Regarding selecting hyperparameters during training

[SoumyaCYZ]

Señor @RaulPPelaez @AdriaPerezCulubret @peastman @stefdoerr @guillemsimeon I'm using the following parameters during training of alanine-dipeptide. But after training, I'm getting some crazy ckpt files. Are these hyperparameters are fine?
rw------- 1 cyz218385 cyz21 3.5M Feb 4 20:00 epoch=1-val_loss=28354817745406260133494784.0000-test_loss=0.0000.ckpt
-rw------- 1 cyz218385 cyz21 3.5M Feb 4 20:01 epoch=3-val_loss=28354817745406260133494784.0000-test_loss=4108202541056.0000.ckpt
-rw------- 1 cyz218385 cyz21 3.5M Feb 4 20:03 epoch=5-val_loss=28354817745406260133494784.0000-test_loss=4108202541056.0000.ckpt
-rw------- 1 cyz218385 cyz21 3.5M Feb 4 20:04 epoch=7-val_loss=28354817745406260133494784.0000-test_loss=4108202541056.0000.ckpt
-rw------- 1 cyz218385 cyz21 3.5M Feb 4 20:05 epoch=9-val_loss=28354817745406260133494784.0000-test_loss=4108202541056.0000.ckpt
-rw------- 1 cyz218385 cyz21 3.5M Feb 4 20:07 epoch=11-val_loss=28354817745406260133494784.0000-test_loss=4108202541056.0000.ckpt
-rw------- 1 cyz218385 cyz21 3.5M Feb 4 20:08 epoch=13-val_loss=28354817745406260133494784.0000-test_loss=4108202541056.0000.ckpt
-rw------- 1 cyz218385 cyz21 3.5M Feb 4 20:10 epoch=15-val_loss=28354817745406260133494784.0000-test_loss=4108202541056.0000.ckpt
-rw------- 1 cyz218385 cyz21 3.5M Feb 4 20:11 epoch=17-val_loss=28354817745406260133494784.0000-test_loss=4108202541056.0000.ckpt
-rw------- 1 cyz218385 cyz21 3.5M Feb 4 20:13 epoch=19-val_loss=28354817745406260133494784.0000-test_loss=4108202541056.0000.ckpt

TRAIN.YAML
batch_size: 256
#inference_batchsize: 256
dataset: Custom
coord_files: "31jan_coords.npy"
embed_files: "31jan_ca_embeddings.npy"
force_files: "31jan_ca_deltaforces.npy"
cutoff_upper: 12.0
cutoff_lower: 3.0
log_dir: /home/chemistry/phd/cyz218385/scratch/aladi_wat/aladi_300k
derivative: true
#distributed_backend: ddp
early_stopping_patience: 30
embedding_dimension: 128

#label:
#- forces
lr: 0.0005
lr_factor: 0.8
lr_min: 1.0e-06
lr_patience: 10
lr_warmup_steps: 0
model: graph-network
neighbor_embedding: false
ngpus: -1
num_epochs: 100
num_layers: 4
num_nodes: 1
num_rbf: 18
num_workers: 8
rbf_type: expnorm
save_interval: 2
seed: 1
test_interval: 2
test_size: 80
trainable_rbf: true
val_size: 20
weight_decay: 0.0