Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Update generation.py #531

Open
wants to merge 1 commit into
base: main
Choose a base branch
from
Open

Update generation.py #531

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
40 changes: 24 additions & 16 deletions bark/generation.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -264,6 +264,8 @@ def load_model(use_gpu=True, use_small=False, force_reload=False, model_type="te
raise NotImplementedError()
global models
global models_devices
gc.collect()
torch.cuda.empty_cache()
device = _grab_best_device(use_gpu=use_gpu)
model_key = f"{model_type}"
if OFFLOAD_CPU:
Expand Down Expand Up @@ -301,6 +303,7 @@ def load_codec_model(use_gpu=True, force_reload=False):


def preload_models(
preload_type="text",
text_use_gpu=True,
text_use_small=False,
coarse_use_gpu=True,
Expand All @@ -315,19 +318,24 @@ def preload_models(
text_use_gpu or coarse_use_gpu or fine_use_gpu or codec_use_gpu
):
logger.warning("No GPU being used. Careful, inference might be very slow!")
_ = load_model(
model_type="text", use_gpu=text_use_gpu, use_small=text_use_small, force_reload=force_reload
)
_ = load_model(
model_type="coarse",
use_gpu=coarse_use_gpu,
use_small=coarse_use_small,
force_reload=force_reload,
)
_ = load_model(
model_type="fine", use_gpu=fine_use_gpu, use_small=fine_use_small, force_reload=force_reload
)
_ = load_codec_model(use_gpu=codec_use_gpu, force_reload=force_reload)
models.clear()
if preload_type == "text":
_ = load_model(
model_type="text", use_gpu=text_use_gpu, use_small=text_use_small, force_reload=force_reload
)
elif preload_type == "coarse":
_ = load_model(
model_type="coarse",
use_gpu=coarse_use_gpu,
use_small=coarse_use_small,
force_reload=force_reload,
)
elif preload_type == "fine":
_ = load_model(
model_type="fine", use_gpu=fine_use_gpu, use_small=fine_use_small, force_reload=force_reload
)
elif preload_type == "codec":
_ = load_codec_model(use_gpu=codec_use_gpu, force_reload=force_reload)


####
Expand Down Expand Up @@ -593,7 +601,7 @@ def generate_coarse(
global models
global models_devices
if "coarse" not in models:
preload_models()
preload_models(preload_type="coarse")
model = models["coarse"]
if OFFLOAD_CPU:
model.to(models_devices["coarse"])
Expand Down Expand Up @@ -721,7 +729,7 @@ def generate_fine(
global models
global models_devices
if "fine" not in models:
preload_models()
preload_models(preload_type="fine")
model = models["fine"]
if OFFLOAD_CPU:
model.to(models_devices["fine"])
Expand Down Expand Up @@ -803,7 +811,7 @@ def codec_decode(fine_tokens):
global models
global models_devices
if "codec" not in models:
preload_models()
preload_models(preload_type="codec")
model = models["codec"]
if OFFLOAD_CPU:
model.to(models_devices["codec"])
Expand Down