microsoft / pdp-solver Goto Github PK

Thanks in advance for any assistance you might be able to provide!

There are important files that Microsoft projects should all have that are not present in this repository. A pull request has been opened to add the missing file(s). When the pr is merged this issue will be closed automatically.

Microsoft teams can learn more about this effort and share feedback within the open source guidance available internally.

Merge this pull request

Thank you very much for your work and this repo.

I face an issue because I do not have the validation set, despite the parameter -g is True

Traceback (most recent call last):
File "src/satyr-train-test.py", line 148, in
args.cpu_mode, args.reset, args.use_generator, args.batch_replication)
File "src/satyr-train-test.py", line 103, in run
train_epoch_size=config['train_epoch_size'])
File "./src/pdp/factorgraph/base.py", line 326, in train
num_workers=self._num_cores, max_cache_size=self._config['max_cache_size'])
File "./src/pdp/factorgraph/dataset.py", line 201, in get_loader
batch_replication=batch_replication)
File "./src/pdp/factorgraph/dataset.py", line 92, in init
with open(self._input_file, 'r') as fh_input:

FileNotFoundError: [Errno 2] No such file or directory: '../../datasets/SAT/M-4SAT-validation-2/M-4SAT-validation-2_0_7.0_10.0.json'

Could you provide the validation and the test set please ?

Thnak you in advance

I have a trouble with multi-GPU training:

graph_map.shape in fn _train_batch file base: torch.Size([2, 5479]) cuda:0
graph_map.shape: in fn forward file solver torch.Size([1, 5479]) cuda:0
graph_map.shape: in fn forward file solver torch.Size([1, 5479]) cuda:1

It seems that the dataparallel divided the graph_map by a wrong way. But it does not work if you simply set dim = 1 for dataparallel obviously.
How can I train the model by multi-GPU?

What is the size of your training set for these two experiments in Section 6.3 and Section 6.4?
In the NeuralSAT paper, they trained with use "millions of problems".

What is the reasonable size to reproduce similar results?