Add files using upload-large-folder tool
Browse files- finetuning/extra/bytes_Extra1_MSA_Dialect_Combined_1x_ckpt/best_model.bin +3 -0
- msa_curriculum/reset/pct50/bytes.complete +0 -0
- msa_curriculum/reset/pct50/experiment.complete +0 -0
- msa_curriculum/reset/pct50/finetuning/bytes/MSA_Dialect_Combined_1x/checkpoint-3000/trainer_state.json +1816 -0
- msa_curriculum/reset/pct50/finetuning/bytes/MSA_Dialect_Combined_1x/checkpoint-4000/trainer_state.json +2416 -0
- msa_curriculum/reset/pct50/finetuning/bytes/MSA_Dialect_Combined_1x/training_complete.json +19 -0
- msa_curriculum/reset/pct50/manifest.json +51 -0
- msa_curriculum/reset/pct50/pretraining/bytes/alignment/checkpoint-20000/optimizer.pt +3 -0
- msa_curriculum/reset/pct50/pretraining/bytes/alignment/checkpoint-20000/pytorch_model.bin +3 -0
- msa_curriculum/reset/pct50/pretraining/bytes/alignment/checkpoint-20000/rng_state_0.pth +3 -0
- msa_curriculum/reset/pct50/pretraining/bytes/alignment/checkpoint-20000/rng_state_1.pth +3 -0
- msa_curriculum/reset/pct50/pretraining/bytes/alignment/checkpoint-20000/rng_state_2.pth +3 -0
- msa_curriculum/reset/pct50/pretraining/bytes/alignment/checkpoint-20000/rng_state_3.pth +3 -0
- msa_curriculum/reset/pct50/pretraining/bytes/alignment/checkpoint-20000/scheduler.pt +3 -0
- msa_curriculum/reset/pct50/pretraining/bytes/alignment/checkpoint-20000/trainer_state.json +0 -0
- msa_curriculum/reset/pct50/pretraining/bytes/alignment/checkpoint-20000/training_args.bin +3 -0
- msa_curriculum/reset/pct50/pretraining/bytes/alignment/checkpoint-22000/optimizer.pt +3 -0
- msa_curriculum/reset/pct50/pretraining/bytes/alignment/checkpoint-22000/pytorch_model.bin +3 -0
- msa_curriculum/reset/pct50/pretraining/bytes/alignment/checkpoint-22000/rng_state_0.pth +3 -0
- msa_curriculum/reset/pct50/pretraining/bytes/alignment/checkpoint-22000/rng_state_1.pth +3 -0
- msa_curriculum/reset/pct50/pretraining/bytes/alignment/checkpoint-22000/rng_state_2.pth +3 -0
- msa_curriculum/reset/pct50/pretraining/bytes/alignment/checkpoint-22000/rng_state_3.pth +3 -0
- msa_curriculum/reset/pct50/pretraining/bytes/alignment/checkpoint-22000/scheduler.pt +3 -0
- msa_curriculum/reset/pct50/pretraining/bytes/alignment/checkpoint-22000/trainer_state.json +0 -0
- msa_curriculum/reset/pct50/pretraining/bytes/alignment/checkpoint-22000/training_args.bin +3 -0
- msa_curriculum/reset/pct50/pretraining/bytes/alignment/final/pytorch_model.bin +3 -0
- msa_curriculum/reset/pct50/pretraining/bytes/alignment/final/training_complete.json +21 -0
- msa_curriculum/reset/pct50/pretraining/bytes/full_epoch1/checkpoint-20000/pytorch_model.bin +3 -0
- msa_curriculum/reset/pct50/pretraining/bytes/full_epoch1/checkpoint-20000/rng_state_0.pth +3 -0
- msa_curriculum/reset/pct50/pretraining/bytes/full_epoch1/checkpoint-20000/rng_state_1.pth +3 -0
- msa_curriculum/reset/pct50/pretraining/bytes/full_epoch1/checkpoint-20000/rng_state_2.pth +3 -0
- msa_curriculum/reset/pct50/pretraining/bytes/full_epoch1/checkpoint-20000/rng_state_3.pth +3 -0
- msa_curriculum/reset/pct50/pretraining/bytes/full_epoch1/checkpoint-20000/scheduler.pt +3 -0
- msa_curriculum/reset/pct50/pretraining/bytes/full_epoch1/checkpoint-20000/trainer_state.json +0 -0
- msa_curriculum/reset/pct50/pretraining/bytes/full_epoch1/checkpoint-20000/training_args.bin +3 -0
- msa_curriculum/reset/pct50/pretraining/bytes/full_epoch1/checkpoint-22000/pytorch_model.bin +3 -0
- msa_curriculum/reset/pct50/pretraining/bytes/full_epoch1/checkpoint-22000/rng_state_0.pth +3 -0
- msa_curriculum/reset/pct50/pretraining/bytes/full_epoch1/checkpoint-22000/rng_state_1.pth +3 -0
- msa_curriculum/reset/pct50/pretraining/bytes/full_epoch1/checkpoint-22000/rng_state_2.pth +3 -0
- msa_curriculum/reset/pct50/pretraining/bytes/full_epoch1/checkpoint-22000/rng_state_3.pth +3 -0
- msa_curriculum/reset/pct50/pretraining/bytes/full_epoch1/checkpoint-22000/scheduler.pt +3 -0
- msa_curriculum/reset/pct50/pretraining/bytes/full_epoch1/checkpoint-22000/trainer_state.json +0 -0
- msa_curriculum/reset/pct50/pretraining/bytes/full_epoch1/checkpoint-22000/training_args.bin +3 -0
- msa_curriculum/reset/pct50/pretraining/bytes/full_epoch1/final/training_complete.json +21 -0
- msa_curriculum/reset/pct50/pretraining/bytes/full_epoch2/checkpoint-20000/rng_state_2.pth +3 -0
- msa_curriculum/reset/pct50/pretraining/bytes/full_epoch2/checkpoint-20000/rng_state_3.pth +3 -0
- msa_curriculum/reset/pct50/pretraining/bytes/full_epoch2/checkpoint-20000/trainer_state.json +0 -0
- msa_curriculum/reset/pct50/pretraining/bytes/full_epoch2/checkpoint-22000/trainer_state.json +0 -0
- msa_curriculum/reset/pct50/pretraining/bytes/full_epoch2/final/training_complete.json +21 -0
- msa_curriculum/reset/pct50/run.lock +0 -0
finetuning/extra/bytes_Extra1_MSA_Dialect_Combined_1x_ckpt/best_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2362f12c7078adcd4f186934cf4423cdfe50bedcb76de4fbad0bed19e530262b
|
| 3 |
+
size 1933289043
|
msa_curriculum/reset/pct50/bytes.complete
ADDED
|
File without changes
|
msa_curriculum/reset/pct50/experiment.complete
ADDED
|
File without changes
|
msa_curriculum/reset/pct50/finetuning/bytes/MSA_Dialect_Combined_1x/checkpoint-3000/trainer_state.json
ADDED
|
@@ -0,0 +1,1816 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.6570661994195915,
|
| 5 |
+
"global_step": 3000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.0,
|
| 12 |
+
"learning_rate": 6.564551422319475e-07,
|
| 13 |
+
"loss": 1.7654,
|
| 14 |
+
"step": 10
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.0,
|
| 18 |
+
"learning_rate": 1.312910284463895e-06,
|
| 19 |
+
"loss": 1.7226,
|
| 20 |
+
"step": 20
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.01,
|
| 24 |
+
"learning_rate": 1.9693654266958425e-06,
|
| 25 |
+
"loss": 1.7008,
|
| 26 |
+
"step": 30
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.01,
|
| 30 |
+
"learning_rate": 2.62582056892779e-06,
|
| 31 |
+
"loss": 1.6721,
|
| 32 |
+
"step": 40
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.01,
|
| 36 |
+
"learning_rate": 3.2822757111597377e-06,
|
| 37 |
+
"loss": 1.7142,
|
| 38 |
+
"step": 50
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.01,
|
| 42 |
+
"learning_rate": 3.938730853391685e-06,
|
| 43 |
+
"loss": 1.7535,
|
| 44 |
+
"step": 60
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.02,
|
| 48 |
+
"learning_rate": 4.595185995623632e-06,
|
| 49 |
+
"loss": 1.6915,
|
| 50 |
+
"step": 70
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.02,
|
| 54 |
+
"learning_rate": 5.25164113785558e-06,
|
| 55 |
+
"loss": 1.6418,
|
| 56 |
+
"step": 80
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.02,
|
| 60 |
+
"learning_rate": 5.908096280087528e-06,
|
| 61 |
+
"loss": 1.6122,
|
| 62 |
+
"step": 90
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.02,
|
| 66 |
+
"learning_rate": 6.564551422319475e-06,
|
| 67 |
+
"loss": 1.6245,
|
| 68 |
+
"step": 100
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.02,
|
| 72 |
+
"learning_rate": 7.221006564551422e-06,
|
| 73 |
+
"loss": 1.6182,
|
| 74 |
+
"step": 110
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.03,
|
| 78 |
+
"learning_rate": 7.87746170678337e-06,
|
| 79 |
+
"loss": 1.4952,
|
| 80 |
+
"step": 120
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.03,
|
| 84 |
+
"learning_rate": 8.533916849015316e-06,
|
| 85 |
+
"loss": 1.6204,
|
| 86 |
+
"step": 130
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.03,
|
| 90 |
+
"learning_rate": 9.190371991247264e-06,
|
| 91 |
+
"loss": 1.5313,
|
| 92 |
+
"step": 140
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.03,
|
| 96 |
+
"learning_rate": 9.846827133479212e-06,
|
| 97 |
+
"loss": 1.5215,
|
| 98 |
+
"step": 150
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.04,
|
| 102 |
+
"learning_rate": 1.050328227571116e-05,
|
| 103 |
+
"loss": 1.5439,
|
| 104 |
+
"step": 160
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.04,
|
| 108 |
+
"learning_rate": 1.1159737417943107e-05,
|
| 109 |
+
"loss": 1.5977,
|
| 110 |
+
"step": 170
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.04,
|
| 114 |
+
"learning_rate": 1.1816192560175055e-05,
|
| 115 |
+
"loss": 1.5625,
|
| 116 |
+
"step": 180
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.04,
|
| 120 |
+
"learning_rate": 1.2472647702407003e-05,
|
| 121 |
+
"loss": 1.5324,
|
| 122 |
+
"step": 190
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.04,
|
| 126 |
+
"learning_rate": 1.312910284463895e-05,
|
| 127 |
+
"loss": 1.479,
|
| 128 |
+
"step": 200
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.05,
|
| 132 |
+
"learning_rate": 1.3785557986870897e-05,
|
| 133 |
+
"loss": 1.4969,
|
| 134 |
+
"step": 210
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.05,
|
| 138 |
+
"learning_rate": 1.4442013129102845e-05,
|
| 139 |
+
"loss": 1.5098,
|
| 140 |
+
"step": 220
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.05,
|
| 144 |
+
"learning_rate": 1.509846827133479e-05,
|
| 145 |
+
"loss": 1.5546,
|
| 146 |
+
"step": 230
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.05,
|
| 150 |
+
"learning_rate": 1.575492341356674e-05,
|
| 151 |
+
"loss": 1.5069,
|
| 152 |
+
"step": 240
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.05,
|
| 156 |
+
"learning_rate": 1.6411378555798686e-05,
|
| 157 |
+
"loss": 1.493,
|
| 158 |
+
"step": 250
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.06,
|
| 162 |
+
"learning_rate": 1.7067833698030632e-05,
|
| 163 |
+
"loss": 1.4697,
|
| 164 |
+
"step": 260
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.06,
|
| 168 |
+
"learning_rate": 1.7724288840262582e-05,
|
| 169 |
+
"loss": 1.4476,
|
| 170 |
+
"step": 270
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.06,
|
| 174 |
+
"learning_rate": 1.8380743982494528e-05,
|
| 175 |
+
"loss": 1.4063,
|
| 176 |
+
"step": 280
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.06,
|
| 180 |
+
"learning_rate": 1.9037199124726478e-05,
|
| 181 |
+
"loss": 1.4337,
|
| 182 |
+
"step": 290
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.07,
|
| 186 |
+
"learning_rate": 1.9693654266958424e-05,
|
| 187 |
+
"loss": 1.4452,
|
| 188 |
+
"step": 300
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.07,
|
| 192 |
+
"learning_rate": 2.0350109409190373e-05,
|
| 193 |
+
"loss": 1.4853,
|
| 194 |
+
"step": 310
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.07,
|
| 198 |
+
"learning_rate": 2.100656455142232e-05,
|
| 199 |
+
"loss": 1.4319,
|
| 200 |
+
"step": 320
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.07,
|
| 204 |
+
"learning_rate": 2.1663019693654265e-05,
|
| 205 |
+
"loss": 1.4208,
|
| 206 |
+
"step": 330
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.07,
|
| 210 |
+
"learning_rate": 2.2319474835886215e-05,
|
| 211 |
+
"loss": 1.4323,
|
| 212 |
+
"step": 340
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.08,
|
| 216 |
+
"learning_rate": 2.297592997811816e-05,
|
| 217 |
+
"loss": 1.4242,
|
| 218 |
+
"step": 350
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.08,
|
| 222 |
+
"learning_rate": 2.363238512035011e-05,
|
| 223 |
+
"loss": 1.4216,
|
| 224 |
+
"step": 360
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.08,
|
| 228 |
+
"learning_rate": 2.4288840262582056e-05,
|
| 229 |
+
"loss": 1.3798,
|
| 230 |
+
"step": 370
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.08,
|
| 234 |
+
"learning_rate": 2.4945295404814006e-05,
|
| 235 |
+
"loss": 1.4481,
|
| 236 |
+
"step": 380
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.09,
|
| 240 |
+
"learning_rate": 2.5601750547045952e-05,
|
| 241 |
+
"loss": 1.448,
|
| 242 |
+
"step": 390
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.09,
|
| 246 |
+
"learning_rate": 2.62582056892779e-05,
|
| 247 |
+
"loss": 1.3897,
|
| 248 |
+
"step": 400
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.09,
|
| 252 |
+
"learning_rate": 2.6914660831509848e-05,
|
| 253 |
+
"loss": 1.4447,
|
| 254 |
+
"step": 410
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.09,
|
| 258 |
+
"learning_rate": 2.7571115973741794e-05,
|
| 259 |
+
"loss": 1.3803,
|
| 260 |
+
"step": 420
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 0.09,
|
| 264 |
+
"learning_rate": 2.8227571115973743e-05,
|
| 265 |
+
"loss": 1.414,
|
| 266 |
+
"step": 430
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.1,
|
| 270 |
+
"learning_rate": 2.888402625820569e-05,
|
| 271 |
+
"loss": 1.3641,
|
| 272 |
+
"step": 440
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.1,
|
| 276 |
+
"learning_rate": 2.954048140043764e-05,
|
| 277 |
+
"loss": 1.3955,
|
| 278 |
+
"step": 450
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 0.1,
|
| 282 |
+
"learning_rate": 2.9999960523153055e-05,
|
| 283 |
+
"loss": 1.4183,
|
| 284 |
+
"step": 460
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 0.1,
|
| 288 |
+
"learning_rate": 2.999925871832113e-05,
|
| 289 |
+
"loss": 1.3418,
|
| 290 |
+
"step": 470
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.11,
|
| 294 |
+
"learning_rate": 2.9997679697467503e-05,
|
| 295 |
+
"loss": 1.4104,
|
| 296 |
+
"step": 480
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.11,
|
| 300 |
+
"learning_rate": 2.9995223552939574e-05,
|
| 301 |
+
"loss": 1.3313,
|
| 302 |
+
"step": 490
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 0.11,
|
| 306 |
+
"learning_rate": 2.9991890428382386e-05,
|
| 307 |
+
"loss": 1.311,
|
| 308 |
+
"step": 500
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 0.11,
|
| 312 |
+
"learning_rate": 2.9987680518730264e-05,
|
| 313 |
+
"loss": 1.3834,
|
| 314 |
+
"step": 510
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 0.11,
|
| 318 |
+
"learning_rate": 2.998259407019538e-05,
|
| 319 |
+
"loss": 1.3476,
|
| 320 |
+
"step": 520
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 0.12,
|
| 324 |
+
"learning_rate": 2.997663138025339e-05,
|
| 325 |
+
"loss": 1.3777,
|
| 326 |
+
"step": 530
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 0.12,
|
| 330 |
+
"learning_rate": 2.9969792797625994e-05,
|
| 331 |
+
"loss": 1.3502,
|
| 332 |
+
"step": 540
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.12,
|
| 336 |
+
"learning_rate": 2.9962078722260575e-05,
|
| 337 |
+
"loss": 1.3632,
|
| 338 |
+
"step": 550
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.12,
|
| 342 |
+
"learning_rate": 2.99534896053068e-05,
|
| 343 |
+
"loss": 1.4011,
|
| 344 |
+
"step": 560
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 0.12,
|
| 348 |
+
"learning_rate": 2.9944025949090224e-05,
|
| 349 |
+
"loss": 1.3331,
|
| 350 |
+
"step": 570
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 0.13,
|
| 354 |
+
"learning_rate": 2.9933688307082926e-05,
|
| 355 |
+
"loss": 1.2906,
|
| 356 |
+
"step": 580
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.13,
|
| 360 |
+
"learning_rate": 2.9922477283871123e-05,
|
| 361 |
+
"loss": 1.3344,
|
| 362 |
+
"step": 590
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 0.13,
|
| 366 |
+
"learning_rate": 2.991039353511983e-05,
|
| 367 |
+
"loss": 1.3746,
|
| 368 |
+
"step": 600
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 0.13,
|
| 372 |
+
"learning_rate": 2.9897437767534515e-05,
|
| 373 |
+
"loss": 1.3082,
|
| 374 |
+
"step": 610
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.14,
|
| 378 |
+
"learning_rate": 2.988361073881974e-05,
|
| 379 |
+
"loss": 1.3623,
|
| 380 |
+
"step": 620
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.14,
|
| 384 |
+
"learning_rate": 2.9868913257634876e-05,
|
| 385 |
+
"loss": 1.2997,
|
| 386 |
+
"step": 630
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 0.14,
|
| 390 |
+
"learning_rate": 2.985334618354681e-05,
|
| 391 |
+
"loss": 1.2911,
|
| 392 |
+
"step": 640
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 0.14,
|
| 396 |
+
"learning_rate": 2.9836910426979645e-05,
|
| 397 |
+
"loss": 1.3226,
|
| 398 |
+
"step": 650
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 0.14,
|
| 402 |
+
"learning_rate": 2.981960694916148e-05,
|
| 403 |
+
"loss": 1.3156,
|
| 404 |
+
"step": 660
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 0.15,
|
| 408 |
+
"learning_rate": 2.98014367620682e-05,
|
| 409 |
+
"loss": 1.3288,
|
| 410 |
+
"step": 670
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 0.15,
|
| 414 |
+
"learning_rate": 2.9782400928364266e-05,
|
| 415 |
+
"loss": 1.3732,
|
| 416 |
+
"step": 680
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.15,
|
| 420 |
+
"learning_rate": 2.976250056134058e-05,
|
| 421 |
+
"loss": 1.368,
|
| 422 |
+
"step": 690
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.15,
|
| 426 |
+
"learning_rate": 2.9741736824849383e-05,
|
| 427 |
+
"loss": 1.3157,
|
| 428 |
+
"step": 700
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 0.16,
|
| 432 |
+
"learning_rate": 2.972011093323617e-05,
|
| 433 |
+
"loss": 1.3349,
|
| 434 |
+
"step": 710
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 0.16,
|
| 438 |
+
"learning_rate": 2.9697624151268684e-05,
