learning_rate: 3e-05
train_batch_size: 16
eval_batch_size: 16
seed: 42
distributed_type: multi-GPU
num_devices: 2
gradient_accumulation_steps: 2
total_train_batch_size: 64
total_eval_batch_size: 32
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 0.03
num_epochs: 3
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss
1.3378	0.2863	500	1.2903
1.2281	0.5726	1000	1.1914
1.1361	0.8589	1500	1.1385
0.9816	1.1449	2000	1.1079
0.9342	1.4311	2500	1.0880
0.9271	1.7174	3000	1.0697
0.9189	2.0034	3500	1.0578
0.7955	2.2897	4000	1.0619
0.8140	2.5760	4500	1.0589
0.8316	2.8623	5000	1.0573

VehaanS/slm-125m-tutor

slm-125m-tutor