text-generation-webui/modules/RWKV.py

111 lines
3.2 KiB
Python
Raw Normal View History

2023-02-27 22:09:11 -05:00
import os
2023-02-27 21:50:16 -05:00
from pathlib import Path
2023-03-07 16:17:56 -05:00
from queue import Queue
from threading import Thread
2023-02-27 22:09:11 -05:00
2023-02-27 21:50:16 -05:00
import numpy as np
2023-03-06 06:45:49 -05:00
from tokenizers import Tokenizer
2023-02-27 22:09:11 -05:00
import modules.shared as shared
2023-02-27 21:50:16 -05:00
np.set_printoptions(precision=4, suppress=True, linewidth=200)
os.environ['RWKV_JIT_ON'] = '1'
os.environ["RWKV_CUDA_ON"] = '1' if shared.args.rwkv_cuda_on else '0' # use CUDA kernel for seq mode (much faster)
2023-02-27 21:50:16 -05:00
from rwkv.model import RWKV
from rwkv.utils import PIPELINE, PIPELINE_ARGS
2023-03-01 10:18:17 -05:00
2023-03-01 10:08:55 -05:00
class RWKVModel:
def __init__(self):
pass
2023-02-27 21:50:16 -05:00
2023-03-01 10:08:55 -05:00
@classmethod
def from_pretrained(self, path, dtype="fp16", device="cuda"):
tokenizer_path = Path(f"{path.parent}/20B_tokenizer.json")
2023-02-27 21:50:16 -05:00
2023-03-01 18:02:48 -05:00
if shared.args.rwkv_strategy is None:
model = RWKV(model=os.path.abspath(path), strategy=f'{device} {dtype}')
else:
model = RWKV(model=os.path.abspath(path), strategy=shared.args.rwkv_strategy)
2023-03-01 17:17:16 -05:00
pipeline = PIPELINE(model, os.path.abspath(tokenizer_path))
2023-02-27 21:50:16 -05:00
2023-03-01 10:08:55 -05:00
result = self()
2023-03-01 10:33:09 -05:00
result.pipeline = pipeline
2023-03-01 10:08:55 -05:00
return result
2023-03-07 16:17:56 -05:00
def generate(self, context="", token_count=20, temperature=1, top_p=1, top_k=50, alpha_frequency=0.1, alpha_presence=0.1, token_ban=[0], token_stop=[], callback=None):
2023-03-01 10:16:11 -05:00
args = PIPELINE_ARGS(
temperature = temperature,
top_p = top_p,
2023-03-07 15:24:28 -05:00
top_k = top_k,
2023-03-01 10:19:37 -05:00
alpha_frequency = alpha_frequency, # Frequency Penalty (as in GPT-3)
alpha_presence = alpha_presence, # Presence Penalty (as in GPT-3)
token_ban = token_ban, # ban the generation of some tokens
token_stop = token_stop
2023-03-01 10:16:11 -05:00
)
2023-03-01 14:40:25 -05:00
return context+self.pipeline.generate(context, token_count=token_count, args=args, callback=callback)
2023-03-06 06:45:49 -05:00
2023-03-07 16:17:56 -05:00
def generate_with_streaming(self, **kwargs):
iterable = Iteratorize(self.generate, kwargs, callback=None)
reply = kwargs['context']
for token in iterable:
reply += token
yield reply
2023-03-06 06:45:49 -05:00
class RWKVTokenizer:
def __init__(self):
pass
@classmethod
def from_pretrained(self, path):
tokenizer_path = path / "20B_tokenizer.json"
tokenizer = Tokenizer.from_file(os.path.abspath(tokenizer_path))
result = self()
result.tokenizer = tokenizer
return result
def encode(self, prompt):
return self.tokenizer.encode(prompt).ids
def decode(self, ids):
return self.tokenizer.decode(ids)
2023-03-07 16:17:56 -05:00
class Iteratorize:
"""
Transforms a function that takes a callback
into a lazy iterator (generator).
"""
def __init__(self, func, kwargs={}, callback=None):
self.mfunc=func
self.c_callback=callback
self.q = Queue(maxsize=1)
self.sentinel = object()
self.kwargs = kwargs
def _callback(val):
self.q.put(val)
def gentask():
ret = self.mfunc(callback=_callback, **self.kwargs)
self.q.put(self.sentinel)
if self.c_callback:
self.c_callback(ret)
Thread(target=gentask).start()
def __iter__(self):
return self
def __next__(self):
obj = self.q.get(True,None)
if obj is self.sentinel:
raise StopIteration
else:
return obj