def main(pytorch: bool = False, gpu_id: int = -1):
global CONFIG
fix_random_seed(0)
if gpu_id >= 0:
require_gpu(gpu_id)
print("Set GPU", gpu_id)
backends = {"pytorch": pytorch}
for name, use_backend in backends.items():
if not use_backend:
print(f"Skipping {name}")
continue
set_backend(name, gpu_id)
C = registry.resolve(Config().from_str(CONFIG))
model = C["model"]
X, Y = get_dummy_data(**C["data"])
print("Copy to device")
X = [model.ops.asarray(x) for x in X]
Y = [model.ops.asarray(y) for y in Y]
print("Begin init", len(X))
model.initialize(X=X[:5])
print("Pre-batch")
n_words = sum(len(x) for x in X)
X = [
model.layers[0].predict(batch)
for batch in model.ops.minibatch(16, X)
]
model.layers.pop(0)
print("Start")
start_time = timer()
end_time = timer()
print(name, n_words, end_time - start_time)
def setup_gpu(use_gpu: int) -> None:
"""Configure the GPU and log info."""
if use_gpu >= 0:
msg.info(f"Using GPU: {use_gpu}")
require_gpu(use_gpu)
else:
msg.info("Using CPU")
def _resolve_gpu(self, use_gpu: int) -> int:
if use_gpu >= 0:
gpu_id = int(os.environ.get("CUDA_VISIBLE_DEVICES", -1))
logger.info(f"Using GPU (isolated): {gpu_id}")
require_gpu(0)
else:
logger.info("Using CPU")
gpu_id = -1
return gpu_id
def setup_gpu(use_gpu: int) -> None:
"""Configure the GPU and log info."""
if use_gpu >= 0:
msg.info(f"Using GPU: {use_gpu}")
require_gpu(use_gpu)
else:
msg.info("Using CPU")
if has_cupy and gpu_is_available():
msg.info("To switch to GPU 0, use the option: --gpu-id 0")
def thread_training(training_step_iterator, print_row, rank, num_workers, gpu_id):
if gpu_id >= 0:
# I don't fully understand why we need to do this within the thread.
# I think 0 is also correct here, because ray sets the available devices?
require_gpu(0)
for batch, info, is_best_checkpoint in training_step_iterator:
if rank == 0 and is_best_checkpoint is not None:
info["words"] *= num_workers
print_row(info)
def setup_gpu(use_gpu: int, silent=None) -> None:
"""Configure the GPU and log info."""
if silent is None:
local_msg = Printer()
else:
local_msg = Printer(no_print=silent, pretty=not silent)
if use_gpu >= 0:
local_msg.info(f"Using GPU: {use_gpu}")
require_gpu(use_gpu)
else:
local_msg.info("Using CPU")
if gpu_is_available():
local_msg.info("To switch to GPU 0, use the option: --gpu-id 0")
def main(numpy: bool = False,
pytorch: bool = False,
generic: bool = False,
gpu_id: int = -1):
global CONFIG
fix_random_seed(0)
if gpu_id >= 0:
require_gpu(gpu_id)
print("Set GPU", gpu_id)
backends = {"pytorch": pytorch, "numpy": numpy, "generic": generic}
for name, use_backend in backends.items():
if not use_backend:
print(f"Skipping {name}")
continue
set_backend(name, gpu_id)
print("Getting data")
C = registry.resolve(Config().from_str(CONFIG))
model = C["model"]
X, Y = get_dummy_data(**C["data"])
print("Copy to device")
X = [model.ops.asarray(x) for x in X]
Y = [model.ops.asarray(y) for y in Y]
print("Begin init", len(X))
model.initialize(X=X[:5])
print("Pre-batch")
n_words = sum(len(x) for x in X)
batches = model.ops.multibatch(16, X, Y)
batches = [(model.layers[0].predict(x), y) for x, y in batches]
model.layers.pop(0)
print("Start")
start_time = timer()
total = run_forward(model, [x for x, y in batches])
end_time = timer()
print(name, n_words, total, end_time - start_time)
start_time = timer()
total = run_forward_backward(model, batches)
end_time = timer()
print(name, n_words, total, end_time - start_time)
import spacy
from copy import copy
from spacy_transformers.pipeline_component import DEFAULT_CONFIG
from thinc.api import Config, set_gpu_allocator, require_gpu
if spacy.prefer_gpu():
print("\n\033[92m" + "✔ Using GPU" + "\033[0m\n")
set_gpu_allocator("pytorch")
require_gpu(0)
else:
print("\n\033[91m" + "✘ NOT Using GPU!" + "\033[0m\n")
config = copy(DEFAULT_CONFIG["transformer"])
config["model"]["name"] = "model/distilbert-base-nli-stsb-mean-tokens"
nlp = spacy.blank("en")
transformer = nlp.add_pipe("transformer", config=config)
transformer.model.initialize()
doc = nlp("hello world")
tokvecs = doc._.trf_data.tensors[-1]
print(tokvecs)
