Skip to content

PC computer using wls ubuntu, have one GPU 4070ti, how to set cuda memory? #1

Closed
@kehan777

Description

@kehan777

Initial GPU Usage
| ID | GPU | MEM |

| 0 | 4% | 6% |
GPU Usage after emptying the cache
| ID | GPU | MEM |

| 0 | 6% | 6% |
Lightning automatically upgraded your loaded checkpoint from v1.9.5 to v2.0.5. To apply the upgrade to your files permanently, run python -m pytorch_lightning.utilities.upgrade_checkpoint --file geodock/weights/dips.ckpt
Completed embedding in 0.77 seconds.
Traceback (most recent call last):
File "/home/adsb/GeoDock/test.py", line 59, in
pred = geodock.dock(
File "/home/adsb/GeoDock/geodock/GeoDockRunner.py", line 72, in dock
dock(
File "/home/adsb/GeoDock/geodock/utils/docking.py", line 22, in dock
model_out = model(model_in)
File "/home/adsb/miniconda3/envs/gk/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "/home/adsb/GeoDock/geodock/model/GeoDock.py", line 66, in forward
lddt_logits, dist_logits, coords, rotat, trans = self.net(
File "/home/adsb/miniconda3/envs/gk/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "/home/adsb/GeoDock/geodock/model/modules/iterative_transformer.py", line 71, in forward
node, edge = self.graph_module(node_0 + recycled_node, edge_0 + recycled_edge)
File "/home/adsb/miniconda3/envs/gk/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "/home/adsb/GeoDock/geodock/model/modules/graph_module.py", line 284, in forward
node, edge = block(node, edge)
File "/home/adsb/miniconda3/envs/gk/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "/home/adsb/GeoDock/geodock/model/modules/graph_module.py", line 264, in forward
edge = edge + self.node_to_edge(node)
File "/home/adsb/miniconda3/envs/gk/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "/home/adsb/GeoDock/geodock/model/modules/graph_module.py", line 236, in forward
x = torch.cat([prod, diff], dim=-1)
torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 772.00 MiB (GPU 0; 11.99 GiB total capacity; 10.60 GiB already allocated; 0 bytes free; 10.77 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions