Modelpublic

zhanxing/ours-crag-movie-sft-v2

sourceHugging Faceotherupdated 22d agoView on Hugging Face

0likes168downloads

Model Card

ours-crag-movie-sft-v2

This model is a fine-tuned version of Qwen/Qwen3-8B on the ourscragmoviesftv2_train dataset. It achieves the following results on the evaluation set:

—Loss: 0.0837

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

—learning_rate: 1e-05
—trainbatchsize: 1
—evalbatchsize: 1
—seed: 42
—distributed_type: multi-GPU
—num_devices: 3
—gradientaccumulationsteps: 10
—totaltrainbatch_size: 30
—totalevalbatch_size: 3
—optimizer: Use adamwtorch with betas=(0.9,0.999) and epsilon=1e-08 and optimizerargs=No additional optimizer arguments
—lrschedulertype: cosine
—lrschedulerwarmup_ratio: 0.1
—num_epochs: 2.0

Training results

Training Loss	Epoch	Step	Validation Loss
2.0056	0.0344	10	1.6154
1.1939	0.0688	20	0.7725
0.6289	0.1032	30	0.4738
0.4165	0.1376	40	0.3359
0.3172	0.1721	50	0.2646
0.2793	0.2065	60	0.2453
0.2499	0.2409	70	0.2251
0.242	0.2753	80	0.2170
0.2437	0.3097	90	0.2050
0.2272	0.3441	100	0.1981
0.2302	0.3785	110	0.1906
0.2155	0.4129	120	0.1850
0.2106	0.4474	130	0.1797
0.2097	0.4818	140	0.1757
0.1859	0.5162	150	0.1705
0.2094	0.5506	160	0.1663
0.1945	0.5850	170	0.1610
0.2102	0.6194	180	0.1597
0.1708	0.6538	190	0.1540
0.1822	0.6882	200	0.1542
0.1736	0.7226	210	0.1501
0.1779	0.7571	220	0.1450
0.1621	0.7915	230	0.1442
0.1548	0.8259	240	0.1383
0.1469	0.8603	250	0.1354
0.1492	0.8947	260	0.1334
0.1398	0.9291	270	0.1299
0.1542	0.9635	280	0.1271
0.1393	0.9979	290	0.1239
0.109	1.0310	300	0.1242
0.0924	1.0654	310	0.1192
0.0946	1.0998	320	0.1187
0.0996	1.1342	330	0.1164
0.0984	1.1686	340	0.1147
0.0887	1.2030	350	0.1112
0.0884	1.2374	360	0.1097
0.0912	1.2719	370	0.1066
0.0816	1.3063	380	0.1065
0.0902	1.3407	390	0.1045
0.0974	1.3751	400	0.1014
0.0829	1.4095	410	0.0987
0.0838	1.4439	420	0.0974
0.09	1.4783	430	0.0955
0.0807	1.5127	440	0.0950
0.0766	1.5471	450	0.0934
0.0726	1.5816	460	0.0921
0.0728	1.6160	470	0.0903
0.0721	1.6504	480	0.0889
0.0708	1.6848	490	0.0878
0.0784	1.7192	500	0.0867
0.0689	1.7536	510	0.0856
0.0752	1.7880	520	0.0848
0.0697	1.8224	530	0.0844
0.0679	1.8568	540	0.0840
0.07	1.8913	550	0.0838
0.0671	1.9257	560	0.0837
0.0648	1.9601	570	0.0837
0.0646	1.9945	580	0.0837

Framework versions

—Transformers 4.51.3
—Pytorch 2.10.0+cu128
—Datasets 3.5.0
—Tokenizers 0.21.1