|
from model_LXM2T5 import T5tokenizer, LXMT52T5, LXMtokenizer |
|
import tqdm |
|
from dataset_val4LXMT5 import KgDatasetVal |
|
model = LXMT52T5() |
|
model.module.load_state_dict(torch.load("xxxx.pth")) |
|
test_dataset = KgDatasetVal(val=False) |
|
|
|
|
|
test_dataloader = DataLoader(dataset=test_dataset, batch_size=args.batch_size, shuffle=False, |
|
num_workers=0, collate_fn=my_val_collate) |
|
|
|
model.eval() |
|
answers = [] |
|
preds = [] |
|
preds_list = [] |
|
answers_list = [] |
|
id2pred_list = {} |
|
for i, batch_data in enumerate(tqdm(test_dataloader)): |
|
with torch.no_grad(): |
|
val_T5_input_id = torch.stack(batch_data['T5_input_ids']).to(device) |
|
val_T5_input_mask = torch.stack(batch_data['T5_input_masks']).to(device) |
|
val_visual_faetures = torch.tensor(np.array(batch_data['img'])).float().to(device) |
|
val_spatial_features = torch.tensor(np.array(batch_data['spatial'])).float().to(device) |
|
|
|
val_LXM_input_id = torch.stack(batch_data['LXM_input_ids']).to(device) |
|
val_LXM_input_mask = torch.stack(batch_data['LXM_input_masks']).to(device) |
|
val_LXM_token_type_ids = torch.stack(batch_data['LXM_token_type_ids']).to(device) |
|
|
|
|
|
|
|
val_outputs = model(train=False, LXM_source_ids=val_LXM_input_id, LXM_source_masks=val_LXM_input_mask,T5_source_ids=val_T5_input_id, T5_source_masks=val_T5_input_mask,token_type_ids=val_LXM_token_type_ids, visual_features=val_visual_faetures, spatial_features=val_spatial_features,T5_target_ids=None) |
|
|
|
|
|
|
|
val_list_predict = T5tokenizer.batch_decode(val_outputs, skip_special_tokens=True) |
|
|
|
|
|
|
|
for i, pre in enumerate(batch_data['ans']): |
|
|
|
preds_list.append(val_list_predict[i]) |
|
|
|
answers_list.append(batch_data['ans'][i]) |
|
|
|
id2pred_list[str(batch_data['id'][i])]=val_list_predict[i] |
|
|
|
|
|
|
|
|
|
f=open("file_to_save.json", 'w') |
|
json.dump(id2pred_list, f) |
|
f.close() |
|
|