learning_rate: 0.0005
train_batch_size: 4
eval_batch_size: 2
seed: 1
optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: cosine
lr_scheduler_warmup_ratio: 0.01
num_epochs: 10

Training Loss	Epoch	Step	Validation Loss
0.6287	0.3319	300	0.6577
0.6573	0.6637	600	0.6499
0.6555	0.9956	900	0.6492
0.6367	1.3274	1200	0.6387
0.6376	1.6593	1500	0.6290
0.6321	1.9912	1800	0.6259
0.6125	2.3230	2100	0.6171
0.6129	2.6549	2400	0.6081
0.6055	2.9867	2700	0.5928
0.5802	3.3186	3000	0.5890
0.5831	3.6504	3300	0.5779
0.5735	3.9823	3600	0.5668
0.5448	4.3142	3900	0.5566
0.5443	4.6460	4200	0.5482
0.5357	4.9779	4500	0.5364
0.5068	5.3097	4800	0.5340
0.5028	5.6416	5100	0.5210
0.497	5.9735	5400	0.5125
0.4683	6.3053	5700	0.5069
0.4657	6.6372	6000	0.4984
0.4589	6.9690	6300	0.4911
0.4336	7.3009	6600	0.4904
0.4319	7.6327	6900	0.4849
0.4274	7.9646	7200	0.4802
0.4106	8.2965	7500	0.4808
0.4093	8.6283	7800	0.4775
0.4059	8.9602	8100	0.4760
0.3974	9.2920	8400	0.4755
0.3969	9.6239	8700	0.4756
0.3928	9.9558	9000	0.4755

Felita/midi_nGram_model_1

midi_nGram_model_1