learning_rate: 0.0001
train_batch_size: 4
eval_batch_size: 4
seed: 42
gradient_accumulation_steps: 8
total_train_batch_size: 32
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_steps: 5
num_epochs: 12

Training Loss	Epoch	Step	Validation Loss
2.4835	0.3094	50	1.6561
1.2918	0.6187	100	1.2015
1.1399	0.9281	150	1.1764
1.0881	1.2351	200	1.1594
1.1221	1.5445	250	1.1659
1.0781	1.8538	300	1.1377
1.0854	2.1609	350	1.1528
1.0849	2.4702	400	1.1304
1.0537	2.7796	450	1.1472
1.0851	3.0866	500	1.1253
1.0725	3.3960	550	1.1166
1.0639	3.7053	600	1.1616
1.0308	4.0124	650	1.1127
1.0616	4.3217	700	1.1094
1.0225	4.6311	750	1.1327
1.0066	4.9404	800	1.1188
1.009	5.2475	850	1.1141
1.0608	5.5568	900	1.1366
1.0516	5.8662	950	1.1308
1.0289	6.1732	1000	1.1256
1.1054	6.4826	1050	1.1159
1.025	6.7920	1100	1.1071
1.0376	7.0990	1150	1.1271
1.0389	7.4084	1200	1.1187
0.9929	7.7177	1250	1.1130
1.0692	8.0247	1300	1.1123
1.0505	8.3341	1350	1.1023
1.0437	8.6435	1400	1.1082
1.0228	8.9528	1450	1.1056
1.0134	9.2599	1500	1.1092
1.073	9.5692	1550	1.1155
1.0354	9.8786	1600	1.1081
1.0316	10.1856	1650	1.1100
1.0558	10.4950	1700	1.1075
0.9927	10.8043	1750	1.1089
1.0543	11.1114	1800	1.1105
1.0364	11.4207	1850	1.1088
1.0407	11.7301	1900	1.1090

Toonies/medsiglip-448-ft-crc100k

medsiglip-448-ft-crc100k