learning_rate: 5e-05
train_batch_size: 512
eval_batch_size: 512
seed: 42
optimizer: Use adamw_torch with betas=(0.9,0.95) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 78125
num_epochs: 9
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss
1.5933	0.1888	500	1.5166
1.3224	0.3775	1000	1.2561
1.1864	0.5663	1500	1.1098
1.0794	0.7550	2000	0.9943
0.9888	0.9438	2500	0.8991
0.9136	1.1325	3000	0.8287
0.8615	1.3213	3500	0.7879
0.8248	1.5100	4000	0.7525
0.7865	1.6988	4500	0.7258
0.7581	1.8875	5000	0.6998
0.7368	2.0763	5500	0.6856
0.7152	2.2650	6000	0.6644
0.701	2.4538	6500	0.6521
0.6883	2.6425	7000	0.6419
0.6732	2.8313	7500	0.6280
0.6645	3.0200	8000	0.6186
0.6551	3.2088	8500	0.6079
0.6472	3.3975	9000	0.6030
0.6407	3.5863	9500	0.5939
0.6289	3.7750	10000	0.5869
0.6185	3.9638	10500	0.5830
0.6219	4.1525	11000	0.5780
0.6102	4.3413	11500	0.5731
0.6115	4.5300	12000	0.5675
0.6023	4.7188	12500	0.5638
0.5937	4.9075	13000	0.5596
0.5928	5.0963	13500	0.5565
0.5874	5.2850	14000	0.5533
0.5822	5.4738	14500	0.5492
0.5782	5.6625	15000	0.5481
0.577	5.8513	15500	0.5445
0.5741	6.0400	16000	0.5413
0.5715	6.2288	16500	0.5380
0.5674	6.4175	17000	0.5365
0.5708	6.6063	17500	0.5344
0.5634	6.7950	18000	0.5323
0.5616	6.9838	18500	0.5313
0.5625	7.1725	19000	0.5285
0.5547	7.3613	19500	0.5270
0.5497	7.5500	20000	0.5257
0.5551	7.7388	20500	0.5234
0.5494	7.9275	21000	0.5211
0.5444	8.1163	21500	0.5192
0.547	8.3050	22000	0.5179
0.5415	8.4938	22500	0.5163
0.5414	8.6825	23000	0.5147
0.5397	8.8713	23500	0.5137

suku9/pretrained_dl_05_17

pretrained_dl_05_17