|
| 439 |
+
"loss": 1.3584,
|
| 440 |
+
"step": 720
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 0.16,
|
| 444 |
+
"learning_rate": 2.967427779406295e-05,
|
| 445 |
+
"loss": 1.3826,
|
| 446 |
+
"step": 730
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 0.16,
|
| 450 |
+
"learning_rate": 2.965007322700634e-05,
|
| 451 |
+
"loss": 1.2787,
|
| 452 |
+
"step": 740
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 0.16,
|
| 456 |
+
"learning_rate": 2.9625011865677757e-05,
|
| 457 |
+
"loss": 1.3617,
|
| 458 |
+
"step": 750
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 0.17,
|
| 462 |
+
"learning_rate": 2.95990951757648e-05,
|
| 463 |
+
"loss": 1.3281,
|
| 464 |
+
"step": 760
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.17,
|
| 468 |
+
"learning_rate": 2.9572324672978096e-05,
|
| 469 |
+
"loss": 1.3359,
|
| 470 |
+
"step": 770
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 0.17,
|
| 474 |
+
"learning_rate": 2.9544701922962598e-05,
|
| 475 |
+
"loss": 1.3062,
|
| 476 |
+
"step": 780
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 0.17,
|
| 480 |
+
"learning_rate": 2.951622854120608e-05,
|
| 481 |
+
"loss": 1.3376,
|
| 482 |
+
"step": 790
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 0.18,
|
| 486 |
+
"learning_rate": 2.948690619294461e-05,
|
| 487 |
+
"loss": 1.3341,
|
| 488 |
+
"step": 800
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 0.18,
|
| 492 |
+
"learning_rate": 2.9456736593065172e-05,
|
| 493 |
+
"loss": 1.3207,
|
| 494 |
+
"step": 810
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 0.18,
|
| 498 |
+
"learning_rate": 2.9425721506005403e-05,
|
| 499 |
+
"loss": 1.2939,
|
| 500 |
+
"step": 820
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 0.18,
|
| 504 |
+
"learning_rate": 2.9393862745650353e-05,
|
| 505 |
+
"loss": 1.3064,
|
| 506 |
+
"step": 830
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.18,
|
| 510 |
+
"learning_rate": 2.9361162175226433e-05,
|
| 511 |
+
"loss": 1.3241,
|
| 512 |
+
"step": 840
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 0.19,
|
| 516 |
+
"learning_rate": 2.932762170719245e-05,
|
| 517 |
+
"loss": 1.3197,
|
| 518 |
+
"step": 850
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 0.19,
|
| 522 |
+
"learning_rate": 2.929324330312773e-05,
|
| 523 |
+
"loss": 1.3123,
|
| 524 |
+
"step": 860
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 0.19,
|
| 528 |
+
"learning_rate": 2.925802897361742e-05,
|
| 529 |
+
"loss": 1.3165,
|
| 530 |
+
"step": 870
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 0.19,
|
| 534 |
+
"learning_rate": 2.9221980778134912e-05,
|
| 535 |
+
"loss": 1.2716,
|
| 536 |
+
"step": 880
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 0.19,
|
| 540 |
+
"learning_rate": 2.9185100824921362e-05,
|
| 541 |
+
"loss": 1.3049,
|
| 542 |
+
"step": 890
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 0.2,
|
| 546 |
+
"learning_rate": 2.9147391270862414e-05,
|
| 547 |
+
"loss": 1.3517,
|
| 548 |
+
"step": 900
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.2,
|
| 552 |
+
"learning_rate": 2.910885432136206e-05,
|
| 553 |
+
"loss": 1.2833,
|
| 554 |
+
"step": 910
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 0.2,
|
| 558 |
+
"learning_rate": 2.9069492230213643e-05,
|
| 559 |
+
"loss": 1.2964,
|
| 560 |
+
"step": 920
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 0.2,
|
| 564 |
+
"learning_rate": 2.9029307299468067e-05,
|
| 565 |
+
"loss": 1.32,
|
| 566 |
+
"step": 930
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 0.21,
|
| 570 |
+
"learning_rate": 2.8988301879299137e-05,
|
| 571 |
+
"loss": 1.2736,
|
| 572 |
+
"step": 940
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 0.21,
|
| 576 |
+
"learning_rate": 2.894647836786615e-05,
|
| 577 |
+
"loss": 1.2839,
|
| 578 |
+
"step": 950
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 0.21,
|
| 582 |
+
"learning_rate": 2.89038392111736e-05,
|
| 583 |
+
"loss": 1.3297,
|
| 584 |
+
"step": 960
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 0.21,
|
| 588 |
+
"learning_rate": 2.8860386902928147e-05,
|
| 589 |
+
"loss": 1.3177,
|
| 590 |
+
"step": 970
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.21,
|
| 594 |
+
"learning_rate": 2.8816123984392777e-05,
|
| 595 |
+
"loss": 1.3007,
|
| 596 |
+
"step": 980
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 0.22,
|
| 600 |
+
"learning_rate": 2.8771053044238174e-05,
|
| 601 |
+
"loss": 1.3428,
|
| 602 |
+
"step": 990
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 0.22,
|
| 606 |
+
"learning_rate": 2.872517671839132e-05,
|
| 607 |
+
"loss": 1.3025,
|
| 608 |
+
"step": 1000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 0.22,
|
| 612 |
+
"learning_rate": 2.8678497689881354e-05,
|
| 613 |
+
"loss": 1.3186,
|
| 614 |
+
"step": 1010
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 0.22,
|
| 618 |
+
"learning_rate": 2.863101868868262e-05,
|
| 619 |
+
"loss": 1.2949,
|
| 620 |
+
"step": 1020
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 0.23,
|
| 624 |
+
"learning_rate": 2.858274249155505e-05,
|
| 625 |
+
"loss": 1.2859,
|
| 626 |
+
"step": 1030
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 0.23,
|
| 630 |
+
"learning_rate": 2.853367192188175e-05,
|
| 631 |
+
"loss": 1.2447,
|
| 632 |
+
"step": 1040
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.23,
|
| 636 |
+
"learning_rate": 2.848380984950386e-05,
|
| 637 |
+
"loss": 1.34,
|
| 638 |
+
"step": 1050
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 0.23,
|
| 642 |
+
"learning_rate": 2.8433159190552752e-05,
|
| 643 |
+
"loss": 1.2706,
|
| 644 |
+
"step": 1060
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 0.23,
|
| 648 |
+
"learning_rate": 2.8381722907279456e-05,
|
| 649 |
+
"loss": 1.2778,
|
| 650 |
+
"step": 1070
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 0.24,
|
| 654 |
+
"learning_rate": 2.8329504007881413e-05,
|
| 655 |
+
"loss": 1.2487,
|
| 656 |
+
"step": 1080
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 0.24,
|
| 660 |
+
"learning_rate": 2.8276505546326554e-05,
|
| 661 |
+
"loss": 1.3054,
|
| 662 |
+
"step": 1090
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 0.24,
|
| 666 |
+
"learning_rate": 2.8222730622174704e-05,
|
| 667 |
+
"loss": 1.3033,
|
| 668 |
+
"step": 1100
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 0.24,
|
| 672 |
+
"learning_rate": 2.8168182380396286e-05,
|
| 673 |
+
"loss": 1.3092,
|
| 674 |
+
"step": 1110
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.25,
|
| 678 |
+
"learning_rate": 2.8112864011188392e-05,
|
| 679 |
+
"loss": 1.2893,
|
| 680 |
+
"step": 1120
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 0.25,
|
| 684 |
+
"learning_rate": 2.8056778749788224e-05,
|
| 685 |
+
"loss": 1.2901,
|
| 686 |
+
"step": 1130
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 0.25,
|
| 690 |
+
"learning_rate": 2.7999929876283874e-05,
|
| 691 |
+
"loss": 1.2615,
|
| 692 |
+
"step": 1140
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 0.25,
|
| 696 |
+
"learning_rate": 2.7942320715422486e-05,
|
| 697 |
+
"loss": 1.3097,
|
| 698 |
+
"step": 1150
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 0.25,
|
| 702 |
+
"learning_rate": 2.788395463641583e-05,
|
| 703 |
+
"loss": 1.2495,
|
| 704 |
+
"step": 1160
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 0.26,
|
| 708 |
+
"learning_rate": 2.7824835052743235e-05,
|
| 709 |
+
"loss": 1.315,
|
| 710 |
+
"step": 1170
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 0.26,
|
| 714 |
+
"learning_rate": 2.7764965421951963e-05,
|
| 715 |
+
"loss": 1.2946,
|
| 716 |
+
"step": 1180
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.26,
|
| 720 |
+
"learning_rate": 2.770434924545501e-05,
|
| 721 |
+
"loss": 1.3103,
|
| 722 |
+
"step": 1190
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 0.26,
|
| 726 |
+
"learning_rate": 2.764299006832631e-05,
|
| 727 |
+
"loss": 1.2656,
|
| 728 |
+
"step": 1200
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 0.27,
|
| 732 |
+
"learning_rate": 2.7580891479093433e-05,
|
| 733 |
+
"loss": 1.3298,
|
| 734 |
+
"step": 1210
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 0.27,
|
| 738 |
+
"learning_rate": 2.7518057109527675e-05,
|
| 739 |
+
"loss": 1.3019,
|
| 740 |
+
"step": 1220
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 0.27,
|
| 744 |
+
"learning_rate": 2.745449063443169e-05,
|
| 745 |
+
"loss": 1.2962,
|
| 746 |
+
"step": 1230
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 0.27,
|
| 750 |
+
"learning_rate": 2.7390195771424554e-05,
|
| 751 |
+
"loss": 1.3314,
|
| 752 |
+
"step": 1240
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 0.27,
|
| 756 |
+
"learning_rate": 2.7325176280724352e-05,
|
| 757 |
+
"loss": 1.2564,
|
| 758 |
+
"step": 1250
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.28,
|
| 762 |
+
"learning_rate": 2.725943596492828e-05,
|
| 763 |
+
"loss": 1.2918,
|
| 764 |
+
"step": 1260
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 0.28,
|
| 768 |
+
"learning_rate": 2.7192978668790217e-05,
|
| 769 |
+
"loss": 1.2837,
|
| 770 |
+
"step": 1270
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 0.28,
|
| 774 |
+
"learning_rate": 2.7125808278995908e-05,
|
| 775 |
+
"loss": 1.2939,
|
| 776 |
+
"step": 1280
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 0.28,
|
| 780 |
+
"learning_rate": 2.705792872393562e-05,
|
| 781 |
+
"loss": 1.2619,
|
| 782 |
+
"step": 1290
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 0.28,
|
| 786 |
+
"learning_rate": 2.6989343973474434e-05,
|
| 787 |
+
"loss": 1.278,
|
| 788 |
+
"step": 1300
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 0.29,
|
| 792 |
+
"learning_rate": 2.6920058038720026e-05,
|
| 793 |
+
"loss": 1.2826,
|
| 794 |
+
"step": 1310
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 0.29,
|
| 798 |
+
"learning_rate": 2.685007497178812e-05,
|
| 799 |
+
"loss": 1.2275,
|
| 800 |
+
"step": 1320
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.29,
|
| 804 |
+
"learning_rate": 2.6779398865565487e-05,
|
| 805 |
+
"loss": 1.3291,
|
| 806 |
+
"step": 1330
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 0.29,
|
| 810 |
+
"learning_rate": 2.670803385347058e-05,
|
| 811 |
+
"loss": 1.2972,
|
| 812 |
+
"step": 1340
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 0.3,
|
| 816 |
+
"learning_rate": 2.6635984109211786e-05,
|
| 817 |
+
"loss": 1.2323,
|
| 818 |
+
"step": 1350
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 0.3,
|
| 822 |
+
"learning_rate": 2.656325384654335e-05,
|
| 823 |
+
"loss": 1.2529,
|
| 824 |
+
"step": 1360
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 0.3,
|
| 828 |
+
"learning_rate": 2.648984731901892e-05,
|
| 829 |
+
"loss": 1.2411,
|
| 830 |
+
"step": 1370
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 0.3,
|
| 834 |
+
"learning_rate": 2.641576881974279e-05,
|
| 835 |
+
"loss": 1.2481,
|
| 836 |
+
"step": 1380
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 0.3,
|
| 840 |
+
"learning_rate": 2.6341022681118828e-05,
|
| 841 |
+
"loss": 1.2779,
|
| 842 |
+
"step": 1390
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.31,
|
| 846 |
+
"learning_rate": 2.6265613274597087e-05,
|
| 847 |
+
"loss": 1.2818,
|
| 848 |
+
"step": 1400
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 0.31,
|
| 852 |
+
"learning_rate": 2.6189545010418157e-05,
|
| 853 |
+
"loss": 1.281,
|
| 854 |
+
"step": 1410
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 0.31,
|
| 858 |
+
"learning_rate": 2.611282233735523e-05,
|
| 859 |
+
"loss": 1.2877,
|
| 860 |
+
"step": 1420
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 0.31,
|
| 864 |
+
"learning_rate": 2.6035449742453915e-05,
|
| 865 |
+
"loss": 1.296,
|
| 866 |
+
"step": 1430
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 0.32,
|
| 870 |
+
"learning_rate": 2.5957431750769836e-05,
|
| 871 |
+
"loss": 1.2733,
|
| 872 |
+
"step": 1440
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 0.32,
|
| 876 |
+
"learning_rate": 2.5878772925103954e-05,
|
| 877 |
+
"loss": 1.214,
|
| 878 |
+
"step": 1450
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 0.32,
|
| 882 |
+
"learning_rate": 2.5799477865735766e-05,
|
| 883 |
+
"loss": 1.2658,
|
| 884 |
+
"step": 1460
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.32,
|
| 888 |
+
"learning_rate": 2.5719551210154207e-05,
|
| 889 |
+
"loss": 1.3124,
|
| 890 |
+
"step": 1470
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 0.32,
|
| 894 |
+
"learning_rate": 2.5638997632786468e-05,
|
| 895 |
+
"loss": 1.3069,
|
| 896 |
+
"step": 1480
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 0.33,
|
| 900 |
+
"learning_rate": 2.5557821844724625e-05,
|
| 901 |
+
"loss": 1.2606,
|
| 902 |
+
"step": 1490
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 0.33,
|
| 906 |
+
"learning_rate": 2.5476028593450075e-05,
|
| 907 |
+
"loss": 1.2651,
|
| 908 |
+
"step": 1500
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 0.33,
|
| 912 |
+
"learning_rate": 2.5393622662555914e-05,
|
| 913 |
+
"loss": 1.2719,
|
| 914 |
+
"step": 1510
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 0.33,
|
| 918 |
+
"learning_rate": 2.5310608871467174e-05,
|
| 919 |
+
"loss": 1.2728,
|
| 920 |
+
"step": 1520
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 0.34,
|
| 924 |
+
"learning_rate": 2.522699207515895e-05,
|
| 925 |
+
"loss": 1.2169,
|
| 926 |
+
"step": 1530
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.34,
|
| 930 |
+
"learning_rate": 2.5142777163872472e-05,
|
| 931 |
+
"loss": 1.3261,
|
| 932 |
+
"step": 1540
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 0.34,
|
| 936 |
+
"learning_rate": 2.5057969062829107e-05,
|
| 937 |
+
"loss": 1.2942,
|
| 938 |
+
"step": 1550
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 0.34,
|
| 942 |
+
"learning_rate": 2.4972572731942294e-05,
|
| 943 |
+
"loss": 1.2167,
|
| 944 |
+
"step": 1560
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 0.34,
|
| 948 |
+
"learning_rate": 2.488659316552749e-05,
|
| 949 |
+
"loss": 1.2768,
|
| 950 |
+
"step": 1570
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 0.35,
|
| 954 |
+
"learning_rate": 2.4800035392010082e-05,
|
| 955 |
+
"loss": 1.2514,
|
| 956 |
+
"step": 1580
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 0.35,
|
| 960 |
+
"learning_rate": 2.471290447363128e-05,
|
| 961 |
+
"loss": 1.2344,
|
| 962 |
+
"step": 1590
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 0.35,
|
| 966 |
+
"learning_rate": 2.4625205506152083e-05,
|
| 967 |
+
"loss": 1.242,
|
| 968 |
+
"step": 1600
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.35,
|
| 972 |
+
"learning_rate": 2.4536943618555256e-05,
|
| 973 |
+
"loss": 1.2482,
|
| 974 |
+
"step": 1610
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 0.35,
|
| 978 |
+
"learning_rate": 2.444812397274535e-05,
|
| 979 |
+
"loss": 1.2741,
|
| 980 |
+
"step": 1620
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 0.36,
|
| 984 |
+
"learning_rate": 2.4358751763246824e-05,
|
| 985 |
+
"loss": 1.2061,
|
| 986 |
+
"step": 1630
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 0.36,
|
| 990 |
+
"learning_rate": 2.4268832216900263e-05,
|
| 991 |
+
"loss": 1.2727,
|
| 992 |
+
"step": 1640
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 0.36,
|
| 996 |
+
"learning_rate": 2.4178370592556657e-05,
|
| 997 |
+
"loss": 1.2432,
|
| 998 |
+
"step": 1650
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 0.36,
|
| 1002 |
+
"learning_rate": 2.4087372180769888e-05,
|
| 1003 |
+
"loss": 1.2156,
|
| 1004 |
+
"step": 1660
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 0.37,
|
| 1008 |
+
"learning_rate": 2.399584230348726e-05,
|
| 1009 |
+
"loss": 1.2286,
|
| 1010 |
+
"step": 1670
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.37,
|
| 1014 |
+
"learning_rate": 2.390378631373832e-05,
|
| 1015 |
+
"loss": 1.264,
|
| 1016 |
+
"step": 1680
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 0.37,
|
| 1020 |
+
"learning_rate": 2.3811209595321716e-05,
|
| 1021 |
+
"loss": 1.3205,
|
| 1022 |
+
"step": 1690
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 0.37,
|
| 1026 |
+
"learning_rate": 2.37181175624904e-05,
|
| 1027 |
+
"loss": 1.2897,
|
| 1028 |
+
"step": 1700
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 0.37,
|
| 1032 |
+
"learning_rate": 2.362451565963495e-05,
|
| 1033 |
+
"loss": 1.233,
|
| 1034 |
+
"step": 1710
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 0.38,
|
| 1038 |
+
"learning_rate": 2.3530409360965126e-05,
|
| 1039 |
+
"loss": 1.2131,
|
| 1040 |
+
"step": 1720
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 0.38,
|
| 1044 |
+
"learning_rate": 2.34358041701898e-05,
|
| 1045 |
+
"loss": 1.2611,
|
| 1046 |
+
"step": 1730
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 0.38,
|
| 1050 |
+
"learning_rate": 2.3340705620194995e-05,
|
| 1051 |
+
"loss": 1.2365,
|
| 1052 |
+
"step": 1740
