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: 1.3698

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
2.5673	0.0402	200	2.3212
2.1646	0.0805	400	2.0731
1.8986	0.1207	600	1.8557
1.6862	0.1609	800	1.6990
1.5341	0.2012	1000	1.5808
1.426	0.2414	1200	1.4958
1.3729	0.2816	1400	1.4311
1.2906	0.3219	1600	1.3739
1.2528	0.3621	1800	1.3325
1.2179	0.4023	2000	1.3026
1.1649	0.4426	2200	1.2771
1.1465	0.4828	2400	1.2520
1.1122	0.5230	2600	1.2309
1.0963	0.5633	2800	1.2096
1.0867	0.6035	3000	1.1950
1.0632	0.6437	3200	1.1824
1.0512	0.6840	3400	1.1730
1.0394	0.7242	3600	1.1646
1.0317	0.7644	3800	1.1513
1.0163	0.8047	4000	1.1481
0.9996	0.8449	4200	1.1462
0.9883	0.8851	4400	1.1372
0.9739	0.9254	4600	1.1317
0.9767	0.9656	4800	1.1267
0.9674	1.0058	5000	1.1256
0.9479	1.0461	5200	1.1318
0.9393	1.0863	5400	1.1232
0.9341	1.1265	5600	1.1241
0.9131	1.1668	5800	1.1207
0.9191	1.2070	6000	1.1190
0.9118	1.2472	6200	1.1219
0.8974	1.2875	6400	1.1192
0.8863	1.3277	6600	1.1180
0.8884	1.3679	6800	1.1206
0.871	1.4082	7000	1.1187
0.8677	1.4484	7200	1.1259
0.8692	1.4886	7400	1.1233
0.8537	1.5289	7600	1.1244
0.8559	1.5691	7800	1.1234
0.8577	1.6093	8000	1.1294
0.8298	1.6496	8200	1.1306
0.8444	1.6898	8400	1.1265
0.83	1.7300	8600	1.1315
0.8183	1.7703	8800	1.1368
0.8199	1.8105	9000	1.1392
0.8006	1.8507	9200	1.1344
0.7958	1.8910	9400	1.1535
0.7916	1.9312	9600	1.1527
0.7799	1.9714	9800	1.1460
0.7794	2.0117	10000	1.1498
0.7754	2.0519	10200	1.1534
0.7723	2.0921	10400	1.1568
0.7579	2.1324	10600	1.1635
0.7571	2.1726	10800	1.1567
0.7566	2.2128	11000	1.1763
0.7469	2.2531	11200	1.1657
0.74	2.2933	11400	1.1744
0.721	2.3335	11600	1.1735
0.7238	2.3738	11800	1.1729
0.7281	2.4140	12000	1.1903
0.7197	2.4542	12200	1.1895
0.7075	2.4945	12400	1.1840
0.7185	2.5347	12600	1.1843
0.7131	2.5749	12800	1.1811
0.7066	2.6152	13000	1.1886
0.7013	2.6554	13200	1.2076
0.6907	2.6956	13400	1.1883
0.6833	2.7359	13600	1.1965
0.6928	2.7761	13800	1.2085
0.6774	2.8163	14000	1.1977
0.6861	2.8566	14200	1.2152
0.6795	2.8968	14400	1.2142
0.6679	2.9370	14600	1.2088
0.6551	2.9773	14800	1.2272
0.6599	3.0175	15000	1.2190
0.6443	3.0577	15200	1.2301
0.6464	3.0980	15400	1.2363
0.6416	3.1382	15600	1.2255
0.6479	3.1784	15800	1.2394
0.6366	3.2187	16000	1.2341
0.6365	3.2589	16200	1.2461
0.6348	3.2991	16400	1.2379
0.6372	3.3394	16600	1.2372
0.6213	3.3796	16800	1.2410
0.6293	3.4198	17000	1.2479
0.6212	3.4601	17200	1.2410
0.6295	3.5003	17400	1.2473
0.6239	3.5405	17600	1.2542
0.6175	3.5808	17800	1.2481
0.6156	3.6210	18000	1.2608
0.614	3.6612	18200	1.2660
0.6072	3.7015	18400	1.2581
0.5974	3.7417	18600	1.2547
0.603	3.7819	18800	1.2671
0.5953	3.8222	19000	1.2588
0.5931	3.8624	19200	1.2762
0.5972	3.9026	19400	1.2587
0.5873	3.9429	19600	1.2870
0.592	3.9831	19800	1.2598
0.5851	4.0233	20000	1.2815
0.5684	4.0636	20200	1.2853
0.5758	4.1038	20400	1.2815
