The RWKV Language Model
https://github.com/BlinkDL/ChatRWKV
https://github.com/BlinkDL/RWKV-LM
# set these before import RWKV
os.environ['RWKV_JIT_ON'] = '1'
os.environ["RWKV_CUDA_ON"] = '0' # if '1' then compile CUDA kernel for seq mode (much faster)
########################################################################################################
#
# Use '/' in model path, instead of '\'. Use ctx4096 models if you need long ctx.
#
# fp16 = good for GPU (!!! DOES NOT support CPU !!!)
# fp32 = good for CPU
# bf16 = worse accuracy, supports CPU
#
# Strategy examples: (device = cpu/cuda/cuda:0/cuda:1/...)
# Here we consider [ln_out+head] to be an extra layer, so L12-D768 model has "13" layers, L24-D2048 model has "25" layers, etc.
#
# 'cpu fp32' = everything on cpu fp32
# 'cuda fp16' = everything on cuda fp16
#
# 'cuda fp16 *6 -> cpu fp32' = first 6 layers on cuda fp16, then on cpu fp32
# 'cuda:0 fp16 *10 -> cuda:1 fp16 *8 -> cpu fp32' = first 10 layers on cuda:0 fp16, then 8 layers on cuda:1 fp16, then on cpu fp32
#
# Use '+' for STREAM mode (do it on your fastest GPU), requires some VRAM to store streamed layers
# 'cuda fp16 *6+' = first 6 layers on cuda fp16, then stream the rest on it
# (for best speed: try *1+ *2+ *3+ ... until you run out of VRAM)
#
# Extreme STREAM: 3G VRAM is enough to run RWKV 14B (slow. will be faster in future)
# 'cuda fp16 *0+ -> cpu fp32 *1' = stream all layers on cuda fp16, then [ln_out+head] on cpu fp32
#
# ########################################################################################################
from rwkv.model import RWKV
from rwkv.utils import PIPELINE, PIPELINE_ARGS
pipeline = PIPELINE(model, "20B_tokenizer.json")
model = RWKV(model='/fsx/BlinkDL/HF-MODEL/rwkv-4-pile-169m/RWKV-4-Pile-169M-20220807-8023', strategy='cpu fp32')
ctx = "\nIn a shocking finding, scientist discovered a herd of dragons living in a remote, previously unexplored valley, in Tibet. Even more surprising to the researchers was the fact that the dragons spoke perfect Chinese."
print(ctx, end='')
def my_print(s):
print(s, end='', flush=True)
# For alpha_frequency and alpha_presence, see "Frequency and presence penalties":
# https://platform.openai.com/docs/api-reference/parameter-details
args = PIPELINE_ARGS(temperature = 1.0, top_p = 0.7,
alpha_frequency = 0.25,
alpha_presence = 0.25,
token_ban = [0], # ban the generation of some tokens
token_stop = []) # stop generation whenever you see any token here
pipeline.generate(ctx, token_count=512, args=args, callback=my_print)
print('\n')
out, state = model.forward([187, 510, 1563, 310, 247], None)
print(out.detach().cpu().numpy()) # get logits
out, state = model.forward([187, 510], None)
out, state = model.forward([1563], state) # RNN has state (use deepcopy if you want to clone it)
out, state = model.forward([310, 247], state)
print(out.detach().cpu().numpy()) # same result as above
print('\n')
1 maintainer
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
rwkv-0.0.4.tar.gz
(13.9 kB
view details)
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
rwkv-0.0.4-py3-none-any.whl
(13.5 kB
view details)
File details
Details for the file rwkv-0.0.4.tar.gz.
File metadata
- Download URL: rwkv-0.0.4.tar.gz
- Upload date:
- Size: 13.9 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via:
twine/4.0.2 CPython/3.9.12
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
57a188b3a2d3b0dfe5ad02f1b47e3379d8bf341e5ab94863213659507441b176
|
|
| MD5 |
a78c99b9fea2c19423bc0b74784f1549
|
|
| BLAKE2b-256 |
d1d35b4d5a889a8b1180ae4aefbb03ea1108478ba0d01043506a47b871e39a54
|
File details
Details for the file rwkv-0.0.4-py3-none-any.whl.
File metadata
- Download URL: rwkv-0.0.4-py3-none-any.whl
- Upload date:
- Size: 13.5 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via:
twine/4.0.2 CPython/3.9.12
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
f540862d964c2fb1bbe94ec9c4a79eccce5ce70770360f37ba8b4e02984876d3
|
|
| MD5 |
76bcd4d06ca253be3f1cc7cf1862bfe6
|
|
| BLAKE2b-256 |
8cae6c3b05947573e106b344db03ee88e0895607eed9896bbe74062ab7d01c1e
|