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.38,
|
| 1056 |
+
"learning_rate": 2.3245119272720344e-05,
|
| 1057 |
+
"loss": 1.208,
|
| 1058 |
+
"step": 1750
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 0.39,
|
| 1062 |
+
"learning_rate": 2.3149050718033815e-05,
|
| 1063 |
+
"loss": 1.2632,
|
| 1064 |
+
"step": 1760
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 0.39,
|
| 1068 |
+
"learning_rate": 2.3052505574604753e-05,
|
| 1069 |
+
"loss": 1.2436,
|
| 1070 |
+
"step": 1770
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 0.39,
|
| 1074 |
+
"learning_rate": 2.29554894887753e-05,
|
| 1075 |
+
"loss": 1.2279,
|
| 1076 |
+
"step": 1780
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 0.39,
|
| 1080 |
+
"learning_rate": 2.2858008134430182e-05,
|
| 1081 |
+
"loss": 1.3023,
|
| 1082 |
+
"step": 1790
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 0.39,
|
| 1086 |
+
"learning_rate": 2.2760067212664852e-05,
|
| 1087 |
+
"loss": 1.2554,
|
| 1088 |
+
"step": 1800
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 0.4,
|
| 1092 |
+
"learning_rate": 2.2661672451452112e-05,
|
| 1093 |
+
"loss": 1.2102,
|
| 1094 |
+
"step": 1810
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 0.4,
|
| 1098 |
+
"learning_rate": 2.2562829605307057e-05,
|
| 1099 |
+
"loss": 1.271,
|
| 1100 |
+
"step": 1820
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 0.4,
|
| 1104 |
+
"learning_rate": 2.2463544454950594e-05,
|
| 1105 |
+
"loss": 1.2741,
|
| 1106 |
+
"step": 1830
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 0.4,
|
| 1110 |
+
"learning_rate": 2.2363822806971322e-05,
|
| 1111 |
+
"loss": 1.2299,
|
| 1112 |
+
"step": 1840
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 0.41,
|
| 1116 |
+
"learning_rate": 2.2263670493485935e-05,
|
| 1117 |
+
"loss": 1.2402,
|
| 1118 |
+
"step": 1850
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 0.41,
|
| 1122 |
+
"learning_rate": 2.2163093371798165e-05,
|
| 1123 |
+
"loss": 1.2592,
|
| 1124 |
+
"step": 1860
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 0.41,
|
| 1128 |
+
"learning_rate": 2.20620973240562e-05,
|
| 1129 |
+
"loss": 1.2148,
|
| 1130 |
+
"step": 1870
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 0.41,
|
| 1134 |
+
"learning_rate": 2.1960688256908693e-05,
|
| 1135 |
+
"loss": 1.3015,
|
| 1136 |
+
"step": 1880
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 0.41,
|
| 1140 |
+
"learning_rate": 2.1858872101159293e-05,
|
| 1141 |
+
"loss": 1.2337,
|
| 1142 |
+
"step": 1890
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 0.42,
|
| 1146 |
+
"learning_rate": 2.175665481141982e-05,
|
| 1147 |
+
"loss": 1.2014,
|
| 1148 |
+
"step": 1900
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 0.42,
|
| 1152 |
+
"learning_rate": 2.1654042365761974e-05,
|
| 1153 |
+
"loss": 1.2187,
|
| 1154 |
+
"step": 1910
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 0.42,
|
| 1158 |
+
"learning_rate": 2.1551040765367757e-05,
|
| 1159 |
+
"loss": 1.2193,
|
| 1160 |
+
"step": 1920
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 0.42,
|
| 1164 |
+
"learning_rate": 2.1447656034178475e-05,
|
| 1165 |
+
"loss": 1.265,
|
| 1166 |
+
"step": 1930
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 0.42,
|
| 1170 |
+
"learning_rate": 2.134389421854243e-05,
|
| 1171 |
+
"loss": 1.2408,
|
| 1172 |
+
"step": 1940
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 0.43,
|
| 1176 |
+
"learning_rate": 2.1239761386861332e-05,
|
| 1177 |
+
"loss": 1.1964,
|
| 1178 |
+
"step": 1950
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 0.43,
|
| 1182 |
+
"learning_rate": 2.1135263629235358e-05,
|
| 1183 |
+
"loss": 1.254,
|
| 1184 |
+
"step": 1960
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 0.43,
|
| 1188 |
+
"learning_rate": 2.103040705710702e-05,
|
| 1189 |
+
"loss": 1.2297,
|
| 1190 |
+
"step": 1970
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 0.43,
|
| 1194 |
+
"learning_rate": 2.0925197802903702e-05,
|
| 1195 |
+
"loss": 1.2397,
|
| 1196 |
+
"step": 1980
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 0.44,
|
| 1200 |
+
"learning_rate": 2.0819642019679053e-05,
|
| 1201 |
+
"loss": 1.2802,
|
| 1202 |
+
"step": 1990
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 0.44,
|
| 1206 |
+
"learning_rate": 2.0713745880753093e-05,
|
| 1207 |
+
"loss": 1.2435,
|
| 1208 |
+
"step": 2000
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 0.44,
|
| 1212 |
+
"learning_rate": 2.0607515579351196e-05,
|
| 1213 |
+
"loss": 1.2494,
|
| 1214 |
+
"step": 2010
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 0.44,
|
| 1218 |
+
"learning_rate": 2.0500957328241884e-05,
|
| 1219 |
+
"loss": 1.2514,
|
| 1220 |
+
"step": 2020
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 0.44,
|
| 1224 |
+
"learning_rate": 2.0394077359373473e-05,
|
| 1225 |
+
"loss": 1.2539,
|
| 1226 |
+
"step": 2030
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 0.45,
|
| 1230 |
+
"learning_rate": 2.0286881923509602e-05,
|
| 1231 |
+
"loss": 1.2823,
|
| 1232 |
+
"step": 2040
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 0.45,
|
| 1236 |
+
"learning_rate": 2.0179377289863683e-05,
|
| 1237 |
+
"loss": 1.2449,
|
| 1238 |
+
"step": 2050
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 0.45,
|
| 1242 |
+
"learning_rate": 2.0071569745732223e-05,
|
| 1243 |
+
"loss": 1.2446,
|
| 1244 |
+
"step": 2060
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 0.45,
|
| 1248 |
+
"learning_rate": 1.9963465596127142e-05,
|
| 1249 |
+
"loss": 1.2498,
|
| 1250 |
+
"step": 2070
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 0.46,
|
| 1254 |
+
"learning_rate": 1.985507116340702e-05,
|
| 1255 |
+
"loss": 1.2032,
|
| 1256 |
+
"step": 2080
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 0.46,
|
| 1260 |
+
"learning_rate": 1.974639278690734e-05,
|
| 1261 |
+
"loss": 1.2074,
|
| 1262 |
+
"step": 2090
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 0.46,
|
| 1266 |
+
"learning_rate": 1.9637436822569742e-05,
|
| 1267 |
+
"loss": 1.2406,
|
| 1268 |
+
"step": 2100
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 0.46,
|
| 1272 |
+
"learning_rate": 1.9528209642570312e-05,
|
| 1273 |
+
"loss": 1.2512,
|
| 1274 |
+
"step": 2110
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 0.46,
|
| 1278 |
+
"learning_rate": 1.9418717634946886e-05,
|
| 1279 |
+
"loss": 1.2406,
|
| 1280 |
+
"step": 2120
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 0.47,
|
| 1284 |
+
"learning_rate": 1.9308967203225465e-05,
|
| 1285 |
+
"loss": 1.2125,
|
| 1286 |
+
"step": 2130
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 0.47,
|
| 1290 |
+
"learning_rate": 1.9198964766045744e-05,
|
| 1291 |
+
"loss": 1.2741,
|
| 1292 |
+
"step": 2140
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 0.47,
|
| 1296 |
+
"learning_rate": 1.9088716756785652e-05,
|
| 1297 |
+
"loss": 1.2423,
|
| 1298 |
+
"step": 2150
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 0.47,
|
| 1302 |
+
"learning_rate": 1.8978229623185185e-05,
|
| 1303 |
+
"loss": 1.245,
|
| 1304 |
+
"step": 2160
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 0.48,
|
| 1308 |
+
"learning_rate": 1.886750982696925e-05,
|
| 1309 |
+
"loss": 1.2462,
|
| 1310 |
+
"step": 2170
|
| 1311 |
+
},
|
| 1312 |
+
{
|
| 1313 |
+
"epoch": 0.48,
|
| 1314 |
+
"learning_rate": 1.8756563843469796e-05,
|
| 1315 |
+
"loss": 1.2165,
|
| 1316 |
+
"step": 2180
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 0.48,
|
| 1320 |
+
"learning_rate": 1.8645398161247095e-05,
|
| 1321 |
+
"loss": 1.241,
|
| 1322 |
+
"step": 2190
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 0.48,
|
| 1326 |
+
"learning_rate": 1.853401928171028e-05,
|
| 1327 |
+
"loss": 1.254,
|
| 1328 |
+
"step": 2200
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 0.48,
|
| 1332 |
+
"learning_rate": 1.8422433718737094e-05,
|
| 1333 |
+
"loss": 1.2337,
|
| 1334 |
+
"step": 2210
|
| 1335 |
+
},
|
| 1336 |
+
{
|
| 1337 |
+
"epoch": 0.49,
|
| 1338 |
+
"learning_rate": 1.8310647998292953e-05,
|
| 1339 |
+
"loss": 1.2657,
|
| 1340 |
+
"step": 2220
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 0.49,
|
| 1344 |
+
"learning_rate": 1.8198668658049263e-05,
|
| 1345 |
+
"loss": 1.2359,
|
| 1346 |
+
"step": 2230
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 0.49,
|
| 1350 |
+
"learning_rate": 1.808650224700111e-05,
|
| 1351 |
+
"loss": 1.2349,
|
| 1352 |
+
"step": 2240
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"epoch": 0.49,
|
| 1356 |
+
"learning_rate": 1.7974155325084183e-05,
|
| 1357 |
+
"loss": 1.2116,
|
| 1358 |
+
"step": 2250
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 0.49,
|
| 1362 |
+
"learning_rate": 1.786163446279119e-05,
|
| 1363 |
+
"loss": 1.2823,
|
| 1364 |
+
"step": 2260
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 0.5,
|
| 1368 |
+
"learning_rate": 1.7748946240787555e-05,
|
| 1369 |
+
"loss": 1.236,
|
| 1370 |
+
"step": 2270
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 0.5,
|
| 1374 |
+
"learning_rate": 1.7636097249526556e-05,
|
| 1375 |
+
"loss": 1.2378,
|
| 1376 |
+
"step": 2280
|
| 1377 |
+
},
|
| 1378 |
+
{
|
| 1379 |
+
"epoch": 0.5,
|
| 1380 |
+
"learning_rate": 1.7523094088863885e-05,
|
| 1381 |
+
"loss": 1.2696,
|
| 1382 |
+
"step": 2290
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 0.5,
|
| 1386 |
+
"learning_rate": 1.7409943367671685e-05,
|
| 1387 |
+
"loss": 1.1972,
|
| 1388 |
+
"step": 2300
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 0.51,
|
| 1392 |
+
"learning_rate": 1.7296651703452022e-05,
|
| 1393 |
+
"loss": 1.235,
|
| 1394 |
+
"step": 2310
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 0.51,
|
| 1398 |
+
"learning_rate": 1.7183225721949857e-05,
|
| 1399 |
+
"loss": 1.252,
|
| 1400 |
+
"step": 2320
|
| 1401 |
+
},
|
| 1402 |
+
{
|
| 1403 |
+
"epoch": 0.51,
|
| 1404 |
+
"learning_rate": 1.7069672056765573e-05,
|
| 1405 |
+
"loss": 1.2441,
|
| 1406 |
+
"step": 2330
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 0.51,
|
| 1410 |
+
"learning_rate": 1.6955997348966986e-05,
|
| 1411 |
+
"loss": 1.2437,
|
| 1412 |
+
"step": 2340
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 0.51,
|
| 1416 |
+
"learning_rate": 1.6842208246700966e-05,
|
| 1417 |
+
"loss": 1.2579,
|
| 1418 |
+
"step": 2350
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 0.52,
|
| 1422 |
+
"learning_rate": 1.6728311404804632e-05,
|
| 1423 |
+
"loss": 1.1966,
|
| 1424 |
+
"step": 2360
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 0.52,
|
| 1428 |
+
"learning_rate": 1.661431348441614e-05,
|
| 1429 |
+
"loss": 1.2409,
|
| 1430 |
+
"step": 2370
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 0.52,
|
| 1434 |
+
"learning_rate": 1.6500221152585114e-05,
|
| 1435 |
+
"loss": 1.1716,
|
| 1436 |
+
"step": 2380
|
| 1437 |
+
},
|
| 1438 |
+
{
|
| 1439 |
+
"epoch": 0.52,
|
| 1440 |
+
"learning_rate": 1.6386041081882727e-05,
|
| 1441 |
+
"loss": 1.3371,
|
| 1442 |
+
"step": 2390
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 0.53,
|
| 1446 |
+
"learning_rate": 1.627177995001148e-05,
|
| 1447 |
+
"loss": 1.2638,
|
| 1448 |
+
"step": 2400
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 0.53,
|
| 1452 |
+
"learning_rate": 1.615744443941464e-05,
|
| 1453 |
+
"loss": 1.2183,
|
| 1454 |
+
"step": 2410
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 0.53,
|
| 1458 |
+
"learning_rate": 1.6043041236885457e-05,
|
| 1459 |
+
"loss": 1.2175,
|
| 1460 |
+
"step": 2420
|
| 1461 |
+
},
|
| 1462 |
+
{
|
| 1463 |
+
"epoch": 0.53,
|
| 1464 |
+
"learning_rate": 1.5928577033176043e-05,
|
| 1465 |
+
"loss": 1.25,
|
| 1466 |
+
"step": 2430
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 0.53,
|
| 1470 |
+
"learning_rate": 1.5814058522606132e-05,
|
| 1471 |
+
"loss": 1.244,
|
| 1472 |
+
"step": 2440
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 0.54,
|
| 1476 |
+
"learning_rate": 1.569949240267151e-05,
|
| 1477 |
+
"loss": 1.1961,
|
| 1478 |
+
"step": 2450
|
| 1479 |
+
},
|
| 1480 |
+
{
|
| 1481 |
+
"epoch": 0.54,
|
| 1482 |
+
"learning_rate": 1.558488537365237e-05,
|
| 1483 |
+
"loss": 1.2389,
|
| 1484 |
+
"step": 2460
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 0.54,
|
| 1488 |
+
"learning_rate": 1.5470244138221415e-05,
|
| 1489 |
+
"loss": 1.2437,
|
| 1490 |
+
"step": 2470
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"epoch": 0.54,
|
| 1494 |
+
"learning_rate": 1.535557540105188e-05,
|
| 1495 |
+
"loss": 1.2218,
|
| 1496 |
+
"step": 2480
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 0.55,
|
| 1500 |
+
"learning_rate": 1.5240885868425409e-05,
|
| 1501 |
+
"loss": 1.2585,
|
| 1502 |
+
"step": 2490
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 0.55,
|
| 1506 |
+
"learning_rate": 1.5126182247839853e-05,
|
| 1507 |
+
"loss": 1.2138,
|
| 1508 |
+
"step": 2500
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 0.55,
|
| 1512 |
+
"learning_rate": 1.5011471247616968e-05,
|
| 1513 |
+
"loss": 1.2291,
|
| 1514 |
+
"step": 2510
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 0.55,
|
| 1518 |
+
"learning_rate": 1.4896759576510118e-05,
|
| 1519 |
+
"loss": 1.246,
|
| 1520 |
+
"step": 2520
|
| 1521 |
+
},
|
| 1522 |
+
{
|
| 1523 |
+
"epoch": 0.55,
|
| 1524 |
+
"learning_rate": 1.4782053943311894e-05,
|
| 1525 |
+
"loss": 1.2412,
|
| 1526 |
+
"step": 2530
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 0.56,
|
| 1530 |
+
"learning_rate": 1.4667361056461766e-05,
|
| 1531 |
+
"loss": 1.2486,
|
| 1532 |
+
"step": 2540
|
| 1533 |
+
},
|
| 1534 |
+
{
|
| 1535 |
+
"epoch": 0.56,
|
| 1536 |
+
"learning_rate": 1.4552687623653744e-05,
|
| 1537 |
+
"loss": 1.2382,
|
| 1538 |
+
"step": 2550
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 0.56,
|
| 1542 |
+
"learning_rate": 1.4438040351444098e-05,
|
| 1543 |
+
"loss": 1.1778,
|
| 1544 |
+
"step": 2560
|
| 1545 |
+
},
|
| 1546 |
+
{
|
| 1547 |
+
"epoch": 0.56,
|
| 1548 |
+
"learning_rate": 1.4323425944859116e-05,
|
| 1549 |
+
"loss": 1.2197,
|
| 1550 |
+
"step": 2570
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 0.57,
|
| 1554 |
+
"learning_rate": 1.4208851107002975e-05,
|
| 1555 |
+
"loss": 1.2182,
|
| 1556 |
+
"step": 2580
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 0.57,
|
| 1560 |
+
"learning_rate": 1.4094322538665707e-05,
|
| 1561 |
+
"loss": 1.2052,
|
| 1562 |
+
"step": 2590
|
| 1563 |
+
},
|
| 1564 |
+
{
|
| 1565 |
+
"epoch": 0.57,
|
| 1566 |
+
"learning_rate": 1.3979846937931348e-05,
|
| 1567 |
+
"loss": 1.1819,
|
| 1568 |
+
"step": 2600
|
| 1569 |
+
},
|
| 1570 |
+
{
|
| 1571 |
+
"epoch": 0.57,
|
| 1572 |
+
"learning_rate": 1.3865430999786147e-05,
|
| 1573 |
+
"loss": 1.1735,
|
| 1574 |
+
"step": 2610
|
| 1575 |
+
},
|
| 1576 |
+
{
|
| 1577 |
+
"epoch": 0.57,
|
| 1578 |
+
"learning_rate": 1.3751081415727068e-05,
|
| 1579 |
+
"loss": 1.2296,
|
| 1580 |
+
"step": 2620
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 0.58,
|
| 1584 |
+
"learning_rate": 1.3636804873370409e-05,
|
| 1585 |
+
"loss": 1.2094,
|
| 1586 |
+
"step": 2630
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 0.58,
|
| 1590 |
+
"learning_rate": 1.3522608056060722e-05,
|
| 1591 |
+
"loss": 1.2369,
|
| 1592 |
+
"step": 2640
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 0.58,
|
| 1596 |
+
"learning_rate": 1.3408497642479905e-05,
|
| 1597 |
+
"loss": 1.2653,
|
| 1598 |
+
"step": 2650
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 0.58,
|
| 1602 |
+
"learning_rate": 1.3294480306256629e-05,
|
| 1603 |
+
"loss": 1.1777,
|
| 1604 |
+
"step": 2660
|
| 1605 |
+
},
|
| 1606 |
+
{
|
| 1607 |
+
"epoch": 0.58,
|
| 1608 |
+
"learning_rate": 1.3180562715576029e-05,
|
| 1609 |
+
"loss": 1.2748,
|
| 1610 |
+
"step": 2670
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 0.59,
|
| 1614 |
+
"learning_rate": 1.306675153278974e-05,
|
| 1615 |
+
"loss": 1.2337,
|
| 1616 |
+
"step": 2680
|
| 1617 |
+
},
|
| 1618 |
+
{
|
| 1619 |
+
"epoch": 0.59,
|
| 1620 |
+
"learning_rate": 1.295305341402623e-05,
|
| 1621 |
+
"loss": 1.1837,
|
| 1622 |
+
"step": 2690
|
| 1623 |
+
},
|
| 1624 |
+
{
|
| 1625 |
+
"epoch": 0.59,
|
| 1626 |
+
"learning_rate": 1.2839475008801526e-05,
|
| 1627 |
+
"loss": 1.1723,
|
| 1628 |
+
"step": 2700
|
| 1629 |
+
},
|
| 1630 |
+
{
|
| 1631 |
+
"epoch": 0.59,
|
| 1632 |
+
"learning_rate": 1.2726022959630363e-05,
|
| 1633 |
+
"loss": 1.1999,
|
| 1634 |
+
"step": 2710
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 0.6,
|
| 1638 |
+
"learning_rate": 1.2612703901637648e-05,
|
| 1639 |
+
"loss": 1.245,
|
| 1640 |
+
"step": 2720
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 0.6,
|
| 1644 |
+
"learning_rate": 1.2499524462170441e-05,
|
| 1645 |
+
"loss": 1.2375,
|
| 1646 |
+
"step": 2730
|
| 1647 |
+
},
|
| 1648 |
+
{
|
| 1649 |
+
"epoch": 0.6,
|
| 1650 |
+
"learning_rate": 1.2386491260410369e-05,
|
| 1651 |
+
"loss": 1.2432,
|
| 1652 |
+
"step": 2740
|
| 1653 |
+
},
|
| 1654 |
+
{
|
| 1655 |
+
"epoch": 0.6,
|
| 1656 |
+
"learning_rate": 1.227361090698649e-05,
