TWON-Agent-OSN-Replies-de

This model is a fine-tuned version of meta-llama/Llama-3.2-3B-Instruct on the generator dataset. It achieves the following results on the evaluation set:

• Loss: 0.5430

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: 5e-05
• train_batch_size: 4
• eval_batch_size: 8
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 8

Training results

Training Loss	Epoch	Step	Validation Loss
2.5602	0.0423	200	2.3015
2.1761	0.0845	400	2.0341
1.8998	0.1268	600	1.7808
1.6962	0.1691	800	1.6139
1.5494	0.2114	1000	1.4910
1.4478	0.2536	1200	1.4047
1.3865	0.2959	1400	1.3415
1.3407	0.3382	1600	1.2891
1.2808	0.3805	1800	1.2429
1.2365	0.4227	2000	1.2054
1.2058	0.4650	2200	1.1766
1.1799	0.5073	2400	1.1484
1.147	0.5496	2600	1.1282
1.1235	0.5918	2800	1.1087
1.1228	0.6341	3000	1.0923
1.0935	0.6764	3200	1.0765
1.0813	0.7187	3400	1.0629
1.0671	0.7609	3600	1.0504
1.0448	0.8032	3800	1.0390
1.041	0.8455	4000	1.0293
1.0167	0.8878	4200	1.0196
1.0293	0.9300	4400	1.0091
1.0254	0.9723	4600	1.0002
1.0053	1.0146	4800	0.9917
0.9957	1.0569	5000	0.9833
0.9976	1.0991	5200	0.9772
0.9891	1.1414	5400	0.9708
0.9648	1.1837	5600	0.9622
0.9711	1.2260	5800	0.9551
0.9569	1.2682	6000	0.9497
0.9526	1.3105	6200	0.9433
0.9571	1.3528	6400	0.9362
0.9265	1.3951	6600	0.9310
0.9274	1.4373	6800	0.9229
0.929	1.4796	7000	0.9190
0.9224	1.5219	7200	0.9128
0.9051	1.5642	7400	0.9071
0.9066	1.6064	7600	0.9014
0.9067	1.6487	7800	0.8962
0.8986	1.6910	8000	0.8905
0.8967	1.7332	8200	0.8847
0.8883	1.7755	8400	0.8806
0.8844	1.8178	8600	0.8745
0.8833	1.8601	8800	0.8710
0.8805	1.9023	9000	0.8649
0.8722	1.9446	9200	0.8600
0.868	1.9869	9400	0.8552
0.8597	2.0292	9600	0.8499
0.8527	2.0714	9800	0.8456
0.8444	2.1137	10000	0.8410
0.8472	2.1560	10200	0.8363
0.8416	2.1983	10400	0.8321
0.8395	2.2405	10600	0.8271
0.8295	2.2828	10800	0.8221
0.8307	2.3251	11000	0.8173
0.8221	2.3674	11200	0.8130
0.8144	2.4096	11400	0.8086
0.8197	2.4519	11600	0.8048
0.8121	2.4942	11800	0.8003
0.8038	2.5365	12000	0.7953
0.8007	2.5787	12200	0.7917
0.808	2.6210	12400	0.7878
0.8047	2.6633	12600	0.7827
0.7908	2.7056	12800	0.7795
0.7963	2.7478	13000	0.7755
0.782	2.7901	13200	0.7718
0.7899	2.8324	13400	0.7664
0.7712	2.8747	13600	0.7623
0.7737	2.9169	13800	0.7581
0.784	2.9592	14000	0.7553
0.7661	3.0015	14200	0.7515
0.759	3.0438	14400	0.7478
0.7534	3.0860	14600	0.7442
0.7526	3.1283	14800	0.7398
0.7526	3.1706	15000	0.7365
0.7413	3.2129	15200	0.7344
0.746	3.2551	15400	0.7296
0.7469	3.2974	15600	0.7264
0.7384	3.3397	15800	0.7221
0.7357	3.3819	16000	0.7191
0.7298	3.4242	16200	0.7173
0.7245	3.4665	16400	0.7122
0.7283	3.5088	16600	0.7087
0.7333	3.5510	16800	0.7062
0.7252	3.5933	17000	0.7040
0.7242	3.6356	17200	0.6987
0.7174	3.6779	17400	0.6956
0.7132	3.7201	17600	0.6931
0.7093	3.7624	17800	0.6898
0.7027	3.8047	18000	0.6869
