learning_rate: 5e-05
train_batch_size: 4
eval_batch_size: 1
seed: 42
gradient_accumulation_steps: 8
total_train_batch_size: 32
optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
num_epochs: 100

Training Loss	Epoch	Step	Validation Loss	Chrf++	Gen Len
10.9286	1.9888	222	0.7423	17.5788	69.4427
5.4170	3.9708	444	0.4620	31.0094	311.5393
3.3034	5.9528	666	0.1601	70.7141	136.5034
1.3585	7.9348	888	0.0870	83.1232	131.9573
0.6166	9.9169	1110	0.0697	85.8737	132.9573
0.4454	11.8989	1332	0.0668	87.2833	132.1124
0.3387	13.8809	1554	0.0826	86.5405	133.6
0.2709	15.8629	1776	0.0672	88.5306	131.791
0.2322	17.8449	1998	0.0671	88.7049	132.2427
0.1853	19.8270	2220	0.0726	88.7893	133.2404
0.1722	21.8090	2442	0.0732	88.8946	134.836
0.1461	23.7910	2664	0.0782	88.7558	132.836
0.1433	25.7730	2886	0.0787	88.5772	132.3169
0.1275	27.7551	3108	0.0758	89.1351	133.1865
0.1216	29.7371	3330	0.0770	89.1286	132.5146
0.1168	31.7191	3552	0.0820	87.9172	134.9775
0.1142	33.7011	3774	0.0900	88.0088	132.6157

Tsedeniya/finetuned_byT5base_ddo

finetuned_byT5base_ddo