|
| 1657 |
+
"loss": 1.1958,
|
| 1658 |
+
"step": 2750
|
| 1659 |
+
},
|
| 1660 |
+
{
|
| 1661 |
+
"epoch": 0.6,
|
| 1662 |
+
"learning_rate": 1.2160890003588686e-05,
|
| 1663 |
+
"loss": 1.2432,
|
| 1664 |
+
"step": 2760
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 0.61,
|
| 1668 |
+
"learning_rate": 1.2048335142581578e-05,
|
| 1669 |
+
"loss": 1.1742,
|
| 1670 |
+
"step": 2770
|
| 1671 |
+
},
|
| 1672 |
+
{
|
| 1673 |
+
"epoch": 0.61,
|
| 1674 |
+
"learning_rate": 1.1935952906618955e-05,
|
| 1675 |
+
"loss": 1.2097,
|
| 1676 |
+
"step": 2780
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 0.61,
|
| 1680 |
+
"learning_rate": 1.1823749868258836e-05,
|
| 1681 |
+
"loss": 1.2993,
|
| 1682 |
+
"step": 2790
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 0.61,
|
| 1686 |
+
"learning_rate": 1.1711732589579032e-05,
|
| 1687 |
+
"loss": 1.1788,
|
| 1688 |
+
"step": 2800
|
| 1689 |
+
},
|
| 1690 |
+
{
|
| 1691 |
+
"epoch": 0.62,
|
| 1692 |
+
"learning_rate": 1.1599907621793398e-05,
|
| 1693 |
+
"loss": 1.2499,
|
| 1694 |
+
"step": 2810
|
| 1695 |
+
},
|
| 1696 |
+
{
|
| 1697 |
+
"epoch": 0.62,
|
| 1698 |
+
"learning_rate": 1.1488281504868695e-05,
|
| 1699 |
+
"loss": 1.2777,
|
| 1700 |
+
"step": 2820
|
| 1701 |
+
},
|
| 1702 |
+
{
|
| 1703 |
+
"epoch": 0.62,
|
| 1704 |
+
"learning_rate": 1.137686076714209e-05,
|
| 1705 |
+
"loss": 1.2305,
|
| 1706 |
+
"step": 2830
|
| 1707 |
+
},
|
| 1708 |
+
{
|
| 1709 |
+
"epoch": 0.62,
|
| 1710 |
+
"learning_rate": 1.1265651924939364e-05,
|
| 1711 |
+
"loss": 1.2181,
|
| 1712 |
+
"step": 2840
|
| 1713 |
+
},
|
| 1714 |
+
{
|
| 1715 |
+
"epoch": 0.62,
|
| 1716 |
+
"learning_rate": 1.11546614821938e-05,
|
| 1717 |
+
"loss": 1.2418,
|
| 1718 |
+
"step": 2850
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 0.63,
|
| 1722 |
+
"learning_rate": 1.1043895930065843e-05,
|
| 1723 |
+
"loss": 1.2372,
|
| 1724 |
+
"step": 2860
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 0.63,
|
| 1728 |
+
"learning_rate": 1.0933361746563418e-05,
|
| 1729 |
+
"loss": 1.2151,
|
| 1730 |
+
"step": 2870
|
| 1731 |
+
},
|
| 1732 |
+
{
|
| 1733 |
+
"epoch": 0.63,
|
| 1734 |
+
"learning_rate": 1.0823065396163104e-05,
|
| 1735 |
+
"loss": 1.1852,
|
| 1736 |
+
"step": 2880
|
| 1737 |
+
},
|
| 1738 |
+
{
|
| 1739 |
+
"epoch": 0.63,
|
| 1740 |
+
"learning_rate": 1.071301332943206e-05,
|
| 1741 |
+
"loss": 1.1956,
|
| 1742 |
+
"step": 2890
|
| 1743 |
+
},
|
| 1744 |
+
{
|
| 1745 |
+
"epoch": 0.64,
|
| 1746 |
+
"learning_rate": 1.0603211982650775e-05,
|
| 1747 |
+
"loss": 1.2443,
|
| 1748 |
+
"step": 2900
|
| 1749 |
+
},
|
| 1750 |
+
{
|
| 1751 |
+
"epoch": 0.64,
|
| 1752 |
+
"learning_rate": 1.0493667777436633e-05,
|
| 1753 |
+
"loss": 1.1962,
|
| 1754 |
+
"step": 2910
|
| 1755 |
+
},
|
| 1756 |
+
{
|
| 1757 |
+
"epoch": 0.64,
|
| 1758 |
+
"learning_rate": 1.0384387120368367e-05,
|
| 1759 |
+
"loss": 1.2033,
|
| 1760 |
+
"step": 2920
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 0.64,
|
| 1764 |
+
"learning_rate": 1.0275376402611364e-05,
|
| 1765 |
+
"loss": 1.279,
|
| 1766 |
+
"step": 2930
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 0.64,
|
| 1770 |
+
"learning_rate": 1.0166641999543912e-05,
|
| 1771 |
+
"loss": 1.2342,
|
| 1772 |
+
"step": 2940
|
| 1773 |
+
},
|
| 1774 |
+
{
|
| 1775 |
+
"epoch": 0.65,
|
| 1776 |
+
"learning_rate": 1.0058190270384298e-05,
|
| 1777 |
+
"loss": 1.2511,
|
| 1778 |
+
"step": 2950
|
| 1779 |
+
},
|
| 1780 |
+
{
|
| 1781 |
+
"epoch": 0.65,
|
| 1782 |
+
"learning_rate": 9.950027557818926e-06,
|
| 1783 |
+
"loss": 1.2051,
|
| 1784 |
+
"step": 2960
|
| 1785 |
+
},
|
| 1786 |
+
{
|
| 1787 |
+
"epoch": 0.65,
|
| 1788 |
+
"learning_rate": 9.84216018763138e-06,
|
| 1789 |
+
"loss": 1.1881,
|
| 1790 |
+
"step": 2970
|
| 1791 |
+
},
|
| 1792 |
+
{
|
| 1793 |
+
"epoch": 0.65,
|
| 1794 |
+
"learning_rate": 9.734594468332436e-06,
|
| 1795 |
+
"loss": 1.2033,
|
| 1796 |
+
"step": 2980
|
| 1797 |
+
},
|
| 1798 |
+
{
|
| 1799 |
+
"epoch": 0.65,
|
| 1800 |
+
"learning_rate": 9.627336690791142e-06,
|
| 1801 |
+
"loss": 1.2522,
|
| 1802 |
+
"step": 2990
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 0.66,
|
| 1806 |
+
"learning_rate": 9.520393127866881e-06,
|
| 1807 |
+
"loss": 1.1684,
|
| 1808 |
+
"step": 3000
|
| 1809 |
+
}
|
| 1810 |
+
],
|
| 1811 |
+
"max_steps": 4565,
|
| 1812 |
+
"num_train_epochs": 1,
|
| 1813 |
+
"total_flos": 0.0,
|
| 1814 |
+
"trial_name": null,
|
| 1815 |
+
"trial_params": null
|
| 1816 |
+
}
|
msa_curriculum/reset/pct50/finetuning/bytes/MSA_Dialect_Combined_1x/checkpoint-4000/trainer_state.json
ADDED
|
@@ -0,0 +1,2416 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.8760882658927887,
|
| 5 |
+
"global_step": 4000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.0,
|
| 12 |
+
"learning_rate": 6.564551422319475e-07,
|
| 13 |
+
"loss": 1.7654,
|
| 14 |
+
"step": 10
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.0,
|
| 18 |
+
"learning_rate": 1.312910284463895e-06,
|
| 19 |
+
"loss": 1.7226,
|
| 20 |
+
"step": 20
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.01,
|
| 24 |
+
"learning_rate": 1.9693654266958425e-06,
|
| 25 |
+
"loss": 1.7008,
|
| 26 |
+
"step": 30
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.01,
|
| 30 |
+
"learning_rate": 2.62582056892779e-06,
|
| 31 |
+
"loss": 1.6721,
|
| 32 |
+
"step": 40
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.01,
|
| 36 |
+
"learning_rate": 3.2822757111597377e-06,
|
| 37 |
+
"loss": 1.7142,
|
| 38 |
+
"step": 50
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.01,
|
| 42 |
+
"learning_rate": 3.938730853391685e-06,
|
| 43 |
+
"loss": 1.7535,
|
| 44 |
+
"step": 60
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.02,
|
| 48 |
+
"learning_rate": 4.595185995623632e-06,
|
| 49 |
+
"loss": 1.6915,
|
| 50 |
+
"step": 70
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.02,
|
| 54 |
+
"learning_rate": 5.25164113785558e-06,
|
| 55 |
+
"loss": 1.6418,
|
| 56 |
+
"step": 80
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.02,
|
| 60 |
+
"learning_rate": 5.908096280087528e-06,
|
| 61 |
+
"loss": 1.6122,
|
| 62 |
+
"step": 90
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.02,
|
| 66 |
+
"learning_rate": 6.564551422319475e-06,
|
| 67 |
+
"loss": 1.6245,
|
| 68 |
+
"step": 100
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.02,
|
| 72 |
+
"learning_rate": 7.221006564551422e-06,
|
| 73 |
+
"loss": 1.6182,
|
| 74 |
+
"step": 110
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.03,
|
| 78 |
+
"learning_rate": 7.87746170678337e-06,
|
| 79 |
+
"loss": 1.4952,
|
| 80 |
+
"step": 120
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.03,
|
| 84 |
+
"learning_rate": 8.533916849015316e-06,
|
| 85 |
+
"loss": 1.6204,
|
| 86 |
+
"step": 130
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.03,
|
| 90 |
+
"learning_rate": 9.190371991247264e-06,
|
| 91 |
+
"loss": 1.5313,
|
| 92 |
+
"step": 140
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.03,
|
| 96 |
+
"learning_rate": 9.846827133479212e-06,
|
| 97 |
+
"loss": 1.5215,
|
| 98 |
+
"step": 150
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.04,
|
| 102 |
+
"learning_rate": 1.050328227571116e-05,
|
| 103 |
+
"loss": 1.5439,
|
| 104 |
+
"step": 160
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.04,
|
| 108 |
+
"learning_rate": 1.1159737417943107e-05,
|
| 109 |
+
"loss": 1.5977,
|
| 110 |
+
"step": 170
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.04,
|
| 114 |
+
"learning_rate": 1.1816192560175055e-05,
|
| 115 |
+
"loss": 1.5625,
|
| 116 |
+
"step": 180
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.04,
|
| 120 |
+
"learning_rate": 1.2472647702407003e-05,
|
| 121 |
+
"loss": 1.5324,
|
| 122 |
+
"step": 190
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.04,
|
| 126 |
+
"learning_rate": 1.312910284463895e-05,
|
| 127 |
+
"loss": 1.479,
|
| 128 |
+
"step": 200
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.05,
|
| 132 |
+
"learning_rate": 1.3785557986870897e-05,
|
| 133 |
+
"loss": 1.4969,
|
| 134 |
+
"step": 210
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.05,
|
| 138 |
+
"learning_rate": 1.4442013129102845e-05,
|
| 139 |
+
"loss": 1.5098,
|
| 140 |
+
"step": 220
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.05,
|
| 144 |
+
"learning_rate": 1.509846827133479e-05,
|
| 145 |
+
"loss": 1.5546,
|
| 146 |
+
"step": 230
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.05,
|
| 150 |
+
"learning_rate": 1.575492341356674e-05,
|
| 151 |
+
"loss": 1.5069,
|
| 152 |
+
"step": 240
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.05,
|
| 156 |
+
"learning_rate": 1.6411378555798686e-05,
|
| 157 |
+
"loss": 1.493,
|
| 158 |
+
"step": 250
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.06,
|
| 162 |
+
"learning_rate": 1.7067833698030632e-05,
|
| 163 |
+
"loss": 1.4697,
|
| 164 |
+
"step": 260
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.06,
|
| 168 |
+
"learning_rate": 1.7724288840262582e-05,
|
| 169 |
+
"loss": 1.4476,
|
| 170 |
+
"step": 270
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.06,
|
| 174 |
+
"learning_rate": 1.8380743982494528e-05,
|
| 175 |
+
"loss": 1.4063,
|
| 176 |
+
"step": 280
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.06,
|
| 180 |
+
"learning_rate": 1.9037199124726478e-05,
|
| 181 |
+
"loss": 1.4337,
|
| 182 |
+
"step": 290
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.07,
|
| 186 |
+
"learning_rate": 1.9693654266958424e-05,
|
| 187 |
+
"loss": 1.4452,
|
| 188 |
+
"step": 300
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.07,
|
| 192 |
+
"learning_rate": 2.0350109409190373e-05,
|
| 193 |
+
"loss": 1.4853,
|
| 194 |
+
"step": 310
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.07,
|
| 198 |
+
"learning_rate": 2.100656455142232e-05,
|
| 199 |
+
"loss": 1.4319,
|
| 200 |
+
"step": 320
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.07,
|
| 204 |
+
"learning_rate": 2.1663019693654265e-05,
|
| 205 |
+
"loss": 1.4208,
|
| 206 |
+
"step": 330
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.07,
|
| 210 |
+
"learning_rate": 2.2319474835886215e-05,
|
| 211 |
+
"loss": 1.4323,
|
| 212 |
+
"step": 340
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.08,
|
| 216 |
+
"learning_rate": 2.297592997811816e-05,
|
| 217 |
+
"loss": 1.4242,
|
| 218 |
+
"step": 350
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.08,
|
| 222 |
+
"learning_rate": 2.363238512035011e-05,
|
| 223 |
+
"loss": 1.4216,
|
| 224 |
+
"step": 360
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.08,
|
| 228 |
+
"learning_rate": 2.4288840262582056e-05,
|
| 229 |
+
"loss": 1.3798,
|
| 230 |
+
"step": 370
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.08,
|
| 234 |
+
"learning_rate": 2.4945295404814006e-05,
|
| 235 |
+
"loss": 1.4481,
|
| 236 |
+
"step": 380
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.09,
|
| 240 |
+
"learning_rate": 2.5601750547045952e-05,
|
| 241 |
+
"loss": 1.448,
|
| 242 |
+
"step": 390
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.09,
|
| 246 |
+
"learning_rate": 2.62582056892779e-05,
|
| 247 |
+
"loss": 1.3897,
|
| 248 |
+
"step": 400
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.09,
|
| 252 |
+
"learning_rate": 2.6914660831509848e-05,
|
| 253 |
+
"loss": 1.4447,
|
| 254 |
+
"step": 410
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.09,
|
| 258 |
+
"learning_rate": 2.7571115973741794e-05,
|
| 259 |
+
"loss": 1.3803,
|
| 260 |
+
"step": 420
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 0.09,
|
| 264 |
+
"learning_rate": 2.8227571115973743e-05,
|
| 265 |
+
"loss": 1.414,
|
| 266 |
+
"step": 430
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.1,
|
| 270 |
+
"learning_rate": 2.888402625820569e-05,
|
| 271 |
+
"loss": 1.3641,
|
| 272 |
+
"step": 440
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.1,
|
| 276 |
+
"learning_rate": 2.954048140043764e-05,
|
| 277 |
+
"loss": 1.3955,
|
| 278 |
+
"step": 450
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 0.1,
|
| 282 |
+
"learning_rate": 2.9999960523153055e-05,
|
| 283 |
+
"loss": 1.4183,
|
| 284 |
+
"step": 460
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 0.1,
|
| 288 |
+
"learning_rate": 2.999925871832113e-05,
|
| 289 |
+
"loss": 1.3418,
|
| 290 |
+
"step": 470
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.11,
|
| 294 |
+
"learning_rate": 2.9997679697467503e-05,
|
| 295 |
+
"loss": 1.4104,
|
| 296 |
+
"step": 480
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.11,
|
| 300 |
+
"learning_rate": 2.9995223552939574e-05,
|
| 301 |
+
"loss": 1.3313,
|
| 302 |
+
"step": 490
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 0.11,
|
| 306 |
+
"learning_rate": 2.9991890428382386e-05,
|
| 307 |
+
"loss": 1.311,
|
| 308 |
+
"step": 500
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 0.11,
|
| 312 |
+
"learning_rate": 2.9987680518730264e-05,
|
| 313 |
+
"loss": 1.3834,
|
| 314 |
+
"step": 510
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 0.11,
|
| 318 |
+
"learning_rate": 2.998259407019538e-05,
|
| 319 |
+
"loss": 1.3476,
|
| 320 |
+
"step": 520
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 0.12,
|
| 324 |
+
"learning_rate": 2.997663138025339e-05,
|
| 325 |
+
"loss": 1.3777,
|
| 326 |
+
"step": 530
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 0.12,
|
| 330 |
+
"learning_rate": 2.9969792797625994e-05,
|
| 331 |
+
"loss": 1.3502,
|
| 332 |
+
"step": 540
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.12,
|
| 336 |
+
"learning_rate": 2.9962078722260575e-05,
|
| 337 |
+
"loss": 1.3632,
|
| 338 |
+
"step": 550
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.12,
|
| 342 |
+
"learning_rate": 2.99534896053068e-05,
|
| 343 |
+
"loss": 1.4011,
|
| 344 |
+
"step": 560
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 0.12,
|
| 348 |
+
"learning_rate": 2.9944025949090224e-05,
|
| 349 |
+
"loss": 1.3331,
|
| 350 |
+
"step": 570
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 0.13,
|
| 354 |
+
"learning_rate": 2.9933688307082926e-05,
|
| 355 |
+
"loss": 1.2906,
|
| 356 |
+
"step": 580
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.13,
|
| 360 |
+
"learning_rate": 2.9922477283871123e-05,
|
| 361 |
+
"loss": 1.3344,
|
| 362 |
+
"step": 590
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 0.13,
|
| 366 |
+
"learning_rate": 2.991039353511983e-05,
|
| 367 |
+
"loss": 1.3746,
|
| 368 |
+
"step": 600
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 0.13,
|
| 372 |
+
"learning_rate": 2.9897437767534515e-05,
|
| 373 |
+
"loss": 1.3082,
|
| 374 |
+
"step": 610
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.14,
|
| 378 |
+
"learning_rate": 2.988361073881974e-05,
|
| 379 |
+
"loss": 1.3623,
|
| 380 |
+
"step": 620
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.14,
|
| 384 |
+
"learning_rate": 2.9868913257634876e-05,
|
| 385 |
+
"loss": 1.2997,
|
| 386 |
+
"step": 630
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 0.14,
|
| 390 |
+
"learning_rate": 2.985334618354681e-05,
|
| 391 |
+
"loss": 1.2911,
|
| 392 |
+
"step": 640
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 0.14,
|
| 396 |
+
"learning_rate": 2.9836910426979645e-05,
|
| 397 |
+
"loss": 1.3226,
|
| 398 |
+
"step": 650
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 0.14,
|
| 402 |
+
"learning_rate": 2.981960694916148e-05,
|
| 403 |
+
"loss": 1.3156,
|
| 404 |
+
"step": 660
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 0.15,
|
| 408 |
+
"learning_rate": 2.98014367620682e-05,
|
| 409 |
+
"loss": 1.3288,
|
| 410 |
+
"step": 670
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 0.15,
|
| 414 |
+
"learning_rate": 2.9782400928364266e-05,
|
| 415 |
+
"loss": 1.3732,
|
| 416 |
+
"step": 680
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.15,
|
| 420 |
+
"learning_rate": 2.976250056134058e-05,
|
| 421 |
+
"loss": 1.368,
|
| 422 |
+
"step": 690
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.15,
|
| 426 |
+
"learning_rate": 2.9741736824849383e-05,
|
| 427 |
+
"loss": 1.3157,
|
| 428 |
+
"step": 700
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 0.16,
|
| 432 |
+
"learning_rate": 2.972011093323617e-05,
|
| 433 |
+
"loss": 1.3349,
|
| 434 |
+
"step": 710
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 0.16,
|
| 438 |
+
"learning_rate": 2.9697624151268684e-05,
|
| 439 |
+
"loss": 1.3584,
|
| 440 |
+
"step": 720
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 0.16,
|
| 444 |
+
"learning_rate": 2.967427779406295e-05,
|
| 445 |
+
"loss": 1.3826,
|
| 446 |
+
"step": 730
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 0.16,
|
| 450 |
+
"learning_rate": 2.965007322700634e-05,
|
| 451 |
+
"loss": 1.2787,
|
| 452 |
+
"step": 740
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 0.16,
|
| 456 |
+
"learning_rate": 2.9625011865677757e-05,
|
| 457 |
+
"loss": 1.3617,
|
| 458 |
+
"step": 750
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 0.17,
