learning_rate: 0.0001
train_batch_size: 2
eval_batch_size: 1
seed: 42
distributed_type: multi-GPU
num_devices: 8
gradient_accumulation_steps: 4
total_train_batch_size: 64
total_eval_batch_size: 8
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.95) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: cosine
lr_scheduler_warmup_ratio: 0.05
num_epochs: 3.0

Training Loss	Epoch	Step	Validation Loss	Token Acc
0.9031	0.1139	50	0.9397	0.7254
0.9153	0.2278	100	0.9367	0.7296
0.8993	0.3417	150	0.9187	0.7346
0.8685	0.4556	200	0.9020	0.7375
0.8662	0.5695	250	0.8845	0.7410
0.8376	0.6834	300	0.8662	0.7451
0.8248	0.7973	350	0.8541	0.7477
0.8197	0.9112	400	0.8303	0.7540
0.7581	1.0251	450	0.8194	0.7561
0.7390	1.1390	500	0.8065	0.7597
0.7164	1.2528	550	0.7953	0.7625
0.7048	1.3667	600	0.7755	0.7664
0.6950	1.4806	650	0.7629	0.7698
0.6906	1.5945	700	0.7444	0.7738
0.6770	1.7084	750	0.7308	0.7767
0.6378	1.8223	800	0.7189	0.7803
0.6497	1.9362	850	0.7050	0.7826
0.5585	2.0501	900	0.6995	0.7853
0.5512	2.1640	950	0.6899	0.7875
0.5506	2.2779	1000	0.6822	0.7905
0.5456	2.3918	1050	0.6711	0.7924
0.5417	2.5057	1100	0.6648	0.7936
0.5179	2.6196	1150	0.6599	0.7948
0.5268	2.7335	1200	0.6574	0.7958
0.5392	2.8474	1250	0.6552	0.7960
0.5345	2.9613	1300	0.6549	0.7961
0.5306	3.0	1317	0.6548	0.7961

boyuzhuGPT/qwen2_5_omni_all_1015_3b

qwen2_5_omni_all_1015_3b