Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,35 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
### TRAINING LOG
|
| 2 |
+
wandb: Run history:
|
| 3 |
+
wandb: eval/loss █▆▅▄▃▃▂▂▁▁▁
|
| 4 |
+
wandb: eval/runtime ▁▃▂▃▃▃▃█▃▄▁
|
| 5 |
+
wandb: eval/samples_per_second █▆▇▆▆▆▆▁▆▄█
|
| 6 |
+
wandb: eval/steps_per_second █▆▇▆▆▆▆▁▆▄█
|
| 7 |
+
wandb: train/epoch ▁▁▁▂▂▂▂▂▂▂▃▃▃▃▃▄▄▄▄▄▅▅▅▅▅▅▆▆▆▆▆▇▇▇▇▇▇███
|
| 8 |
+
wandb: train/global_step ▁▁▁▂▂▂▂▂▂▃▃▃▃▃▄▄▄▄▄▄▅▅▅▅▅▅▆▆▆▆▆▇▇▇▇▇▇███
|
| 9 |
+
wandb: train/learning_rate ▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁
|
| 10 |
+
wandb: train/loss █▄▄▅▃▅▃▃▄▅▃▃▃▄▃▃▃▃▂▂▂▂▃▂▄▂▃▂▂▂▂▂▃▂▁▃▂▂▂▁
|
| 11 |
+
wandb: train/total_flos ▁
|
| 12 |
+
wandb: train/train_loss ▁
|
| 13 |
+
wandb: train/train_runtime ▁
|
| 14 |
+
wandb: train/train_samples_per_second ▁
|
| 15 |
+
wandb: train/train_steps_per_second ▁
|
| 16 |
+
wandb:
|
| 17 |
+
wandb: Run summary:
|
| 18 |
+
wandb: eval/loss 0.27314
|
| 19 |
+
wandb: eval/runtime 129.6563
|
| 20 |
+
wandb: eval/samples_per_second 7.713
|
| 21 |
+
wandb: eval/steps_per_second 7.713
|
| 22 |
+
wandb: train/epoch 0.53
|
| 23 |
+
wandb: train/global_step 1875
|
| 24 |
+
wandb: train/learning_rate 0.0002
|
| 25 |
+
wandb: train/loss 0.258
|
| 26 |
+
wandb: train/total_flos 1.9547706216175334e+17
|
| 27 |
+
wandb: train/train_loss 0.30445
|
| 28 |
+
wandb: train/train_runtime 13368.3721
|
| 29 |
+
wandb: train/train_samples_per_second 2.244
|
| 30 |
+
wandb: train/train_steps_per_second 0.14
|
| 31 |
+
wandb:
|
| 32 |
+
wandb: 🚀 View run happy-deluge-17 at: https://wandb.ai/metric/llm_finetune_multiwoz22.sh/runs/4epf9h85
|
| 33 |
+
|
| 34 |
+
### INFERENCE LOG
|
| 35 |
+
TODO
|