|
| 462 |
+
"learning_rate": 2.95990951757648e-05,
|
| 463 |
+
"loss": 1.3281,
|
| 464 |
+
"step": 760
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.17,
|
| 468 |
+
"learning_rate": 2.9572324672978096e-05,
|
| 469 |
+
"loss": 1.3359,
|
| 470 |
+
"step": 770
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 0.17,
|
| 474 |
+
"learning_rate": 2.9544701922962598e-05,
|
| 475 |
+
"loss": 1.3062,
|
| 476 |
+
"step": 780
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 0.17,
|
| 480 |
+
"learning_rate": 2.951622854120608e-05,
|
| 481 |
+
"loss": 1.3376,
|
| 482 |
+
"step": 790
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 0.18,
|
| 486 |
+
"learning_rate": 2.948690619294461e-05,
|
| 487 |
+
"loss": 1.3341,
|
| 488 |
+
"step": 800
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 0.18,
|
| 492 |
+
"learning_rate": 2.9456736593065172e-05,
|
| 493 |
+
"loss": 1.3207,
|
| 494 |
+
"step": 810
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 0.18,
|
| 498 |
+
"learning_rate": 2.9425721506005403e-05,
|
| 499 |
+
"loss": 1.2939,
|
| 500 |
+
"step": 820
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 0.18,
|
| 504 |
+
"learning_rate": 2.9393862745650353e-05,
|
| 505 |
+
"loss": 1.3064,
|
| 506 |
+
"step": 830
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.18,
|
| 510 |
+
"learning_rate": 2.9361162175226433e-05,
|
| 511 |
+
"loss": 1.3241,
|
| 512 |
+
"step": 840
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 0.19,
|
| 516 |
+
"learning_rate": 2.932762170719245e-05,
|
| 517 |
+
"loss": 1.3197,
|
| 518 |
+
"step": 850
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 0.19,
|
| 522 |
+
"learning_rate": 2.929324330312773e-05,
|
| 523 |
+
"loss": 1.3123,
|
| 524 |
+
"step": 860
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 0.19,
|
| 528 |
+
"learning_rate": 2.925802897361742e-05,
|
| 529 |
+
"loss": 1.3165,
|
| 530 |
+
"step": 870
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 0.19,
|
| 534 |
+
"learning_rate": 2.9221980778134912e-05,
|
| 535 |
+
"loss": 1.2716,
|
| 536 |
+
"step": 880
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 0.19,
|
| 540 |
+
"learning_rate": 2.9185100824921362e-05,
|
| 541 |
+
"loss": 1.3049,
|
| 542 |
+
"step": 890
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 0.2,
|
| 546 |
+
"learning_rate": 2.9147391270862414e-05,
|
| 547 |
+
"loss": 1.3517,
|
| 548 |
+
"step": 900
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.2,
|
| 552 |
+
"learning_rate": 2.910885432136206e-05,
|
| 553 |
+
"loss": 1.2833,
|
| 554 |
+
"step": 910
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 0.2,
|
| 558 |
+
"learning_rate": 2.9069492230213643e-05,
|
| 559 |
+
"loss": 1.2964,
|
| 560 |
+
"step": 920
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 0.2,
|
| 564 |
+
"learning_rate": 2.9029307299468067e-05,
|
| 565 |
+
"loss": 1.32,
|
| 566 |
+
"step": 930
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 0.21,
|
| 570 |
+
"learning_rate": 2.8988301879299137e-05,
|
| 571 |
+
"loss": 1.2736,
|
| 572 |
+
"step": 940
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 0.21,
|
| 576 |
+
"learning_rate": 2.894647836786615e-05,
|
| 577 |
+
"loss": 1.2839,
|
| 578 |
+
"step": 950
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 0.21,
|
| 582 |
+
"learning_rate": 2.89038392111736e-05,
|
| 583 |
+
"loss": 1.3297,
|
| 584 |
+
"step": 960
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 0.21,
|
| 588 |
+
"learning_rate": 2.8860386902928147e-05,
|
| 589 |
+
"loss": 1.3177,
|
| 590 |
+
"step": 970
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.21,
|
| 594 |
+
"learning_rate": 2.8816123984392777e-05,
|
| 595 |
+
"loss": 1.3007,
|
| 596 |
+
"step": 980
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 0.22,
|
| 600 |
+
"learning_rate": 2.8771053044238174e-05,
|
| 601 |
+
"loss": 1.3428,
|
| 602 |
+
"step": 990
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 0.22,
|
| 606 |
+
"learning_rate": 2.872517671839132e-05,
|
| 607 |
+
"loss": 1.3025,
|
| 608 |
+
"step": 1000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 0.22,
|
| 612 |
+
"learning_rate": 2.8678497689881354e-05,
|
| 613 |
+
"loss": 1.3186,
|
| 614 |
+
"step": 1010
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 0.22,
|
| 618 |
+
"learning_rate": 2.863101868868262e-05,
|
| 619 |
+
"loss": 1.2949,
|
| 620 |
+
"step": 1020
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 0.23,
|
| 624 |
+
"learning_rate": 2.858274249155505e-05,
|
| 625 |
+
"loss": 1.2859,
|
| 626 |
+
"step": 1030
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 0.23,
|
| 630 |
+
"learning_rate": 2.853367192188175e-05,
|
| 631 |
+
"loss": 1.2447,
|
| 632 |
+
"step": 1040
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.23,
|
| 636 |
+
"learning_rate": 2.848380984950386e-05,
|
| 637 |
+
"loss": 1.34,
|
| 638 |
+
"step": 1050
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 0.23,
|
| 642 |
+
"learning_rate": 2.8433159190552752e-05,
|
| 643 |
+
"loss": 1.2706,
|
| 644 |
+
"step": 1060
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 0.23,
|
| 648 |
+
"learning_rate": 2.8381722907279456e-05,
|
| 649 |
+
"loss": 1.2778,
|
| 650 |
+
"step": 1070
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 0.24,
|
| 654 |
+
"learning_rate": 2.8329504007881413e-05,
|
| 655 |
+
"loss": 1.2487,
|
| 656 |
+
"step": 1080
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 0.24,
|
| 660 |
+
"learning_rate": 2.8276505546326554e-05,
|
| 661 |
+
"loss": 1.3054,
|
| 662 |
+
"step": 1090
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 0.24,
|
| 666 |
+
"learning_rate": 2.8222730622174704e-05,
|
| 667 |
+
"loss": 1.3033,
|
| 668 |
+
"step": 1100
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 0.24,
|
| 672 |
+
"learning_rate": 2.8168182380396286e-05,
|
| 673 |
+
"loss": 1.3092,
|
| 674 |
+
"step": 1110
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.25,
|
| 678 |
+
"learning_rate": 2.8112864011188392e-05,
|
| 679 |
+
"loss": 1.2893,
|
| 680 |
+
"step": 1120
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 0.25,
|
| 684 |
+
"learning_rate": 2.8056778749788224e-05,
|
| 685 |
+
"loss": 1.2901,
|
| 686 |
+
"step": 1130
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 0.25,
|
| 690 |
+
"learning_rate": 2.7999929876283874e-05,
|
| 691 |
+
"loss": 1.2615,
|
| 692 |
+
"step": 1140
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 0.25,
|
| 696 |
+
"learning_rate": 2.7942320715422486e-05,
|
| 697 |
+
"loss": 1.3097,
|
| 698 |
+
"step": 1150
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 0.25,
|
| 702 |
+
"learning_rate": 2.788395463641583e-05,
|
| 703 |
+
"loss": 1.2495,
|
| 704 |
+
"step": 1160
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 0.26,
|
| 708 |
+
"learning_rate": 2.7824835052743235e-05,
|
| 709 |
+
"loss": 1.315,
|
| 710 |
+
"step": 1170
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 0.26,
|
| 714 |
+
"learning_rate": 2.7764965421951963e-05,
|
| 715 |
+
"loss": 1.2946,
|
| 716 |
+
"step": 1180
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.26,
|
| 720 |
+
"learning_rate": 2.770434924545501e-05,
|
| 721 |
+
"loss": 1.3103,
|
| 722 |
+
"step": 1190
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 0.26,
|
| 726 |
+
"learning_rate": 2.764299006832631e-05,
|
| 727 |
+
"loss": 1.2656,
|
| 728 |
+
"step": 1200
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 0.27,
|
| 732 |
+
"learning_rate": 2.7580891479093433e-05,
|
| 733 |
+
"loss": 1.3298,
|
| 734 |
+
"step": 1210
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 0.27,
|
| 738 |
+
"learning_rate": 2.7518057109527675e-05,
|
| 739 |
+
"loss": 1.3019,
|
| 740 |
+
"step": 1220
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 0.27,
|
| 744 |
+
"learning_rate": 2.745449063443169e-05,
|
| 745 |
+
"loss": 1.2962,
|
| 746 |
+
"step": 1230
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 0.27,
|
| 750 |
+
"learning_rate": 2.7390195771424554e-05,
|
| 751 |
+
"loss": 1.3314,
|
| 752 |
+
"step": 1240
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 0.27,
|
| 756 |
+
"learning_rate": 2.7325176280724352e-05,
|
| 757 |
+
"loss": 1.2564,
|
| 758 |
+
"step": 1250
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.28,
|
| 762 |
+
"learning_rate": 2.725943596492828e-05,
|
| 763 |
+
"loss": 1.2918,
|
| 764 |
+
"step": 1260
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 0.28,
|
| 768 |
+
"learning_rate": 2.7192978668790217e-05,
|
| 769 |
+
"loss": 1.2837,
|
| 770 |
+
"step": 1270
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 0.28,
|
| 774 |
+
"learning_rate": 2.7125808278995908e-05,
|
| 775 |
+
"loss": 1.2939,
|
| 776 |
+
"step": 1280
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 0.28,
|
| 780 |
+
"learning_rate": 2.705792872393562e-05,
|
| 781 |
+
"loss": 1.2619,
|
| 782 |
+
"step": 1290
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 0.28,
|
| 786 |
+
"learning_rate": 2.6989343973474434e-05,
|
| 787 |
+
"loss": 1.278,
|
| 788 |
+
"step": 1300
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 0.29,
|
| 792 |
+
"learning_rate": 2.6920058038720026e-05,
|
| 793 |
+
"loss": 1.2826,
|
| 794 |
+
"step": 1310
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 0.29,
|
| 798 |
+
"learning_rate": 2.685007497178812e-05,
|
| 799 |
+
"loss": 1.2275,
|
| 800 |
+
"step": 1320
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.29,
|
| 804 |
+
"learning_rate": 2.6779398865565487e-05,
|
| 805 |
+
"loss": 1.3291,
|
| 806 |
+
"step": 1330
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 0.29,
|
| 810 |
+
"learning_rate": 2.670803385347058e-05,
|
| 811 |
+
"loss": 1.2972,
|
| 812 |
+
"step": 1340
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 0.3,
|
| 816 |
+
"learning_rate": 2.6635984109211786e-05,
|
| 817 |
+
"loss": 1.2323,
|
| 818 |
+
"step": 1350
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 0.3,
|
| 822 |
+
"learning_rate": 2.656325384654335e-05,
|
| 823 |
+
"loss": 1.2529,
|
| 824 |
+
"step": 1360
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 0.3,
|
| 828 |
+
"learning_rate": 2.648984731901892e-05,
|
| 829 |
+
"loss": 1.2411,
|
| 830 |
+
"step": 1370
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 0.3,
|
| 834 |
+
"learning_rate": 2.641576881974279e-05,
|
| 835 |
+
"loss": 1.2481,
|
| 836 |
+
"step": 1380
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 0.3,
|
| 840 |
+
"learning_rate": 2.6341022681118828e-05,
|
| 841 |
+
"loss": 1.2779,
|
| 842 |
+
"step": 1390
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.31,
|
| 846 |
+
"learning_rate": 2.6265613274597087e-05,
|
| 847 |
+
"loss": 1.2818,
|
| 848 |
+
"step": 1400
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 0.31,
|
| 852 |
+
"learning_rate": 2.6189545010418157e-05,
|
| 853 |
+
"loss": 1.281,
|
| 854 |
+
"step": 1410
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 0.31,
|
| 858 |
+
"learning_rate": 2.611282233735523e-05,
|
| 859 |
+
"loss": 1.2877,
|
| 860 |
+
"step": 1420
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 0.31,
|
| 864 |
+
"learning_rate": 2.6035449742453915e-05,
|
| 865 |
+
"loss": 1.296,
|
| 866 |
+
"step": 1430
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 0.32,
|
| 870 |
+
"learning_rate": 2.5957431750769836e-05,
|
| 871 |
+
"loss": 1.2733,
|
| 872 |
+
"step": 1440
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 0.32,
|
| 876 |
+
"learning_rate": 2.5878772925103954e-05,
|
| 877 |
+
"loss": 1.214,
|
| 878 |
+
"step": 1450
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 0.32,
|
| 882 |
+
"learning_rate": 2.5799477865735766e-05,
|
| 883 |
+
"loss": 1.2658,
|
| 884 |
+
"step": 1460
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.32,
|
| 888 |
+
"learning_rate": 2.5719551210154207e-05,
|
| 889 |
+
"loss": 1.3124,
|
| 890 |
+
"step": 1470
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 0.32,
|
| 894 |
+
"learning_rate": 2.5638997632786468e-05,
|
| 895 |
+
"loss": 1.3069,
|
| 896 |
+
"step": 1480
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 0.33,
|
| 900 |
+
"learning_rate": 2.5557821844724625e-05,
|
| 901 |
+
"loss": 1.2606,
|
| 902 |
+
"step": 1490
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 0.33,
|
| 906 |
+
"learning_rate": 2.5476028593450075e-05,
|
| 907 |
+
"loss": 1.2651,
|
| 908 |
+
"step": 1500
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 0.33,
|
| 912 |
+
"learning_rate": 2.5393622662555914e-05,
|
| 913 |
+
"loss": 1.2719,
|
| 914 |
+
"step": 1510
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 0.33,
|
| 918 |
+
"learning_rate": 2.5310608871467174e-05,
|
| 919 |
+
"loss": 1.2728,
|
| 920 |
+
"step": 1520
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 0.34,
|
| 924 |
+
"learning_rate": 2.522699207515895e-05,
|
| 925 |
+
"loss": 1.2169,
|
| 926 |
+
"step": 1530
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.34,
|
| 930 |
+
"learning_rate": 2.5142777163872472e-05,
|
| 931 |
+
"loss": 1.3261,
|
| 932 |
+
"step": 1540
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 0.34,
|
| 936 |
+
"learning_rate": 2.5057969062829107e-05,
|
| 937 |
+
"loss": 1.2942,
|
| 938 |
+
"step": 1550
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 0.34,
|
| 942 |
+
"learning_rate": 2.4972572731942294e-05,
|
| 943 |
+
"loss": 1.2167,
|
| 944 |
+
"step": 1560
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 0.34,
|
| 948 |
+
"learning_rate": 2.488659316552749e-05,
|
| 949 |
+
"loss": 1.2768,
|
| 950 |
+
"step": 1570
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 0.35,
|
| 954 |
+
"learning_rate": 2.4800035392010082e-05,
|
| 955 |
+
"loss": 1.2514,
|
| 956 |
+
"step": 1580
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 0.35,
|
| 960 |
+
"learning_rate": 2.471290447363128e-05,
|
| 961 |
+
"loss": 1.2344,
|
| 962 |
+
"step": 1590
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 0.35,
|
| 966 |
+
"learning_rate": 2.4625205506152083e-05,
|
| 967 |
+
"loss": 1.242,
|
| 968 |
+
"step": 1600
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.35,
|
| 972 |
+
"learning_rate": 2.4536943618555256e-05,
|
| 973 |
+
"loss": 1.2482,
|
| 974 |
+
"step": 1610
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 0.35,
|
| 978 |
+
"learning_rate": 2.444812397274535e-05,
|
| 979 |
+
"loss": 1.2741,
|
| 980 |
+
"step": 1620
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 0.36,
|
| 984 |
+
"learning_rate": 2.4358751763246824e-05,
|
| 985 |
+
"loss": 1.2061,
|
| 986 |
+
"step": 1630
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 0.36,
|
| 990 |
+
"learning_rate": 2.4268832216900263e-05,
|
| 991 |
+
"loss": 1.2727,
|
| 992 |
+
"step": 1640
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 0.36,
|
| 996 |
+
"learning_rate": 2.4178370592556657e-05,
|
| 997 |
+
"loss": 1.2432,
|
| 998 |
+
"step": 1650
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 0.36,
|
| 1002 |
+
"learning_rate": 2.4087372180769888e-05,
|
| 1003 |
+
"loss": 1.2156,
|
| 1004 |
+
"step": 1660
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 0.37,
|
| 1008 |
+
"learning_rate": 2.399584230348726e-05,
|
| 1009 |
+
"loss": 1.2286,
|
| 1010 |
+
"step": 1670
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.37,
|
| 1014 |
+
"learning_rate": 2.390378631373832e-05,
|
| 1015 |
+
"loss": 1.264,
|
| 1016 |
+
"step": 1680
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 0.37,
|
| 1020 |
+
"learning_rate": 2.3811209595321716e-05,
|
| 1021 |
+
"loss": 1.3205,
|
| 1022 |
+
"step": 1690
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 0.37,
|
| 1026 |
+
"learning_rate": 2.37181175624904e-05,
|
| 1027 |
+
"loss": 1.2897,
|
| 1028 |
+
"step": 1700
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 0.37,
|
| 1032 |
+
"learning_rate": 2.362451565963495e-05,
|
| 1033 |
+
"loss": 1.233,
|
| 1034 |
+
"step": 1710
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 0.38,
|
| 1038 |
+
"learning_rate": 2.3530409360965126e-05,
|
| 1039 |
+
"loss": 1.2131,
|
| 1040 |
+
"step": 1720
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 0.38,
|
| 1044 |
+
"learning_rate": 2.34358041701898e-05,
|
| 1045 |
+
"loss": 1.2611,
|
| 1046 |
+
"step": 1730
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 0.38,
|
| 1050 |
+
"learning_rate": 2.3340705620194995e-05,
|
| 1051 |
+
"loss": 1.2365,
|
| 1052 |
+
"step": 1740
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.38,
|
| 1056 |
+
"learning_rate": 2.3245119272720344e-05,
|
| 1057 |
+
"loss": 1.208,
|
| 1058 |
+
"step": 1750
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 0.39,
|
| 1062 |
+
"learning_rate": 2.3149050718033815e-05,
|
| 1063 |
+
"loss": 1.2632,
|
| 1064 |
+
"step": 1760
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 0.39,
|
| 1068 |
+
"learning_rate": 2.3052505574604753e-05,
|
| 1069 |
+
"loss": 1.2436,
|
| 1070 |
+
"step": 1770
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 0.39,
|
| 1074 |
+
"learning_rate": 2.29554894887753e-05,
|
| 1075 |
+