0.7177	3.8470	18200	0.6838
0.707	3.8892	18400	0.6805
0.7091	3.9315	18600	0.6786
0.7031	3.9738	18800	0.6749
0.6913	4.0161	19000	0.6723
0.6895	4.0583	19200	0.6697
0.6858	4.1006	19400	0.6666
0.678	4.1429	19600	0.6645
0.6852	4.1852	19800	0.6622
0.6787	4.2274	20000	0.6586
0.6784	4.2697	20200	0.6568
0.6771	4.3120	20400	0.6528
0.6697	4.3543	20600	0.6509
0.6698	4.3965	20800	0.6481
0.6792	4.4388	21000	0.6455
0.6741	4.4811	21200	0.6436
0.6582	4.5234	21400	0.6402
0.6648	4.5656	21600	0.6380
0.6606	4.6079	21800	0.6363
0.6598	4.6502	22000	0.6341
0.6696	4.6925	22200	0.6312
0.6604	4.7347	22400	0.6298
0.6611	4.7770	22600	0.6274
0.6515	4.8193	22800	0.6260
0.6528	4.8616	23000	0.6228
0.6557	4.9038	23200	0.6201
0.6473	4.9461	23400	0.6184
0.6506	4.9884	23600	0.6168
0.6387	5.0306	23800	0.6146
0.638	5.0729	24000	0.6139
0.6389	5.1152	24200	0.6111
0.641	5.1575	24400	0.6103
0.6278	5.1997	24600	0.6080
0.6332	5.2420	24800	0.6068
0.6214	5.2843	25000	0.6047
0.6325	5.3266	25200	0.6020
0.6312	5.3688	25400	0.6000
0.6278	5.4111	25600	0.5979
0.6237	5.4534	25800	0.5962
0.6263	5.4957	26000	0.5942
0.6228	5.5379	26200	0.5938
0.625	5.5802	26400	0.5919
0.6271	5.6225	26600	0.5905
0.6206	5.6648	26800	0.5880
0.6204	5.7070	27000	0.5872
0.617	5.7493	27200	0.5857
0.6138	5.7916	27400	0.5844
0.6216	5.8339	27600	0.5817
0.6122	5.8761	27800	0.5805
0.6186	5.9184	28000	0.5793
0.6164	5.9607	28200	0.5783
0.6117	6.0030	28400	0.5769
0.6068	6.0452	28600	0.5749
0.6034	6.0875	28800	0.5740
0.6085	6.1298	29000	0.5720
0.6068	6.1721	29200	0.5710
0.603	6.2143	29400	0.5697
0.5979	6.2566	29600	0.5694
0.5941	6.2989	29800	0.5686
0.6022	6.3412	30000	0.5673
0.5999	6.3834	30200	0.5656
0.6064	6.4257	30400	0.5645
0.5982	6.4680	30600	0.5633
0.5863	6.5103	30800	0.5630
0.5942	6.5525	31000	0.5619
0.5969	6.5948	31200	0.5607
0.595	6.6371	31400	0.5600
0.6004	6.6793	31600	0.5583
0.593	6.7216	31800	0.5584
0.5903	6.7639	32000	0.5579
0.5917	6.8062	32200	0.5568
0.5946	6.8484	32400	0.5563
0.5876	6.8907	32600	0.5560
0.593	6.9330	32800	0.5543
0.5848	6.9753	33000	0.5527
0.5865	7.0175	33200	0.5523
0.5872	7.0598	33400	0.5524
0.583	7.1021	33600	0.5511
0.5838	7.1444	33800	0.5512
0.5819	7.1866	34000	0.5502
0.5833	7.2289	34200	0.5492
0.5826	7.2712	34400	0.5493
0.5813	7.3135	34600	0.5487
0.5815	7.3557	34800	0.5477
0.582	7.3980	35000	0.5472
0.5729	7.4403	35200	0.5466
0.5769	7.4826	35400	0.5457
0.5864	7.5248	35600	0.5457
0.5853	7.5671	35800	0.5453
0.5771	7.6094	36000	0.5449
0.5786	7.6517	36200	0.5447
0.5839	7.6939	36400	0.5443
0.5759	7.7362	36600	0.5435
0.5804	7.7785	36800	0.5436
0.5826	7.8208	37000	0.5437
0.5829	7.8630	37200	0.5434
0.574	7.9053	37400	0.5431
0.5756	7.9476	37600	0.5432
0.5722	7.9899	37800	0.5430

Framework versions

• PEFT 0.12.0
• Transformers 4.47.1
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0