search

capstone-2024-T91 / Image-Processing-and-NLP-for-Brand-Protection

0 stars 0 forks source link

no progress display #13

Closed Youssef2430 closed 1 month ago

Youssef2430 commented 1 month ago

When using the flag -v there is no progress bar display showing updates

Youssef2430 commented 1 month ago

When running with the -v flag:

python main.py -r -train -v

You should see output similar to:

Loading tokenizer from: roberta-base
Loading Tokenizer: 100%|██████████| 1/1 [00:00<00:00,  2.50step/s]
Loading model from: roberta-base
Loading Model: 100%|██████████| 1/1 [00:01<00:00,  1.00s/step]
Loading dataset...
Loading Dataset: 100%|██████████| 1/1 [00:00<00:00,  2.00step/s]
Tokenizing Training Data: 100%|██████████| 800/800 [00:10<00:00, 77.00batch/s]
Tokenizing Evaluation Data: 100%|██████████| 200/200 [00:02<00:00, 80.00batch/s]
Starting training...
[INFO|trainer.py:1235] ***** Running training *****
[INFO|trainer.py:1236]   Num examples = 800
[INFO|trainer.py:1237]   Num Epochs = 3
[INFO|trainer.py:1238]   Instantaneous batch size per device = 8
[INFO|trainer.py:1239]   Total train batch size (w. parallel, distributed & accumulation) = 8
[INFO|trainer.py:1240]   Gradient Accumulation steps = 1
[INFO|trainer.py:1241]   Total optimization steps = 300
Epoch:   0%|          | 0/3 [00:00<?, ?epoch/s]
  0%|          | 0/100 [00:00<?, ?batch/s]
...
[INFO|trainer.py:1929] Saving model checkpoint to ./models/roberta-base_fine_tuned
Model fine-tuned and saved at './models/roberta-base_fine_tuned'