"loss": 1.2279,
|
| 1076 |
+
"step": 1780
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 0.39,
|
| 1080 |
+
"learning_rate": 2.2858008134430182e-05,
|
| 1081 |
+
"loss": 1.3023,
|
| 1082 |
+
"step": 1790
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 0.39,
|
| 1086 |
+
"learning_rate": 2.2760067212664852e-05,
|
| 1087 |
+
"loss": 1.2554,
|
| 1088 |
+
"step": 1800
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 0.4,
|
| 1092 |
+
"learning_rate": 2.2661672451452112e-05,
|
| 1093 |
+
"loss": 1.2102,
|
| 1094 |
+
"step": 1810
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 0.4,
|
| 1098 |
+
"learning_rate": 2.2562829605307057e-05,
|
| 1099 |
+
"loss": 1.271,
|
| 1100 |
+
"step": 1820
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 0.4,
|
| 1104 |
+
"learning_rate": 2.2463544454950594e-05,
|
| 1105 |
+
"loss": 1.2741,
|
| 1106 |
+
"step": 1830
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 0.4,
|
| 1110 |
+
"learning_rate": 2.2363822806971322e-05,
|
| 1111 |
+
"loss": 1.2299,
|
| 1112 |
+
"step": 1840
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 0.41,
|
| 1116 |
+
"learning_rate": 2.2263670493485935e-05,
|
| 1117 |
+
"loss": 1.2402,
|
| 1118 |
+
"step": 1850
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 0.41,
|
| 1122 |
+
"learning_rate": 2.2163093371798165e-05,
|
| 1123 |
+
"loss": 1.2592,
|
| 1124 |
+
"step": 1860
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 0.41,
|
| 1128 |
+
"learning_rate": 2.20620973240562e-05,
|
| 1129 |
+
"loss": 1.2148,
|
| 1130 |
+
"step": 1870
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 0.41,
|
| 1134 |
+
"learning_rate": 2.1960688256908693e-05,
|
| 1135 |
+
"loss": 1.3015,
|
| 1136 |
+
"step": 1880
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 0.41,
|
| 1140 |
+
"learning_rate": 2.1858872101159293e-05,
|
| 1141 |
+
"loss": 1.2337,
|
| 1142 |
+
"step": 1890
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 0.42,
|
| 1146 |
+
"learning_rate": 2.175665481141982e-05,
|
| 1147 |
+
"loss": 1.2014,
|
| 1148 |
+
"step": 1900
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 0.42,
|
| 1152 |
+
"learning_rate": 2.1654042365761974e-05,
|
| 1153 |
+
"loss": 1.2187,
|
| 1154 |
+
"step": 1910
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 0.42,
|
| 1158 |
+
"learning_rate": 2.1551040765367757e-05,
|
| 1159 |
+
"loss": 1.2193,
|
| 1160 |
+
"step": 1920
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 0.42,
|
| 1164 |
+
"learning_rate": 2.1447656034178475e-05,
|
| 1165 |
+
"loss": 1.265,
|
| 1166 |
+
"step": 1930
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 0.42,
|
| 1170 |
+
"learning_rate": 2.134389421854243e-05,
|
| 1171 |
+
"loss": 1.2408,
|
| 1172 |
+
"step": 1940
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 0.43,
|
| 1176 |
+
"learning_rate": 2.1239761386861332e-05,
|
| 1177 |
+
"loss": 1.1964,
|
| 1178 |
+
"step": 1950
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 0.43,
|
| 1182 |
+
"learning_rate": 2.1135263629235358e-05,
|
| 1183 |
+
"loss": 1.254,
|
| 1184 |
+
"step": 1960
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 0.43,
|
| 1188 |
+
"learning_rate": 2.103040705710702e-05,
|
| 1189 |
+
"loss": 1.2297,
|
| 1190 |
+
"step": 1970
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 0.43,
|
| 1194 |
+
"learning_rate": 2.0925197802903702e-05,
|
| 1195 |
+
"loss": 1.2397,
|
| 1196 |
+
"step": 1980
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 0.44,
|
| 1200 |
+
"learning_rate": 2.0819642019679053e-05,
|
| 1201 |
+
"loss": 1.2802,
|
| 1202 |
+
"step": 1990
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 0.44,
|
| 1206 |
+
"learning_rate": 2.0713745880753093e-05,
|
| 1207 |
+
"loss": 1.2435,
|
| 1208 |
+
"step": 2000
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 0.44,
|
| 1212 |
+
"learning_rate": 2.0607515579351196e-05,
|
| 1213 |
+
"loss": 1.2494,
|
| 1214 |
+
"step": 2010
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 0.44,
|
| 1218 |
+
"learning_rate": 2.0500957328241884e-05,
|
| 1219 |
+
"loss": 1.2514,
|
| 1220 |
+
"step": 2020
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 0.44,
|
| 1224 |
+
"learning_rate": 2.0394077359373473e-05,
|
| 1225 |
+
"loss": 1.2539,
|
| 1226 |
+
"step": 2030
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 0.45,
|
| 1230 |
+
"learning_rate": 2.0286881923509602e-05,
|
| 1231 |
+
"loss": 1.2823,
|
| 1232 |
+
"step": 2040
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 0.45,
|
| 1236 |
+
"learning_rate": 2.0179377289863683e-05,
|
| 1237 |
+
"loss": 1.2449,
|
| 1238 |
+
"step": 2050
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 0.45,
|
| 1242 |
+
"learning_rate": 2.0071569745732223e-05,
|
| 1243 |
+
"loss": 1.2446,
|
| 1244 |
+
"step": 2060
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 0.45,
|
| 1248 |
+
"learning_rate": 1.9963465596127142e-05,
|
| 1249 |
+
"loss": 1.2498,
|
| 1250 |
+
"step": 2070
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 0.46,
|
| 1254 |
+
"learning_rate": 1.985507116340702e-05,
|
| 1255 |
+
"loss": 1.2032,
|
| 1256 |
+
"step": 2080
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 0.46,
|
| 1260 |
+
"learning_rate": 1.974639278690734e-05,
|
| 1261 |
+
"loss": 1.2074,
|
| 1262 |
+
"step": 2090
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 0.46,
|
| 1266 |
+
"learning_rate": 1.9637436822569742e-05,
|
| 1267 |
+
"loss": 1.2406,
|
| 1268 |
+
"step": 2100
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 0.46,
|
| 1272 |
+
"learning_rate": 1.9528209642570312e-05,
|
| 1273 |
+
"loss": 1.2512,
|
| 1274 |
+
"step": 2110
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 0.46,
|
| 1278 |
+
"learning_rate": 1.9418717634946886e-05,
|
| 1279 |
+
"loss": 1.2406,
|
| 1280 |
+
"step": 2120
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 0.47,
|
| 1284 |
+
"learning_rate": 1.9308967203225465e-05,
|
| 1285 |
+
"loss": 1.2125,
|
| 1286 |
+
"step": 2130
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 0.47,
|
| 1290 |
+
"learning_rate": 1.9198964766045744e-05,
|
| 1291 |
+
"loss": 1.2741,
|
| 1292 |
+
"step": 2140
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 0.47,
|
| 1296 |
+
"learning_rate": 1.9088716756785652e-05,
|
| 1297 |
+
"loss": 1.2423,
|
| 1298 |
+
"step": 2150
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 0.47,
|
| 1302 |
+
"learning_rate": 1.8978229623185185e-05,
|
| 1303 |
+
"loss": 1.245,
|
| 1304 |
+
"step": 2160
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 0.48,
|
| 1308 |
+
"learning_rate": 1.886750982696925e-05,
|
| 1309 |
+
"loss": 1.2462,
|
| 1310 |
+
"step": 2170
|
| 1311 |
+
},
|
| 1312 |
+
{
|
| 1313 |
+
"epoch": 0.48,
|
| 1314 |
+
"learning_rate": 1.8756563843469796e-05,
|
| 1315 |
+
"loss": 1.2165,
|
| 1316 |
+
"step": 2180
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 0.48,
|
| 1320 |
+
"learning_rate": 1.8645398161247095e-05,
|
| 1321 |
+
"loss": 1.241,
|
| 1322 |
+
"step": 2190
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 0.48,
|
| 1326 |
+
"learning_rate": 1.853401928171028e-05,
|
| 1327 |
+
"loss": 1.254,
|
| 1328 |
+
"step": 2200
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 0.48,
|
| 1332 |
+
"learning_rate": 1.8422433718737094e-05,
|
| 1333 |
+
"loss": 1.2337,
|
| 1334 |
+
"step": 2210
|
| 1335 |
+
},
|
| 1336 |
+
{
|
| 1337 |
+
"epoch": 0.49,
|
| 1338 |
+
"learning_rate": 1.8310647998292953e-05,
|
| 1339 |
+
"loss": 1.2657,
|
| 1340 |
+
"step": 2220
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 0.49,
|
| 1344 |
+
"learning_rate": 1.8198668658049263e-05,
|
| 1345 |
+
"loss": 1.2359,
|
| 1346 |
+
"step": 2230
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 0.49,
|
| 1350 |
+
"learning_rate": 1.808650224700111e-05,
|
| 1351 |
+
"loss": 1.2349,
|
| 1352 |
+
"step": 2240
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"epoch": 0.49,
|
| 1356 |
+
"learning_rate": 1.7974155325084183e-05,
|
| 1357 |
+
"loss": 1.2116,
|
| 1358 |
+
"step": 2250
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 0.49,
|
| 1362 |
+
"learning_rate": 1.786163446279119e-05,
|
| 1363 |
+
"loss": 1.2823,
|
| 1364 |
+
"step": 2260
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 0.5,
|
| 1368 |
+
"learning_rate": 1.7748946240787555e-05,
|
| 1369 |
+
"loss": 1.236,
|
| 1370 |
+
"step": 2270
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 0.5,
|
| 1374 |
+
"learning_rate": 1.7636097249526556e-05,
|
| 1375 |
+
"loss": 1.2378,
|
| 1376 |
+
"step": 2280
|
| 1377 |
+
},
|
| 1378 |
+
{
|
| 1379 |
+
"epoch": 0.5,
|
| 1380 |
+
"learning_rate": 1.7523094088863885e-05,
|
| 1381 |
+
"loss": 1.2696,
|
| 1382 |
+
"step": 2290
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 0.5,
|
| 1386 |
+
"learning_rate": 1.7409943367671685e-05,
|
| 1387 |
+
"loss": 1.1972,
|
| 1388 |
+
"step": 2300
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 0.51,
|
| 1392 |
+
"learning_rate": 1.7296651703452022e-05,
|
| 1393 |
+
"loss": 1.235,
|
| 1394 |
+
"step": 2310
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 0.51,
|
| 1398 |
+
"learning_rate": 1.7183225721949857e-05,
|
| 1399 |
+
"loss": 1.252,
|
| 1400 |
+
"step": 2320
|
| 1401 |
+
},
|
| 1402 |
+
{
|
| 1403 |
+
"epoch": 0.51,
|
| 1404 |
+
"learning_rate": 1.7069672056765573e-05,
|
| 1405 |
+
"loss": 1.2441,
|
| 1406 |
+
"step": 2330
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 0.51,
|
| 1410 |
+
"learning_rate": 1.6955997348966986e-05,
|
| 1411 |
+
"loss": 1.2437,
|
| 1412 |
+
"step": 2340
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 0.51,
|
| 1416 |
+
"learning_rate": 1.6842208246700966e-05,
|
| 1417 |
+
"loss": 1.2579,
|
| 1418 |
+
"step": 2350
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 0.52,
|
| 1422 |
+
"learning_rate": 1.6728311404804632e-05,
|
| 1423 |
+
"loss": 1.1966,
|
| 1424 |
+
"step": 2360
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 0.52,
|
| 1428 |
+
"learning_rate": 1.661431348441614e-05,
|
| 1429 |
+
"loss": 1.2409,
|
| 1430 |
+
"step": 2370
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 0.52,
|
| 1434 |
+
"learning_rate": 1.6500221152585114e-05,
|
| 1435 |
+
"loss": 1.1716,
|
| 1436 |
+
"step": 2380
|
| 1437 |
+
},
|
| 1438 |
+
{
|
| 1439 |
+
"epoch": 0.52,
|
| 1440 |
+
"learning_rate": 1.6386041081882727e-05,
|
| 1441 |
+
"loss": 1.3371,
|
| 1442 |
+
"step": 2390
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 0.53,
|
| 1446 |
+
"learning_rate": 1.627177995001148e-05,
|
| 1447 |
+
"loss": 1.2638,
|
| 1448 |
+
"step": 2400
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 0.53,
|
| 1452 |
+
"learning_rate": 1.615744443941464e-05,
|
| 1453 |
+
"loss": 1.2183,
|
| 1454 |
+
"step": 2410
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 0.53,
|
| 1458 |
+
"learning_rate": 1.6043041236885457e-05,
|
| 1459 |
+
"loss": 1.2175,
|
| 1460 |
+
"step": 2420
|
| 1461 |
+
},
|
| 1462 |
+
{
|
| 1463 |
+
"epoch": 0.53,
|
| 1464 |
+
"learning_rate": 1.5928577033176043e-05,
|
| 1465 |
+
"loss": 1.25,
|
| 1466 |
+
"step": 2430
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 0.53,
|
| 1470 |
+
"learning_rate": 1.5814058522606132e-05,
|
| 1471 |
+
"loss": 1.244,
|
| 1472 |
+
"step": 2440
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 0.54,
|
| 1476 |
+
"learning_rate": 1.569949240267151e-05,
|
| 1477 |
+
"loss": 1.1961,
|
| 1478 |
+
"step": 2450
|
| 1479 |
+
},
|
| 1480 |
+
{
|
| 1481 |
+
"epoch": 0.54,
|
| 1482 |
+
"learning_rate": 1.558488537365237e-05,
|
| 1483 |
+
"loss": 1.2389,
|
| 1484 |
+
"step": 2460
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 0.54,
|
| 1488 |
+
"learning_rate": 1.5470244138221415e-05,
|
| 1489 |
+
"loss": 1.2437,
|
| 1490 |
+
"step": 2470
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"epoch": 0.54,
|
| 1494 |
+
"learning_rate": 1.535557540105188e-05,
|
| 1495 |
+
"loss": 1.2218,
|
| 1496 |
+
"step": 2480
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 0.55,
|
| 1500 |
+
"learning_rate": 1.5240885868425409e-05,
|
| 1501 |
+
"loss": 1.2585,
|
| 1502 |
+
"step": 2490
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 0.55,
|
| 1506 |
+
"learning_rate": 1.5126182247839853e-05,
|
| 1507 |
+
"loss": 1.2138,
|
| 1508 |
+
"step": 2500
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 0.55,
|
| 1512 |
+
"learning_rate": 1.5011471247616968e-05,
|
| 1513 |
+
"loss": 1.2291,
|
| 1514 |
+
"step": 2510
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 0.55,
|
| 1518 |
+
"learning_rate": 1.4896759576510118e-05,
|
| 1519 |
+
"loss": 1.246,
|
| 1520 |
+
"step": 2520
|
| 1521 |
+
},
|
| 1522 |
+
{
|
| 1523 |
+
"epoch": 0.55,
|
| 1524 |
+
"learning_rate": 1.4782053943311894e-05,
|
| 1525 |
+
"loss": 1.2412,
|
| 1526 |
+
"step": 2530
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 0.56,
|
| 1530 |
+
"learning_rate": 1.4667361056461766e-05,
|
| 1531 |
+
"loss": 1.2486,
|
| 1532 |
+
"step": 2540
|
| 1533 |
+
},
|
| 1534 |
+
{
|
| 1535 |
+
"epoch": 0.56,
|
| 1536 |
+
"learning_rate": 1.4552687623653744e-05,
|
| 1537 |
+
"loss": 1.2382,
|
| 1538 |
+
"step": 2550
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 0.56,
|
| 1542 |
+
"learning_rate": 1.4438040351444098e-05,
|
| 1543 |
+
"loss": 1.1778,
|
| 1544 |
+
"step": 2560
|
| 1545 |
+
},
|
| 1546 |
+
{
|
| 1547 |
+
"epoch": 0.56,
|
| 1548 |
+
"learning_rate": 1.4323425944859116e-05,
|
| 1549 |
+
"loss": 1.2197,
|
| 1550 |
+
"step": 2570
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 0.57,
|
| 1554 |
+
"learning_rate": 1.4208851107002975e-05,
|
| 1555 |
+
"loss": 1.2182,
|
| 1556 |
+
"step": 2580
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 0.57,
|
| 1560 |
+
"learning_rate": 1.4094322538665707e-05,
|
| 1561 |
+
"loss": 1.2052,
|
| 1562 |
+
"step": 2590
|
| 1563 |
+
},
|
| 1564 |
+
{
|
| 1565 |
+
"epoch": 0.57,
|
| 1566 |
+
"learning_rate": 1.3979846937931348e-05,
|
| 1567 |
+
"loss": 1.1819,
|
| 1568 |
+
"step": 2600
|
| 1569 |
+
},
|
| 1570 |
+
{
|
| 1571 |
+
"epoch": 0.57,
|
| 1572 |
+
"learning_rate": 1.3865430999786147e-05,
|
| 1573 |
+
"loss": 1.1735,
|
| 1574 |
+
"step": 2610
|
| 1575 |
+
},
|
| 1576 |
+
{
|
| 1577 |
+
"epoch": 0.57,
|
| 1578 |
+
"learning_rate": 1.3751081415727068e-05,
|
| 1579 |
+
"loss": 1.2296,
|
| 1580 |
+
"step": 2620
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 0.58,
|
| 1584 |
+
"learning_rate": 1.3636804873370409e-05,
|
| 1585 |
+
"loss": 1.2094,
|
| 1586 |
+
"step": 2630
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 0.58,
|
| 1590 |
+
"learning_rate": 1.3522608056060722e-05,
|
| 1591 |
+
"loss": 1.2369,
|
| 1592 |
+
"step": 2640
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 0.58,
|
| 1596 |
+
"learning_rate": 1.3408497642479905e-05,
|
| 1597 |
+
"loss": 1.2653,
|
| 1598 |
+
"step": 2650
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 0.58,
|
| 1602 |
+
"learning_rate": 1.3294480306256629e-05,
|
| 1603 |
+
"loss": 1.1777,
|
| 1604 |
+
"step": 2660
|
| 1605 |
+
},
|
| 1606 |
+
{
|
| 1607 |
+
"epoch": 0.58,
|
| 1608 |
+
"learning_rate": 1.3180562715576029e-05,
|
| 1609 |
+
"loss": 1.2748,
|
| 1610 |
+
"step": 2670
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 0.59,
|
| 1614 |
+
"learning_rate": 1.306675153278974e-05,
|
| 1615 |
+
"loss": 1.2337,
|
| 1616 |
+
"step": 2680
|
| 1617 |
+
},
|
| 1618 |
+
{
|
| 1619 |
+
"epoch": 0.59,
|
| 1620 |
+
"learning_rate": 1.295305341402623e-05,
|
| 1621 |
+
"loss": 1.1837,
|
| 1622 |
+
"step": 2690
|
| 1623 |
+
},
|
| 1624 |
+
{
|
| 1625 |
+
"epoch": 0.59,
|
| 1626 |
+
"learning_rate": 1.2839475008801526e-05,
|
| 1627 |
+
"loss": 1.1723,
|
| 1628 |
+
"step": 2700
|
| 1629 |
+
},
|
| 1630 |
+
{
|
| 1631 |
+
"epoch": 0.59,
|
| 1632 |
+
"learning_rate": 1.2726022959630363e-05,
|
| 1633 |
+
"loss": 1.1999,
|
| 1634 |
+
"step": 2710
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 0.6,
|
| 1638 |
+
"learning_rate": 1.2612703901637648e-05,
|
| 1639 |
+
"loss": 1.245,
|
| 1640 |
+
"step": 2720
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 0.6,
|
| 1644 |
+
"learning_rate": 1.2499524462170441e-05,
|
| 1645 |
+
"loss": 1.2375,
|
| 1646 |
+
"step": 2730
|
| 1647 |
+
},
|
| 1648 |
+
{
|
| 1649 |
+
"epoch": 0.6,
|
| 1650 |
+
"learning_rate": 1.2386491260410369e-05,
|
| 1651 |
+
"loss": 1.2432,
|
| 1652 |
+
"step": 2740
|
| 1653 |
+
},
|
| 1654 |
+
{
|
| 1655 |
+
"epoch": 0.6,
|
| 1656 |
+
"learning_rate": 1.227361090698649e-05,
|
| 1657 |
+
"loss": 1.1958,
|
| 1658 |
+
"step": 2750
|
| 1659 |
+
},
|
| 1660 |
+
{
|
| 1661 |
+
"epoch": 0.6,
|
| 1662 |
+
"learning_rate": 1.2160890003588686e-05,
|
| 1663 |
+
"loss": 1.2432,
|
| 1664 |
+
"step": 2760
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 0.61,
|
| 1668 |
+
"learning_rate": 1.2048335142581578e-05,
|
| 1669 |
+
"loss": 1.1742,
|
| 1670 |
+
"step": 2770
|
| 1671 |
+
},
|
| 1672 |
+
{
|
| 1673 |
+
"epoch": 0.61,
|
| 1674 |
+
"learning_rate": 1.1935952906618955e-05,
|
| 1675 |
+
"loss": 1.2097,
|
| 1676 |
+
"step": 2780
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 0.61,
