learning_rate: 5e-06
train_batch_size: 8
eval_batch_size: 16
seed: 42
gradient_accumulation_steps: 4
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
lr_scheduler_warmup_ratio: 0.1
num_epochs: 10
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Bertscore Precision	Bertscore Recall	Bertscore F1	Gen Len
2.1943	1.0	487	1.2857	0.2067	0.1381	0.1971	0.1969	0.7842	0.6609	0.7164	21.0
1.9194	2.0	974	1.2561	0.2102	0.1423	0.2015	0.2013	0.7874	0.6616	0.7182	21.0
1.7597	3.0	1461	1.2300	0.2099	0.1436	0.2017	0.2015	0.7884	0.6613	0.7185	21.0
1.7099	4.0	1948	1.1986	0.208	0.1413	0.2002	0.1999	0.7875	0.6612	0.718	21.0
1.676	5.0	2435	1.1750	0.2073	0.1407	0.1994	0.1991	0.7868	0.6609	0.7175	21.0

aku47z/mt5-small-nepali-v3.1

mt5-small-nepali-v3.1