0.5803	4.1440	20600	1.2781
0.5668	4.1843	20800	1.2832
0.5659	4.2245	21000	1.2807
0.5687	4.2647	21200	1.2854
0.5711	4.3050	21400	1.2997
0.5668	4.3452	21600	1.2976
0.5601	4.3854	21800	1.2830
0.5631	4.4257	22000	1.2915
0.5633	4.4659	22200	1.3009
0.5596	4.5061	22400	1.2926
0.5572	4.5464	22600	1.2954
0.5497	4.5866	22800	1.3009
0.5523	4.6268	23000	1.3114
0.544	4.6671	23200	1.3007
0.5465	4.7073	23400	1.2887
0.5452	4.7475	23600	1.3136
0.5435	4.7878	23800	1.3094
0.5368	4.8280	24000	1.3141
0.5359	4.8682	24200	1.3112
0.5352	4.9085	24400	1.3126
0.5411	4.9487	24600	1.3149
0.5357	4.9889	24800	1.3144
0.5245	5.0292	25000	1.3235
0.5211	5.0694	25200	1.3211
0.5226	5.1096	25400	1.3162
0.5263	5.1499	25600	1.3308
0.5242	5.1901	25800	1.3286
0.5253	5.2303	26000	1.3320
0.5215	5.2706	26200	1.3249
0.519	5.3108	26400	1.3330
0.5162	5.3510	26600	1.3224
0.5123	5.3913	26800	1.3270
0.5107	5.4315	27000	1.3291
0.5161	5.4717	27200	1.3360
0.515	5.5120	27400	1.3358
0.5137	5.5522	27600	1.3360
0.5201	5.5924	27800	1.3405
0.5001	5.6327	28000	1.3359
0.5032	5.6729	28200	1.3253
0.4985	5.7131	28400	1.3420
0.4993	5.7534	28600	1.3410
0.4964	5.7936	28800	1.3407
0.5148	5.8338	29000	1.3304
0.4968	5.8741	29200	1.3385
0.4998	5.9143	29400	1.3413
0.4905	5.9545	29600	1.3524
0.4937	5.9948	29800	1.3509
0.4899	6.0350	30000	1.3423
0.4985	6.0752	30200	1.3526
0.4914	6.1155	30400	1.3515
0.4885	6.1557	30600	1.3554
0.4904	6.1959	30800	1.3446
0.4839	6.2362	31000	1.3584
0.4854	6.2764	31200	1.3497
0.4828	6.3166	31400	1.3624
0.4878	6.3569	31600	1.3430
0.4862	6.3971	31800	1.3530
0.4844	6.4373	32000	1.3559
0.4713	6.4776	32200	1.3592
0.4841	6.5178	32400	1.3537
0.4834	6.5580	32600	1.3569
0.4774	6.5983	32800	1.3620
0.4808	6.6385	33000	1.3557
0.481	6.6787	33200	1.3602
0.4725	6.7190	33400	1.3667
0.4752	6.7592	33600	1.3612
0.4698	6.7994	33800	1.3568
0.4717	6.8397	34000	1.3703
0.4723	6.8799	34200	1.3598
0.4721	6.9201	34400	1.3538
0.4777	6.9604	34600	1.3646
0.4815	7.0006	34800	1.3581
0.4674	7.0408	35000	1.3688
0.4658	7.0811	35200	1.3728
0.4634	7.1213	35400	1.3690
0.4713	7.1615	35600	1.3664
0.4709	7.2018	35800	1.3719
0.4606	7.2420	36000	1.3700
0.4583	7.2822	36200	1.3702
0.4599	7.3225	36400	1.3719
0.469	7.3627	36600	1.3646
0.4662	7.4029	36800	1.3622
0.4682	7.4432	37000	1.3662
0.47	7.4834	37200	1.3695
0.4653	7.5236	37400	1.3731
0.4676	7.5639	37600	1.3667
0.4689	7.6041	37800	1.3702
0.4675	7.6443	38000	1.3699
0.4614	7.6846	38200	1.3753
0.4622	7.7248	38400	1.3687
0.4662	7.7650	38600	1.3731
0.4609	7.8053	38800	1.3667
0.4661	7.8455	39000	1.3732
0.4605	7.8857	39200	1.3692
0.4649	7.9260	39400	1.3716
0.463	7.9662	39600	1.3692

Framework versions

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

simon-muenker/TWON-Agent-OSN-Replies-de

TWON-Agent-OSN-Replies-de

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

simon-muenker/
TWON-Agent-OSN-Replies-de