|
| 1680 |
+
"learning_rate": 1.1823749868258836e-05,
|
| 1681 |
+
"loss": 1.2993,
|
| 1682 |
+
"step": 2790
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 0.61,
|
| 1686 |
+
"learning_rate": 1.1711732589579032e-05,
|
| 1687 |
+
"loss": 1.1788,
|
| 1688 |
+
"step": 2800
|
| 1689 |
+
},
|
| 1690 |
+
{
|
| 1691 |
+
"epoch": 0.62,
|
| 1692 |
+
"learning_rate": 1.1599907621793398e-05,
|
| 1693 |
+
"loss": 1.2499,
|
| 1694 |
+
"step": 2810
|
| 1695 |
+
},
|
| 1696 |
+
{
|
| 1697 |
+
"epoch": 0.62,
|
| 1698 |
+
"learning_rate": 1.1488281504868695e-05,
|
| 1699 |
+
"loss": 1.2777,
|
| 1700 |
+
"step": 2820
|
| 1701 |
+
},
|
| 1702 |
+
{
|
| 1703 |
+
"epoch": 0.62,
|
| 1704 |
+
"learning_rate": 1.137686076714209e-05,
|
| 1705 |
+
"loss": 1.2305,
|
| 1706 |
+
"step": 2830
|
| 1707 |
+
},
|
| 1708 |
+
{
|
| 1709 |
+
"epoch": 0.62,
|
| 1710 |
+
"learning_rate": 1.1265651924939364e-05,
|
| 1711 |
+
"loss": 1.2181,
|
| 1712 |
+
"step": 2840
|
| 1713 |
+
},
|
| 1714 |
+
{
|
| 1715 |
+
"epoch": 0.62,
|
| 1716 |
+
"learning_rate": 1.11546614821938e-05,
|
| 1717 |
+
"loss": 1.2418,
|
| 1718 |
+
"step": 2850
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 0.63,
|
| 1722 |
+
"learning_rate": 1.1043895930065843e-05,
|
| 1723 |
+
"loss": 1.2372,
|
| 1724 |
+
"step": 2860
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 0.63,
|
| 1728 |
+
"learning_rate": 1.0933361746563418e-05,
|
| 1729 |
+
"loss": 1.2151,
|
| 1730 |
+
"step": 2870
|
| 1731 |
+
},
|
| 1732 |
+
{
|
| 1733 |
+
"epoch": 0.63,
|
| 1734 |
+
"learning_rate": 1.0823065396163104e-05,
|
| 1735 |
+
"loss": 1.1852,
|
| 1736 |
+
"step": 2880
|
| 1737 |
+
},
|
| 1738 |
+
{
|
| 1739 |
+
"epoch": 0.63,
|
| 1740 |
+
"learning_rate": 1.071301332943206e-05,
|
| 1741 |
+
"loss": 1.1956,
|
| 1742 |
+
"step": 2890
|
| 1743 |
+
},
|
| 1744 |
+
{
|
| 1745 |
+
"epoch": 0.64,
|
| 1746 |
+
"learning_rate": 1.0603211982650775e-05,
|
| 1747 |
+
"loss": 1.2443,
|
| 1748 |
+
"step": 2900
|
| 1749 |
+
},
|
| 1750 |
+
{
|
| 1751 |
+
"epoch": 0.64,
|
| 1752 |
+
"learning_rate": 1.0493667777436633e-05,
|
| 1753 |
+
"loss": 1.1962,
|
| 1754 |
+
"step": 2910
|
| 1755 |
+
},
|
| 1756 |
+
{
|
| 1757 |
+
"epoch": 0.64,
|
| 1758 |
+
"learning_rate": 1.0384387120368367e-05,
|
| 1759 |
+
"loss": 1.2033,
|
| 1760 |
+
"step": 2920
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 0.64,
|
| 1764 |
+
"learning_rate": 1.0275376402611364e-05,
|
| 1765 |
+
"loss": 1.279,
|
| 1766 |
+
"step": 2930
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 0.64,
|
| 1770 |
+
"learning_rate": 1.0166641999543912e-05,
|
| 1771 |
+
"loss": 1.2342,
|
| 1772 |
+
"step": 2940
|
| 1773 |
+
},
|
| 1774 |
+
{
|
| 1775 |
+
"epoch": 0.65,
|
| 1776 |
+
"learning_rate": 1.0058190270384298e-05,
|
| 1777 |
+
"loss": 1.2511,
|
| 1778 |
+
"step": 2950
|
| 1779 |
+
},
|
| 1780 |
+
{
|
| 1781 |
+
"epoch": 0.65,
|
| 1782 |
+
"learning_rate": 9.950027557818926e-06,
|
| 1783 |
+
"loss": 1.2051,
|
| 1784 |
+
"step": 2960
|
| 1785 |
+
},
|
| 1786 |
+
{
|
| 1787 |
+
"epoch": 0.65,
|
| 1788 |
+
"learning_rate": 9.84216018763138e-06,
|
| 1789 |
+
"loss": 1.1881,
|
| 1790 |
+
"step": 2970
|
| 1791 |
+
},
|
| 1792 |
+
{
|
| 1793 |
+
"epoch": 0.65,
|
| 1794 |
+
"learning_rate": 9.734594468332436e-06,
|
| 1795 |
+
"loss": 1.2033,
|
| 1796 |
+
"step": 2980
|
| 1797 |
+
},
|
| 1798 |
+
{
|
| 1799 |
+
"epoch": 0.65,
|
| 1800 |
+
"learning_rate": 9.627336690791142e-06,
|
| 1801 |
+
"loss": 1.2522,
|
| 1802 |
+
"step": 2990
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 0.66,
|
| 1806 |
+
"learning_rate": 9.520393127866881e-06,
|
| 1807 |
+
"loss": 1.1684,
|
| 1808 |
+
"step": 3000
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 0.66,
|
| 1812 |
+
"learning_rate": 9.41377003404254e-06,
|
| 1813 |
+
"loss": 1.182,
|
| 1814 |
+
"step": 3010
|
| 1815 |
+
},
|
| 1816 |
+
{
|
| 1817 |
+
"epoch": 0.66,
|
| 1818 |
+
"learning_rate": 9.307473645058684e-06,
|
| 1819 |
+
"loss": 1.2219,
|
| 1820 |
+
"step": 3020
|
| 1821 |
+
},
|
| 1822 |
+
{
|
| 1823 |
+
"epoch": 0.66,
|
| 1824 |
+
"learning_rate": 9.201510177548903e-06,
|
| 1825 |
+
"loss": 1.271,
|
| 1826 |
+
"step": 3030
|
| 1827 |
+
},
|
| 1828 |
+
{
|
| 1829 |
+
"epoch": 0.67,
|
| 1830 |
+
"learning_rate": 9.095885828676206e-06,
|
| 1831 |
+
"loss": 1.2031,
|
| 1832 |
+
"step": 3040
|
| 1833 |
+
},
|
| 1834 |
+
{
|
| 1835 |
+
"epoch": 0.67,
|
| 1836 |
+
"learning_rate": 8.990606775770622e-06,
|
| 1837 |
+
"loss": 1.2099,
|
| 1838 |
+
"step": 3050
|
| 1839 |
+
},
|
| 1840 |
+
{
|
| 1841 |
+
"epoch": 0.67,
|
| 1842 |
+
"learning_rate": 8.885679175967886e-06,
|
| 1843 |
+
"loss": 1.1929,
|
| 1844 |
+
"step": 3060
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 0.67,
|
| 1848 |
+
"learning_rate": 8.78110916584937e-06,
|
| 1849 |
+
"loss": 1.1839,
|
| 1850 |
+
"step": 3070
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 0.67,
|
| 1854 |
+
"learning_rate": 8.676902861083185e-06,
|
| 1855 |
+
"loss": 1.1661,
|
| 1856 |
+
"step": 3080
|
| 1857 |
+
},
|
| 1858 |
+
{
|
| 1859 |
+
"epoch": 0.68,
|
| 1860 |
+
"learning_rate": 8.573066356066528e-06,
|
| 1861 |
+
"loss": 1.2045,
|
| 1862 |
+
"step": 3090
|
| 1863 |
+
},
|
| 1864 |
+
{
|
| 1865 |
+
"epoch": 0.68,
|
| 1866 |
+
"learning_rate": 8.469605723569216e-06,
|
| 1867 |
+
"loss": 1.2661,
|
| 1868 |
+
"step": 3100
|
| 1869 |
+
},
|
| 1870 |
+
{
|
| 1871 |
+
"epoch": 0.68,
|
| 1872 |
+
"learning_rate": 8.366527014378575e-06,
|
| 1873 |
+
"loss": 1.2367,
|
| 1874 |
+
"step": 3110
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 0.68,
|
| 1878 |
+
"learning_rate": 8.263836256945541e-06,
|
| 1879 |
+
"loss": 1.2267,
|
| 1880 |
+
"step": 3120
|
| 1881 |
+
},
|
| 1882 |
+
{
|
| 1883 |
+
"epoch": 0.69,
|
| 1884 |
+
"learning_rate": 8.16153945703208e-06,
|
| 1885 |
+
"loss": 1.1673,
|
| 1886 |
+
"step": 3130
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"epoch": 0.69,
|
| 1890 |
+
"learning_rate": 8.059642597359986e-06,
|
| 1891 |
+
"loss": 1.2081,
|
| 1892 |
+
"step": 3140
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 0.69,
|
| 1896 |
+
"learning_rate": 7.958151637260934e-06,
|
| 1897 |
+
"loss": 1.2117,
|
| 1898 |
+
"step": 3150
|
| 1899 |
+
},
|
| 1900 |
+
{
|
| 1901 |
+
"epoch": 0.69,
|
| 1902 |
+
"learning_rate": 7.85707251232803e-06,
|
| 1903 |
+
"loss": 1.1661,
|
| 1904 |
+
"step": 3160
|
| 1905 |
+
},
|
| 1906 |
+
{
|
| 1907 |
+
"epoch": 0.69,
|
| 1908 |
+
"learning_rate": 7.756411134068586e-06,
|
| 1909 |
+
"loss": 1.2254,
|
| 1910 |
+
"step": 3170
|
| 1911 |
+
},
|
| 1912 |
+
{
|
| 1913 |
+
"epoch": 0.7,
|
| 1914 |
+
"learning_rate": 7.656173389558464e-06,
|
| 1915 |
+
"loss": 1.2201,
|
| 1916 |
+
"step": 3180
|
| 1917 |
+
},
|
| 1918 |
+
{
|
| 1919 |
+
"epoch": 0.7,
|
| 1920 |
+
"learning_rate": 7.556365141097715e-06,
|
| 1921 |
+
"loss": 1.2059,
|
| 1922 |
+
"step": 3190
|
| 1923 |
+
},
|
| 1924 |
+
{
|
| 1925 |
+
"epoch": 0.7,
|
| 1926 |
+
"learning_rate": 7.456992225867811e-06,
|
| 1927 |
+
"loss": 1.2234,
|
| 1928 |
+
"step": 3200
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 0.7,
|
| 1932 |
+
"learning_rate": 7.358060455590162e-06,
|
| 1933 |
+
"loss": 1.27,
|
| 1934 |
+
"step": 3210
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 0.71,
|
| 1938 |
+
"learning_rate": 7.2595756161863095e-06,
|
| 1939 |
+
"loss": 1.2303,
|
| 1940 |
+
"step": 3220
|
| 1941 |
+
},
|
| 1942 |
+
{
|
| 1943 |
+
"epoch": 0.71,
|
| 1944 |
+
"learning_rate": 7.161543467439487e-06,
|
| 1945 |
+
"loss": 1.228,
|
| 1946 |
+
"step": 3230
|
| 1947 |
+
},
|
| 1948 |
+
{
|
| 1949 |
+
"epoch": 0.71,
|
| 1950 |
+
"learning_rate": 7.0639697426578e-06,
|
| 1951 |
+
"loss": 1.2104,
|
| 1952 |
+
"step": 3240
|
| 1953 |
+
},
|
| 1954 |
+
{
|
| 1955 |
+
"epoch": 0.71,
|
| 1956 |
+
"learning_rate": 6.966860148338909e-06,
|
| 1957 |
+
"loss": 1.1964,
|
| 1958 |
+
"step": 3250
|
| 1959 |
+
},
|
| 1960 |
+
{
|
| 1961 |
+
"epoch": 0.71,
|
| 1962 |
+
"learning_rate": 6.87022036383626e-06,
|
| 1963 |
+
"loss": 1.1855,
|
| 1964 |
+
"step": 3260
|
| 1965 |
+
},
|
| 1966 |
+
{
|
| 1967 |
+
"epoch": 0.72,
|
| 1968 |
+
"learning_rate": 6.7740560410270095e-06,
|
| 1969 |
+
"loss": 1.2372,
|
| 1970 |
+
"step": 3270
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"epoch": 0.72,
|
| 1974 |
+
"learning_rate": 6.678372803981382e-06,
|
| 1975 |
+
"loss": 1.2251,
|
| 1976 |
+
"step": 3280
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 0.72,
|
| 1980 |
+
"learning_rate": 6.58317624863384e-06,
|
| 1981 |
+
"loss": 1.178,
|
| 1982 |
+
"step": 3290
|
| 1983 |
+
},
|
| 1984 |
+
{
|
| 1985 |
+
"epoch": 0.72,
|
| 1986 |
+
"learning_rate": 6.488471942455745e-06,
|
| 1987 |
+
"loss": 1.2406,
|
| 1988 |
+
"step": 3300
|
| 1989 |
+
},
|
| 1990 |
+
{
|
| 1991 |
+
"epoch": 0.72,
|
| 1992 |
+
"learning_rate": 6.3942654241298e-06,
|
| 1993 |
+
"loss": 1.1814,
|
| 1994 |
+
"step": 3310
|
| 1995 |
+
},
|
| 1996 |
+
{
|
| 1997 |
+
"epoch": 0.73,
|
| 1998 |
+
"learning_rate": 6.300562203226101e-06,
|
| 1999 |
+
"loss": 1.1829,
|
| 2000 |
+
"step": 3320
|
| 2001 |
+
},
|
| 2002 |
+
{
|
| 2003 |
+
"epoch": 0.73,
|
| 2004 |
+
"learning_rate": 6.2073677598799e-06,
|
| 2005 |
+
"loss": 1.1957,
|
| 2006 |
+
"step": 3330
|
| 2007 |
+
},
|
| 2008 |
+
{
|
| 2009 |
+
"epoch": 0.73,
|
| 2010 |
+
"learning_rate": 6.114687544471146e-06,
|
| 2011 |
+
"loss": 1.2314,
|
| 2012 |
+
"step": 3340
|
| 2013 |
+
},
|
| 2014 |
+
{
|
| 2015 |
+
"epoch": 0.73,
|
| 2016 |
+
"learning_rate": 6.022526977305692e-06,
|
| 2017 |
+
"loss": 1.2095,
|
| 2018 |
+
"step": 3350
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 0.74,
|
| 2022 |
+
"learning_rate": 5.930891448298314e-06,
|
| 2023 |
+
"loss": 1.1936,
|
| 2024 |
+
"step": 3360
|
| 2025 |
+
},
|
| 2026 |
+
{
|
| 2027 |
+
"epoch": 0.74,
|
| 2028 |
+
"learning_rate": 5.83978631665746e-06,
|
| 2029 |
+
"loss": 1.1752,
|
| 2030 |
+
"step": 3370
|
| 2031 |
+
},
|
| 2032 |
+
{
|
| 2033 |
+
"epoch": 0.74,
|
| 2034 |
+
"learning_rate": 5.749216910571854e-06,
|
| 2035 |
+
"loss": 1.2165,
|
| 2036 |
+
"step": 3380
|
| 2037 |
+
},
|
| 2038 |
+
{
|
| 2039 |
+
"epoch": 0.74,
|
| 2040 |
+
"learning_rate": 5.659188526898875e-06,
|
| 2041 |
+
"loss": 1.1822,
|
| 2042 |
+
"step": 3390
|
| 2043 |
+
},
|
| 2044 |
+
{
|
| 2045 |
+
"epoch": 0.74,
|
| 2046 |
+
"learning_rate": 5.5697064308547495e-06,
|
| 2047 |
+
"loss": 1.1735,
|
| 2048 |
+
"step": 3400
|
| 2049 |
+
},
|
| 2050 |
+
{
|
| 2051 |
+
"epoch": 0.75,
|
| 2052 |
+
"learning_rate": 5.480775855706659e-06,
|
| 2053 |
+
"loss": 1.2234,
|
| 2054 |
+
"step": 3410
|
| 2055 |
+
},
|
| 2056 |
+
{
|
| 2057 |
+
"epoch": 0.75,
|
| 2058 |
+
"learning_rate": 5.392402002466651e-06,
|
| 2059 |
+
"loss": 1.2713,
|
| 2060 |
+
"step": 3420
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 0.75,
|
| 2064 |
+
"learning_rate": 5.304590039587483e-06,
|
| 2065 |
+
"loss": 1.2403,
|
| 2066 |
+
"step": 3430
|
| 2067 |
+
},
|
| 2068 |
+
{
|
| 2069 |
+
"epoch": 0.75,
|
| 2070 |
+
"learning_rate": 5.21734510266032e-06,
|
| 2071 |
+
"loss": 1.218,
|
| 2072 |
+
"step": 3440
|
| 2073 |
+
},
|
| 2074 |
+
{
|
| 2075 |
+
"epoch": 0.76,
|
| 2076 |
+
"learning_rate": 5.1306722941144235e-06,
|
| 2077 |
+
"loss": 1.2235,
|
| 2078 |
+
"step": 3450
|
| 2079 |
+
},
|
| 2080 |
+
{
|
| 2081 |
+
"epoch": 0.76,
|
| 2082 |
+
"learning_rate": 5.044576682918714e-06,
|
| 2083 |
+
"loss": 1.2126,
|
| 2084 |
+
"step": 3460
|
| 2085 |
+
},
|
| 2086 |
+
{
|
| 2087 |
+
"epoch": 0.76,
|
| 2088 |
+
"learning_rate": 4.959063304285335e-06,
|
| 2089 |
+
"loss": 1.2443,
|
| 2090 |
+
"step": 3470
|
| 2091 |
+
},
|
| 2092 |
+
{
|
| 2093 |
+
"epoch": 0.76,
|
| 2094 |
+
"learning_rate": 4.874137159375147e-06,
|
| 2095 |
+
"loss": 1.2198,
|
| 2096 |
+
"step": 3480
|
| 2097 |
+
},
|
| 2098 |
+
{
|
| 2099 |
+
"epoch": 0.76,
|
| 2100 |
+
"learning_rate": 4.789803215005274e-06,
|
| 2101 |
+
"loss": 1.2177,
|
| 2102 |
+
"step": 3490
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 0.77,
|
| 2106 |
+
"learning_rate": 4.70606640335861e-06,
|
| 2107 |
+
"loss": 1.2277,
|
| 2108 |
+
"step": 3500
|
| 2109 |
+
},
|
| 2110 |
+
{
|
| 2111 |
+
"epoch": 0.77,
|
| 2112 |
+
"learning_rate": 4.622931621695345e-06,
|
| 2113 |
+
"loss": 1.2223,
|
| 2114 |
+
"step": 3510
|
| 2115 |
+
},
|
| 2116 |
+
{
|
| 2117 |
+
"epoch": 0.77,
|
| 2118 |
+
"learning_rate": 4.540403732066591e-06,
|
| 2119 |
+
"loss": 1.2198,
|
| 2120 |
+
"step": 3520
|
| 2121 |
+
},
|
| 2122 |
+
{
|
| 2123 |
+
"epoch": 0.77,
|
| 2124 |
+
"learning_rate": 4.458487561030004e-06,
|
| 2125 |
+
"loss": 1.2308,
|
| 2126 |
+
"step": 3530
|
| 2127 |
+
},
|
| 2128 |
+
{
|
| 2129 |
+
"epoch": 0.78,
|
| 2130 |
+
"learning_rate": 4.377187899367523e-06,
|
| 2131 |
+
"loss": 1.1712,
|
| 2132 |
+
"step": 3540
|
| 2133 |
+
},
|
| 2134 |
+
{
|
| 2135 |
+
"epoch": 0.78,
|
| 2136 |
+
"learning_rate": 4.296509501805161e-06,
|
| 2137 |
+
"loss": 1.2379,
|
| 2138 |
+
"step": 3550
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"epoch": 0.78,
|
| 2142 |
+
"learning_rate": 4.2164570867349635e-06,
|
| 2143 |
+
"loss": 1.2465,
|
| 2144 |
+
"step": 3560
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 0.78,
|
| 2148 |
+
"learning_rate": 4.137035335939042e-06,
|
| 2149 |
+
"loss": 1.2093,
|
| 2150 |
+
"step": 3570
|
| 2151 |
+
},
|
| 2152 |
+
{
|
| 2153 |
+
"epoch": 0.78,
|
| 2154 |
+
"learning_rate": 4.05824889431575e-06,
|
| 2155 |
+
"loss": 1.2278,
|
| 2156 |
+
"step": 3580
|
| 2157 |
+
},
|
| 2158 |
+
{
|
| 2159 |
+
"epoch": 0.79,
|
| 2160 |
+
"learning_rate": 3.980102369608065e-06,
|
| 2161 |
+
"loss": 1.2187,
|
| 2162 |
+
"step": 3590
|
| 2163 |
+
},
|
| 2164 |
+
{
|
| 2165 |
+
"epoch": 0.79,
|
| 2166 |
+
"learning_rate": 3.902600332134067e-06,
|
| 2167 |
+
"loss": 1.2284,
|
| 2168 |
+
"step": 3600
|
| 2169 |
+
},
|
| 2170 |
+
{
|
| 2171 |
+
"epoch": 0.79,
|
| 2172 |
+
"learning_rate": 3.825747314519704e-06,
|
| 2173 |
+
"loss": 1.1992,
|
| 2174 |
+
"step": 3610
|
| 2175 |
+
},
|
| 2176 |
+
{
|
| 2177 |
+
"epoch": 0.79,
|
| 2178 |
+
"learning_rate": 3.749547811433639e-06,
|
| 2179 |
+
"loss": 1.1699,
|
| 2180 |
+
"step": 3620
|
| 2181 |
+
},
|
| 2182 |
+
{
|
| 2183 |
+
"epoch": 0.8,
|
| 2184 |
+
"learning_rate": 3.674006279324444e-06,
|
| 2185 |
+
"loss": 1.2103,
|
| 2186 |
+
"step": 3630
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 0.8,
|
| 2190 |
+
"learning_rate": 3.599127136159928e-06,
|
| 2191 |
+
"loss": 1.1975,
|
| 2192 |
+
"step": 3640
|
| 2193 |
+
},
|
| 2194 |
+
{
|
| 2195 |
+
"epoch": 0.8,
|
| 2196 |
+
"learning_rate": 3.5249147611687815e-06,
|
| 2197 |
+
"loss": 1.2151,
|
| 2198 |
+
"step": 3650
|
| 2199 |
+
},
|
| 2200 |
+
{
|
| 2201 |
+
"epoch": 0.8,
|
| 2202 |
+
"learning_rate": 3.4513734945844535e-06,
|
| 2203 |
+
"loss": 1.208,
|
| 2204 |
+
"step": 3660
|
| 2205 |
+
},
|
| 2206 |
+
{
|
| 2207 |
+
"epoch": 0.8,
|
| 2208 |
+
"learning_rate": 3.378507637391302e-06,
|
| 2209 |
+
"loss": 1.243,
|
| 2210 |
+
"step": 3670
|
| 2211 |
+
},
|
| 2212 |
+
{
|
| 2213 |
+
"epoch": 0.81,
|
| 2214 |
+
"learning_rate": 3.306321451073099e-06,
|
| 2215 |
+
"loss": 1.2255,
|
| 2216 |
+
"step": 3680
|
| 2217 |
+
},
|
| 2218 |
+
{
|
| 2219 |
+
"epoch": 0.81,
|
| 2220 |
+
"learning_rate": 3.234819157363744e-06,
|
| 2221 |
+
"loss": 1.2241,
|
| 2222 |
+
"step": 3690
|
| 2223 |
+
},
|
| 2224 |
+
{
|
| 2225 |
+
"epoch": 0.81,
|
| 2226 |
+
"learning_rate": 3.16400493800041e-06,
|
| 2227 |
+
"loss": 1.1896,
|
| 2228 |
+
"step": 3700
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 0.81,
|
| 2232 |
+
"learning_rate": 3.0938829344789373e-06,
|
| 2233 |
+
"loss": 1.2198,
|
| 2234 |
+
"step": 3710
|
| 2235 |
+
},
|
| 2236 |
+
{
|
| 2237 |
+
"epoch": 0.81,
|
| 2238 |
+
"learning_rate": 3.024457247811674e-06,
|
| 2239 |
+
"loss": 1.2019,
|
| 2240 |
+
"step": 3720
|
| 2241 |
+
},
|
| 2242 |
+
{
|
| 2243 |
+
"epoch": 0.82,
|
| 2244 |
+
"learning_rate": 2.9557319382875753e-06,
|
| 2245 |
+
"loss": 1.1925,
|
| 2246 |
+
"step": 3730
|
| 2247 |
+
},
|
| 2248 |
+
{
|
| 2249 |
+
"epoch": 0.82,
|
| 2250 |
+
"learning_rate": 2.887711025234783e-06,
|
| 2251 |
+
"loss": 1.2032,
|
| 2252 |
+
"step": 3740
|
| 2253 |
+
},
|
| 2254 |
+
{
|
| 2255 |
+
"epoch": 0.82,
|
| 2256 |
+
"learning_rate": 2.8203984867855366e-06,
|
| 2257 |
+
"loss": 1.2181,
|
| 2258 |
+
"step": 3750
|
| 2259 |
+
},
|
| 2260 |
+
{
|
| 2261 |
+
"epoch": 0.82,
|
| 2262 |
+
"learning_rate": 2.753798259643531e-06,
|
| 2263 |
+
"loss": 1.2471,
|
| 2264 |
+
"step": 3760
|
| 2265 |
+
},
|
| 2266 |
+
{
|
| 2267 |
+
"epoch": 0.83,
|
| 2268 |
+
"learning_rate": 2.6879142388536774e-06,
|
| 2269 |
+
"loss": 1.2515,
|
| 2270 |
+
"step": 3770
|
| 2271 |
+
},
|
| 2272 |
+
{
|
| 2273 |
+
"epoch": 0.83,
|
| 2274 |
+
"learning_rate": 2.6227502775742977e-06,
|
| 2275 |
+
"loss": 1.2309,
|
| 2276 |
+
"step": 3780
|
| 2277 |
+
},
|
| 2278 |
+
{
|
| 2279 |
+
"epoch": 0.83,
|
| 2280 |
+
"learning_rate": 2.5583101868517874e-06,
|
| 2281 |
+
"loss": 1.2277,
|
| 2282 |
+
"step": 3790
|
| 2283 |
+
},
|
| 2284 |
+
{
|
| 2285 |
+
"epoch": 0.83,
