Adam Goodge

Trouble downloading datasets

1

When I click to download the .csv files of the datasets, I am sent to a raw version of the data in the browser and no file is downloaded

During training: "ValueError: could not convert string to float" because of delimiters

5

"ValueError: could not convert string to float: '0.0\t1.0\t11.238836854\t3.746958' " As you can see, the delimiters are set to \t and so the data becomes a string which cannot be converted...

D_data_loss and G_discriminator_loss don't change

15

As in the title, the adversarial losses don't change at all from 1.398 and 0.693 resepectively after roughly epoch 2 until end. Though G_l2_loss does change. Any ideas whats wrong?...

Pre-trained checkpoints

5

Thanks for the repo, Are there pre-trained model checkpoints available? Thanks

Hi, thanks for the repo. Have you made any model checkpoints available? Thanks

AttributeError: 'STLlamaConfig' object has no attribute 'lin_hidden_size'

2

When I execute run_urbangpt.py with --model-name facebook/opt350m, I get the following error: AttributeError: 'STLlamaConfig' object has no attribute 'lin_hidden_size' All of the other arguments are left unchanged from the .sh...

Model weights

1

Hello, Thank you the code release. Are you planning to release the pre-trained model weights soon?

Hi, When I call train/val/test, the input is appropriately pre-processed in the train/val/test_step function. If I call predict after doing this, the input is also pre-processed similarly. However, if I...

Using pre-trained model without prompts

4

Hello, I have a few questions: 1) ` if args.prompt_ST==1: if args.file_load_path != '': model.load_state_dict(torch.load('{}'.format(args.file_load_path),map_location=device), strict=False) ` I see that if using prompt_ST, then the model weights will be loaded...