TWON-Agent-OSN-Replies-en

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: 1.8358

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
2.3503	0.0497	200	2.1682
2.1036	0.0994	400	2.1058
1.9649	0.1491	600	2.0560
1.8396	0.1988	800	2.0452
1.7276	0.2484	1000	2.0329
1.6524	0.2981	1200	2.0138
1.5645	0.3478	1400	2.0187
1.4801	0.3975	1600	1.9912
1.4005	0.4472	1800	1.9669
1.3587	0.4969	2000	1.9804
1.2921	0.5466	2200	1.9398
1.2397	0.5963	2400	1.9562
1.1929	0.6460	2600	1.9550
1.1304	0.6957	2800	1.9113
1.0899	0.7453	3000	1.9063
1.0553	0.7950	3200	1.9009
1.0314	0.8447	3400	1.8910
0.9999	0.8944	3600	1.9052
0.9609	0.9441	3800	1.8723
0.9628	0.9938	4000	1.9101
0.9334	1.0435	4200	1.9030
0.9248	1.0932	4400	1.9266
0.8926	1.1429	4600	1.8806
0.8571	1.1925	4800	1.8865
0.8307	1.2422	5000	1.9113
0.7843	1.2919	5200	1.9155
0.7805	1.3416	5400	1.8569
0.812	1.3913	5600	1.8873
0.7783	1.4410	5800	1.8931
0.7489	1.4907	6000	1.8741
0.7386	1.5404	6200	1.8477
0.7217	1.5901	6400	1.8675
0.7177	1.6398	6600	1.8315
0.7051	1.6894	6800	1.8340
0.6848	1.7391	7000	1.8802
0.6959	1.7888	7200	1.8618
0.6938	1.8385	7400	1.8422
0.6847	1.8882	7600	1.8499
0.6601	1.9379	7800	1.8307
0.6423	1.9876	8000	1.8136
0.6275	2.0373	8200	1.8181
0.6361	2.0870	8400	1.8578
0.6323	2.1366	8600	1.8541
0.6192	2.1863	8800	1.8347
0.6273	2.2360	9000	1.8254
0.5939	2.2857	9200	1.8243
0.5994	2.3354	9400	1.8133
0.591	2.3851	9600	1.8312
0.5619	2.4348	9800	1.8509
0.6013	2.4845	10000	1.8538
0.5538	2.5342	10200	1.8633
0.5623	2.5839	10400	1.7793
0.5824	2.6335	10600	1.8132
0.5611	2.6832	10800	1.8083
0.5493	2.7329	11000	1.8147
0.5329	2.7826	11200	1.7890
0.5374	2.8323	11400	1.7984
0.5389	2.8820	11600	1.8228
0.5193	2.9317	11800	1.7950
0.5222	2.9814	12000	1.8798
0.523	3.0311	12200	1.8300
0.4889	3.0807	12400	1.8481
0.5143	3.1304	12600	1.8807
0.4961	3.1801	12800	1.7871
0.4961	3.2298	13000	1.8220
0.503	3.2795	13200	1.8354
0.4936	3.3292	13400	1.8162
0.4753	3.3789	13600	1.8069
0.4971	3.4286	13800	1.8034
0.4901	3.4783	14000	1.8229
0.4921	3.5280	14200	1.8046
0.4873	3.5776	14400	1.8074
0.4697	3.6273	14600	1.7865
0.478	3.6770	14800	1.7935
0.4657	3.7267	15000	1.8454
0.4616	3.7764	15200	1.8294
0.4463	3.8261	15400	1.8229
0.4489	3.8758	15600	1.8061
0.4628	3.9255	15800	1.8125
0.424	3.9752	16000	1.7936
0.4536	4.0248	16200	1.8191
0.4347	4.0745	16400	1.8064
0.4333	4.1242	16600	1.8251
0.4611	4.1739	16800	1.8013
0.4381	4.2236	17000	1.8054
0.4491	4.2733	17200	1.8044
0.4262	4.3230	17400	1.8105
0.4356	4.3727	17600	1.8472
0.4315	4.4224	17800	1.8449
0.4364	4.4720	18000	1.7980
0.4134	4.5217	18200	1.8057
0.4417	4.5714	18400	1.8060
0.4082	4.6211	18600	1.8169
0.4155	4.6708	18800	1.7955
0.4146	4.7205	19000	1.7947
0.4011	4.7702	19200	1.7869
0.4107	4.8199	19400	1.8057
0.4099	4.8696	19600	1.8007
0.4186	4.9193	19800	1.7996
0.3943	4.9689	20000	1.8203
0.4066	5.0186	20200	1.8108
0.3899	5.0683	20400	1.8313
0.404	5.1180	20600	1.8058
0.3946	5.1677	20800	1.8053
0.4003	5.2174	21000	1.8303
0.3865	5.2671	21200	1.8430
0.3917	5.3168	21400	1.8160
0.3952	5.3665	21600	1.8379
0.3983	5.4161	21800	1.8183
0.3704	5.4658	22000	1.8574
0.3923	5.5155	22200	1.8571
0.404	5.5652	22400	1.8146
0.3892	5.6149	22600	1.8386
0.3923	5.6646	22800	1.8043
0.3879	5.7143	23000	1.8278
0.3768	5.7640	23200	1.8183
0.3817	5.8137	23400	1.8415
0.3817	5.8634	23600	1.8258
0.3901	5.9130	23800	1.8612
0.39	5.9627	24000	1.7939
0.3704	6.0124	24200	1.8254
0.3759	6.0621	24400	1.8531
0.3642	6.1118	24600	1.8282
0.3692	6.1615	24800	1.8415
0.3755	6.2112	25000	1.8282
0.3786	6.2609	25200	1.8480
0.3752	6.3106	25400	1.8108
0.3567	6.3602	25600	1.8374
0.3685	6.4099	25800	1.8302
0.3677	6.4596	26000	1.8064
0.3608	6.5093	26200	1.8223
0.3772	6.5590	26400	1.8755
0.3778	6.6087	26600	1.8252
0.3708	6.6584	26800	1.8462
0.3629	6.7081	27000	1.8216
0.3735	6.7578	27200	1.8410
0.3627	6.8075	27400	1.8334
0.353	6.8571	27600	1.8406
0.3711	6.9068	27800	1.8674
0.3543	6.9565	28000	1.8352
0.3709	7.0062	28200	1.8477
0.3565	7.0559	28400	1.8331
0.3549	7.1056	28600	1.8234
0.3629	7.1553	28800	1.8365
0.3541	7.2050	29000	1.8365
0.3468	7.2547	29200	1.8321
0.3527	7.3043	29400	1.8489
0.3608	7.3540	29600	1.8420
0.365	7.4037	29800	1.8372
0.3668	7.4534	30000	1.8356
0.3553	7.5031	30200	1.8418
0.3456	7.5528	30400	1.8356
0.3578	7.6025	30600	1.8243
0.353	7.6522	30800	1.8492
0.342	7.7019	31000	1.8441
0.3493	7.7516	31200	1.8292
0.3553	7.8012	31400	1.8441
0.3529	7.8509	31600	1.8383
0.3402	7.9006	31800	1.8447
0.3562	7.9503	32000	1.8356
0.3735	8.0	32200	1.8358

Framework versions

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