| Before load dataset, RAM used: 0.36 GB | Avaiable: 200.44 GB | Left: 200.08 GB | |
| After load dataset, RAM used: 0.61 GB | Avaiable: 200.42 GB | Left: 199.80 GB | |
| After Prepare Dataloader, RAM used: 6.94 GB | Avaiable: 203.21 GB | Left: 196.27 GB | |
| After epoch 1, RAM used: 27.96 GB | Avaiable: 206.69 GB | Left: 178.73 GB | |
| >>> Epoch 1: Perplexity: 8.71756492102604 Loss: 1.987068991905017 | |
| Loss improved inf -> 1.987068991905017 | |
| Saved training checkpoint | |
| After epoch 2, RAM used: 32.61 GB | Avaiable: 217.10 GB | Left: 184.49 GB | |
| >>> Epoch 2: Perplexity: 6.334900481327126 Loss: 1.7609998571673966 | |
| Loss improved 1.987068991905017 -> 1.7609998571673966 | |
| Saved training checkpoint | |
| After epoch 3, RAM used: 32.60 GB | Avaiable: 220.86 GB | Left: 188.26 GB | |
| >>> Epoch 3: Perplexity: 5.783759349968259 Loss: 1.714263437903529 | |
| Loss improved 1.7609998571673966 -> 1.714263437903529 | |
| Saved training checkpoint | |
| After epoch 4, RAM used: 32.33 GB | Avaiable: 192.09 GB | Left: 159.77 GB | |
| >>> Epoch 4: Perplexity: 4.530374708434244 Loss: 1.5004713556098934 | |
| Loss improved 1.714263437903529 -> 1.5004713556098934 | |
| Saved training checkpoint | |