|
| 2286 |
+
"learning_rate": 2.4945977353977283e-06,
|
| 2287 |
+
"loss": 1.193,
|
| 2288 |
+
"step": 3800
|
| 2289 |
+
},
|
| 2290 |
+
{
|
| 2291 |
+
"epoch": 0.83,
|
| 2292 |
+
"learning_rate": 2.4316166493684783e-06,
|
| 2293 |
+
"loss": 1.1837,
|
| 2294 |
+
"step": 3810
|
| 2295 |
+
},
|
| 2296 |
+
{
|
| 2297 |
+
"epoch": 0.84,
|
| 2298 |
+
"learning_rate": 2.3693706121472404e-06,
|
| 2299 |
+
"loss": 1.2142,
|
| 2300 |
+
"step": 3820
|
| 2301 |
+
},
|
| 2302 |
+
{
|
| 2303 |
+
"epoch": 0.84,
|
| 2304 |
+
"learning_rate": 2.3078632641286653e-06,
|
| 2305 |
+
"loss": 1.203,
|
| 2306 |
+
"step": 3830
|
| 2307 |
+
},
|
| 2308 |
+
{
|
| 2309 |
+
"epoch": 0.84,
|
| 2310 |
+
"learning_rate": 2.247098202505933e-06,
|
| 2311 |
+
"loss": 1.1808,
|
| 2312 |
+
"step": 3840
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 0.84,
|
| 2316 |
+
"learning_rate": 2.1870789810603694e-06,
|
| 2317 |
+
"loss": 1.2221,
|
| 2318 |
+
"step": 3850
|
| 2319 |
+
},
|
| 2320 |
+
{
|
| 2321 |
+
"epoch": 0.85,
|
| 2322 |
+
"learning_rate": 2.1278091099536224e-06,
|
| 2323 |
+
"loss": 1.2447,
|
| 2324 |
+
"step": 3860
|
| 2325 |
+
},
|
| 2326 |
+
{
|
| 2327 |
+
"epoch": 0.85,
|
| 2328 |
+
"learning_rate": 2.0692920555223647e-06,
|
| 2329 |
+
"loss": 1.1919,
|
| 2330 |
+
"step": 3870
|
| 2331 |
+
},
|
| 2332 |
+
{
|
| 2333 |
+
"epoch": 0.85,
|
| 2334 |
+
"learning_rate": 2.0115312400755697e-06,
|
| 2335 |
+
"loss": 1.1988,
|
| 2336 |
+
"step": 3880
|
| 2337 |
+
},
|
| 2338 |
+
{
|
| 2339 |
+
"epoch": 0.85,
|
| 2340 |
+
"learning_rate": 1.954530041694355e-06,
|
| 2341 |
+
"loss": 1.1982,
|
| 2342 |
+
"step": 3890
|
| 2343 |
+
},
|
| 2344 |
+
{
|
| 2345 |
+
"epoch": 0.85,
|
| 2346 |
+
"learning_rate": 1.898291794034434e-06,
|
| 2347 |
+
"loss": 1.2069,
|
| 2348 |
+
"step": 3900
|
| 2349 |
+
},
|
| 2350 |
+
{
|
| 2351 |
+
"epoch": 0.86,
|
| 2352 |
+
"learning_rate": 1.8428197861311425e-06,
|
| 2353 |
+
"loss": 1.2017,
|
| 2354 |
+
"step": 3910
|
| 2355 |
+
},
|
| 2356 |
+
{
|
| 2357 |
+
"epoch": 0.86,
|
| 2358 |
+
"learning_rate": 1.7881172622070725e-06,
|
| 2359 |
+
"loss": 1.2166,
|
| 2360 |
+
"step": 3920
|
| 2361 |
+
},
|
| 2362 |
+
{
|
| 2363 |
+
"epoch": 0.86,
|
| 2364 |
+
"learning_rate": 1.7341874214823578e-06,
|
| 2365 |
+
"loss": 1.2365,
|
| 2366 |
+
"step": 3930
|
| 2367 |
+
},
|
| 2368 |
+
{
|
| 2369 |
+
"epoch": 0.86,
|
| 2370 |
+
"learning_rate": 1.6810334179875469e-06,
|
| 2371 |
+
"loss": 1.213,
|
| 2372 |
+
"step": 3940
|
| 2373 |
+
},
|
| 2374 |
+
{
|
| 2375 |
+
"epoch": 0.87,
|
| 2376 |
+
"learning_rate": 1.6286583603791772e-06,
|
| 2377 |
+
"loss": 1.2168,
|
| 2378 |
+
"step": 3950
|
| 2379 |
+
},
|
| 2380 |
+
{
|
| 2381 |
+
"epoch": 0.87,
|
| 2382 |
+
"learning_rate": 1.5770653117579265e-06,
|
| 2383 |
+
"loss": 1.2531,
|
| 2384 |
+
"step": 3960
|
| 2385 |
+
},
|
| 2386 |
+
{
|
| 2387 |
+
"epoch": 0.87,
|
| 2388 |
+
"learning_rate": 1.5262572894895021e-06,
|
| 2389 |
+
"loss": 1.2265,
|
| 2390 |
+
"step": 3970
|
| 2391 |
+
},
|
| 2392 |
+
{
|
| 2393 |
+
"epoch": 0.87,
|
| 2394 |
+
"learning_rate": 1.4762372650281602e-06,
|
| 2395 |
+
"loss": 1.2678,
|
| 2396 |
+
"step": 3980
|
| 2397 |
+
},
|
| 2398 |
+
{
|
| 2399 |
+
"epoch": 0.87,
|
| 2400 |
+
"learning_rate": 1.4270081637429293e-06,
|
| 2401 |
+
"loss": 1.2325,
|
| 2402 |
+
"step": 3990
|
| 2403 |
+
},
|
| 2404 |
+
{
|
| 2405 |
+
"epoch": 0.88,
|
| 2406 |
+
"learning_rate": 1.3785728647465128e-06,
|
| 2407 |
+
"loss": 1.1785,
|
| 2408 |
+
"step": 4000
|
| 2409 |
+
}
|
| 2410 |
+
],
|
| 2411 |
+
"max_steps": 4565,
|
| 2412 |
+
"num_train_epochs": 1,
|
| 2413 |
+
"total_flos": 0.0,
|
| 2414 |
+
"trial_name": null,
|
| 2415 |
+
"trial_params": null
|
| 2416 |
+
}
|
msa_curriculum/reset/pct50/finetuning/bytes/MSA_Dialect_Combined_1x/training_complete.json
ADDED
|
@@ -0,0 +1,19 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"dataset_path": "/root/baseline_cloud/work/prepared/msa_curriculum/train/bytes/MSA_Dialect_Combined_1x",
|
| 3 |
+
"dataset_rows": 584406,
|
| 4 |
+
"epochs": 1,
|
| 5 |
+
"global_step": 4565,
|
| 6 |
+
"input_checkpoint": "/root/baseline_cloud/work/checkpoints/msa_curriculum/reset/pct50/pretraining/bytes/full_epoch2/final/pytorch_model.bin",
|
| 7 |
+
"learning_rate": 3e-05,
|
| 8 |
+
"metrics": {
|
| 9 |
+
"epoch": 1.0,
|
| 10 |
+
"train_loss": 1.273271749105798,
|
| 11 |
+
"train_runtime": 3014.5799,
|
| 12 |
+
"train_samples_per_second": 193.86,
|
| 13 |
+
"train_steps_per_second": 1.514
|
| 14 |
+
},
|
| 15 |
+
"model_type": "bytes",
|
| 16 |
+
"scheduler": "cosine",
|
| 17 |
+
"seed": 42,
|
| 18 |
+
"warmup_ratio": 0.1
|
| 19 |
+
}
|
msa_curriculum/reset/pct50/manifest.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"created_at_utc": "2026-09-13T08:04:15.470657+00:00",
|
| 3 |
+
"datasets": {
|
| 4 |
+
"checked_train_indices": [
|
| 5 |
+
0,
|
| 6 |
+
1,
|
| 7 |
+
1717288,
|
| 8 |
+
6010510,
|
| 9 |
+
12021018,
|
| 10 |
+
12021019
|
| 11 |
+
],
|
| 12 |
+
"checked_validation_indices": [
|
| 13 |
+
0,
|
| 14 |
+
1999
|
| 15 |
+
],
|
| 16 |
+
"compact_path": "/root/baseline_cloud/work/data/byte_msa_compact_pct50",
|
| 17 |
+
"compact_train_fingerprint": "e30ae8e054884aad",
|
| 18 |
+
"raw_path": "/workspace/baseline_cloud/data/pixel_translation_MSA_filtered_32",
|
| 19 |
+
"raw_train_fingerprint": "7bad3a4e9c544674"
|
| 20 |
+
},
|
| 21 |
+
"percentage": 50,
|
| 22 |
+
"schedule": "reset",
|
| 23 |
+
"schema_version": 1,
|
| 24 |
+
"seed": 42,
|
| 25 |
+
"selection": {
|
| 26 |
+
"compact_max_percentage": 50,
|
| 27 |
+
"compact_train_rows": 12021020,
|
| 28 |
+
"method": "prefix_of_already_shuffled_filtered_MSA",
|
| 29 |
+
"original_train_rows": 24042041,
|
| 30 |
+
"selected_train_rows": 12021020
|
| 31 |
+
},
|
| 32 |
+
"training": {
|
| 33 |
+
"alignment_epochs": {
|
| 34 |
+
"bytes": 1,
|
| 35 |
+
"decoder": 0,
|
| 36 |
+
"vocab": 1
|
| 37 |
+
},
|
| 38 |
+
"effective_global_batch": 512,
|
| 39 |
+
"finetuning_mixture": "MSA_Dialect_Combined_1x",
|
| 40 |
+
"full_epochs": 2,
|
| 41 |
+
"learning_rate": 0.0003,
|
| 42 |
+
"models": [
|
| 43 |
+
"decoder",
|
| 44 |
+
"vocab",
|
| 45 |
+
"bytes"
|
| 46 |
+
],
|
| 47 |
+
"scheduler": "cosine",
|
| 48 |
+
"seed": 42,
|
| 49 |
+
"warmup_ratio": 0.1
|
| 50 |
+
}
|
| 51 |
+
}
|
msa_curriculum/reset/pct50/pretraining/bytes/alignment/checkpoint-20000/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d4fad895bb4e7f0fdc995399c1008c2c31c40d118a290e154b07b659ee61453b
|
| 3 |
+
size 971693050
|
msa_curriculum/reset/pct50/pretraining/bytes/alignment/checkpoint-20000/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a9b757b90fa971719464ce5fa5b1a38a25a67f075ae5b2ad31c2799b66f16d07
|
| 3 |
+
size 1933290270
|
msa_curriculum/reset/pct50/pretraining/bytes/alignment/checkpoint-20000/rng_state_0.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ae78313eb528c8d3695eebaf4de3539bd0a0bc6ee18c66af1ee183442f1758a0
|
| 3 |
+
size 15024
|
msa_curriculum/reset/pct50/pretraining/bytes/alignment/checkpoint-20000/rng_state_1.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1b38031f60d9e88601d369ef46bcdcf2b5b03f2cb4ba93853bcb2328df7ebb7c
|
| 3 |
+
size 15024
|
msa_curriculum/reset/pct50/pretraining/bytes/alignment/checkpoint-20000/rng_state_2.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f58092375c93d237cd0e3149aecfbf83e2acdae46279e07a32920d01cb507e64
|
| 3 |
+
size 15024
|
msa_curriculum/reset/pct50/pretraining/bytes/alignment/checkpoint-20000/rng_state_3.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:83cd4bbff9962da7ec6787fcea8d65df7096917f9a5902e249ba7aee8887fe5f
|
| 3 |
+
size 15024
|
msa_curriculum/reset/pct50/pretraining/bytes/alignment/checkpoint-20000/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:eacb5fe5dda8a27ae87613ffd1d4846f8a3e357fdd062ab6d40e11ee2c27460f
|
| 3 |
+
size 1064
|
msa_curriculum/reset/pct50/pretraining/bytes/alignment/checkpoint-20000/trainer_state.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
msa_curriculum/reset/pct50/pretraining/bytes/alignment/checkpoint-20000/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ccd9b76da328bb0c8b2b13fe2a9eec9c7fe7b8794c98eb4474a18da0709f17cc
|
| 3 |
+
size 4600
|
msa_curriculum/reset/pct50/pretraining/bytes/alignment/checkpoint-22000/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:00f73178d5e14ecd2f4ce4ca0875728232e1922b13bd41d0286f63ff53f4d687
|
| 3 |
+
size 971693050
|
msa_curriculum/reset/pct50/pretraining/bytes/alignment/checkpoint-22000/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4997d9d99ea5d69c736f2bde7bf5560c798d2bbc51114ab0bdae246c2a53b474
|
| 3 |
+
size 1933290270
|
msa_curriculum/reset/pct50/pretraining/bytes/alignment/checkpoint-22000/rng_state_0.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a81e3916b1392c4c49afb171dee5415c15f5a5a5af8749b28195fcfa0596699c
|
| 3 |
+
size 15024
|
msa_curriculum/reset/pct50/pretraining/bytes/alignment/checkpoint-22000/rng_state_1.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9a781038dd714b87b8adb1aac8dbc8217ceb607428a992133954ad522365236e
|
| 3 |
+
size 15024
|
msa_curriculum/reset/pct50/pretraining/bytes/alignment/checkpoint-22000/rng_state_2.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9446c3db15f382a5546f13622787fc99392a5e0bc8a9ca2da1838de7ab621a37
|
| 3 |
+
size 15024
|
msa_curriculum/reset/pct50/pretraining/bytes/alignment/checkpoint-22000/rng_state_3.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1f11e7a6b3faa884fc23044e3772ff9dd72c257f02e121665061e2a03d518bd9
|
| 3 |
+
size 15024
|
msa_curriculum/reset/pct50/pretraining/bytes/alignment/checkpoint-22000/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:50c36e10fd6a411914cecdd29029b8031583697291efdec87e433173663a171d
|
| 3 |
+
size 1064
|
msa_curriculum/reset/pct50/pretraining/bytes/alignment/checkpoint-22000/trainer_state.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
msa_curriculum/reset/pct50/pretraining/bytes/alignment/checkpoint-22000/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ccd9b76da328bb0c8b2b13fe2a9eec9c7fe7b8794c98eb4474a18da0709f17cc
|
| 3 |
+
size 4600
|
msa_curriculum/reset/pct50/pretraining/bytes/alignment/final/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:62e912f1efb9370d028c092e4c266b79ceaf685eb00f9077c290d3e52ba78ea8
|
| 3 |
+
size 1933294015
|
msa_curriculum/reset/pct50/pretraining/bytes/alignment/final/training_complete.json
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epochs": 1,
|
| 3 |
+
"global_step": 23479,
|
| 4 |
+
"initial_weights": null,
|
| 5 |
+
"learning_rate": 0.0003,
|
| 6 |
+
"metrics": {
|
| 7 |
+
"epoch": 1.0,
|
| 8 |
+
"train_loss": 2.878765146413173,
|
| 9 |
+
"train_runtime": 23214.6005,
|
| 10 |
+
"train_samples_per_second": 517.822,
|
| 11 |
+
"train_steps_per_second": 1.011
|
| 12 |
+
},
|
| 13 |
+
"model_type": "bytes",
|
| 14 |
+
"percentage": 50,
|
| 15 |
+
"scheduler": "cosine",
|
| 16 |
+
"seed": 42,
|
| 17 |
+
"stage": "alignment",
|
| 18 |
+
"subset_rows": 12021020,
|
| 19 |
+
"total_rows": 24042041,
|
| 20 |
+
"warmup_ratio": 0.1
|
| 21 |
+
}
|
msa_curriculum/reset/pct50/pretraining/bytes/full_epoch1/checkpoint-20000/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9e65bd60756e5db1159c4cd6830588b3f2947d2f74d62cc67cd4904a58f9c9ec
|
| 3 |
+
size 1933290270
|
msa_curriculum/reset/pct50/pretraining/bytes/full_epoch1/checkpoint-20000/rng_state_0.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ae78313eb528c8d3695eebaf4de3539bd0a0bc6ee18c66af1ee183442f1758a0
|
| 3 |
+
size 15024
|
msa_curriculum/reset/pct50/pretraining/bytes/full_epoch1/checkpoint-20000/rng_state_1.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1b38031f60d9e88601d369ef46bcdcf2b5b03f2cb4ba93853bcb2328df7ebb7c
|
| 3 |
+
size 15024
|
msa_curriculum/reset/pct50/pretraining/bytes/full_epoch1/checkpoint-20000/rng_state_2.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f58092375c93d237cd0e3149aecfbf83e2acdae46279e07a32920d01cb507e64
|
| 3 |
+
size 15024
|
msa_curriculum/reset/pct50/pretraining/bytes/full_epoch1/checkpoint-20000/rng_state_3.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:83cd4bbff9962da7ec6787fcea8d65df7096917f9a5902e249ba7aee8887fe5f
|
| 3 |
+
size 15024
|
msa_curriculum/reset/pct50/pretraining/bytes/full_epoch1/checkpoint-20000/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:eacb5fe5dda8a27ae87613ffd1d4846f8a3e357fdd062ab6d40e11ee2c27460f
|
| 3 |
+
size 1064
|
msa_curriculum/reset/pct50/pretraining/bytes/full_epoch1/checkpoint-20000/trainer_state.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
msa_curriculum/reset/pct50/pretraining/bytes/full_epoch1/checkpoint-20000/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e3fa413c6681b80d764d7fef0f0b8064f2dac22d434af17914771944522b591f
|
| 3 |
+
size 4600
|
msa_curriculum/reset/pct50/pretraining/bytes/full_epoch1/checkpoint-22000/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e08e192526c1c189dfb9f55b8ddf00b87834d751929f856ccdb8c0af4a6952b8
|
| 3 |
+
size 1933290270
|
msa_curriculum/reset/pct50/pretraining/bytes/full_epoch1/checkpoint-22000/rng_state_0.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a81e3916b1392c4c49afb171dee5415c15f5a5a5af8749b28195fcfa0596699c
|
| 3 |
+
size 15024
|
msa_curriculum/reset/pct50/pretraining/bytes/full_epoch1/checkpoint-22000/rng_state_1.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9a781038dd714b87b8adb1aac8dbc8217ceb607428a992133954ad522365236e
|
| 3 |
+
size 15024
|
msa_curriculum/reset/pct50/pretraining/bytes/full_epoch1/checkpoint-22000/rng_state_2.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9446c3db15f382a5546f13622787fc99392a5e0bc8a9ca2da1838de7ab621a37
|
| 3 |
+
size 15024
|
msa_curriculum/reset/pct50/pretraining/bytes/full_epoch1/checkpoint-22000/rng_state_3.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1f11e7a6b3faa884fc23044e3772ff9dd72c257f02e121665061e2a03d518bd9
|
| 3 |
+
size 15024
|
msa_curriculum/reset/pct50/pretraining/bytes/full_epoch1/checkpoint-22000/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:50c36e10fd6a411914cecdd29029b8031583697291efdec87e433173663a171d
|
| 3 |
+
size 1064
|
msa_curriculum/reset/pct50/pretraining/bytes/full_epoch1/checkpoint-22000/trainer_state.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
msa_curriculum/reset/pct50/pretraining/bytes/full_epoch1/checkpoint-22000/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e3fa413c6681b80d764d7fef0f0b8064f2dac22d434af17914771944522b591f
|
| 3 |
+
size 4600
|
msa_curriculum/reset/pct50/pretraining/bytes/full_epoch1/final/training_complete.json
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epochs": 1,
|
| 3 |
+
"global_step": 23479,
|
| 4 |
+
"initial_weights": "/root/baseline_cloud/work/checkpoints/msa_curriculum/reset/pct50/pretraining/bytes/alignment/final/pytorch_model.bin",
|
| 5 |
+
"learning_rate": 0.0003,
|
| 6 |
+
"metrics": {
|
| 7 |
+
"epoch": 1.0,
|
| 8 |
+
"train_loss": 1.1009242805053003,
|
| 9 |
+
"train_runtime": 25724.6077,
|
| 10 |
+
"train_samples_per_second": 467.297,
|
| 11 |
+
"train_steps_per_second": 0.913
|
| 12 |
+
},
|
| 13 |
+
"model_type": "bytes",
|
| 14 |
+
"percentage": 50,
|
| 15 |
+
"scheduler": "cosine",
|
| 16 |
+
"seed": 42,
|
| 17 |
+
"stage": "full",
|
| 18 |
+
"subset_rows": 12021020,
|
| 19 |
+
"total_rows": 24042041,
|
| 20 |
+
"warmup_ratio": 0.1
|
| 21 |
+
}
|
msa_curriculum/reset/pct50/pretraining/bytes/full_epoch2/checkpoint-20000/rng_state_2.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f58092375c93d237cd0e3149aecfbf83e2acdae46279e07a32920d01cb507e64
|
| 3 |
+
size 15024
|
msa_curriculum/reset/pct50/pretraining/bytes/full_epoch2/checkpoint-20000/rng_state_3.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:83cd4bbff9962da7ec6787fcea8d65df7096917f9a5902e249ba7aee8887fe5f
|
| 3 |
+
size 15024
|
msa_curriculum/reset/pct50/pretraining/bytes/full_epoch2/checkpoint-20000/trainer_state.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
msa_curriculum/reset/pct50/pretraining/bytes/full_epoch2/checkpoint-22000/trainer_state.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
msa_curriculum/reset/pct50/pretraining/bytes/full_epoch2/final/training_complete.json
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epochs": 1,
|
| 3 |
+
"global_step": 23479,
|
| 4 |
+
"initial_weights": "/root/baseline_cloud/work/checkpoints/msa_curriculum/reset/pct50/pretraining/bytes/full_epoch1/final/pytorch_model.bin",
|
| 5 |
+
"learning_rate": 0.0003,
|
| 6 |
+
"metrics": {
|
| 7 |
+
"epoch": 1.0,
|
| 8 |
+
"train_loss": 0.8970891325451141,
|
| 9 |
+
"train_runtime": 25698.5965,
|
| 10 |
+
"train_samples_per_second": 467.77,
|
| 11 |
+
"train_steps_per_second": 0.914
|
| 12 |
+
},
|
| 13 |
+
"model_type": "bytes",
|
| 14 |
+
"percentage": 50,
|
| 15 |
+
"scheduler": "cosine",
|
| 16 |
+
"seed": 42,
|
| 17 |
+
"stage": "full",
|
| 18 |
+
"subset_rows": 12021020,
|
| 19 |
+
"total_rows": 24042041,
|
| 20 |
+
"warmup_ratio": 0.1
|
| 21 |
+
}
|
msa_curriculum/reset/pct50/run.lock
ADDED
|
File without changes
|