
# returns list of string completions
["What is Python?", "What is C++?"] | kapi.complete().all()
# returns list of string completions. The prompts sent to the server are ["\n\n\nPlease summarize the above paragraph", ...]
["", " "] | kapi.complete("Please summarize the above paragraph").all()
- VRAM: 22GB
- Throughput: 8/s
def inner(*args, **kwargs): # batchify
output = [...]; idx, event = con.add([output, args, kwargs]) # batchify
t = threading.Thread(target=lambda: event.wait()); t.start(); t.join() # batchify
return output[0] # batchify