gemma-2-2b-it-bs-2-lr-5e-05-ep-3-wp-0.1-gacc-16-gnm-1.0-FP16-mx-2048-v2.3

This model is a fine-tuned version of google/gemma-2-2b-it on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 0.6431
Bleu: 14.6503
Chrf: 32.9918
Ter: 84.9561
Gen Len: 1.0

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: 2
eval_batch_size: 2
seed: 42
gradient_accumulation_steps: 16
total_train_batch_size: 32
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
lr_scheduler_warmup_ratio: 0.1
num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Chrf	Ter	Gen Len
1.6305	0.0289	20	1.6468	4.8799	18.9039	154.0306	1.0
1.2481	0.0578	40	0.6159	5.9161	20.616	94.795	1.0
1.1889	0.0867	60	0.5880	6.3486	21.6392	95.4402	1.0
1.0469	0.1157	80	0.5815	7.1401	22.4082	92.8465	1.0
1.1488	0.1446	100	0.5766	7.2559	24.0665	93.819	1.0
1.1192	0.1735	120	0.5579	8.1656	24.6362	91.974	1.0
1.2241	0.2024	140	0.5569	8.5328	25.2912	92.9546	1.0
1.1121	0.2313	160	0.5570	8.2855	24.9762	91.7013	1.0
1.1824	0.2602	180	0.5574	7.9792	24.5218	90.7155	1.0
1.1207	0.2891	200	0.5591	8.5123	24.9312	90.7648	1.0
1.1408	0.3181	220	0.5651	9.263	25.8002	90.2563	1.0
1.1243	0.3470	240	0.5532	9.1717	24.5861	89.8496	1.0
1.1997	0.3759	260	0.5472	9.7144	26.5564	88.2731	1.0
1.2488	0.4048	280	0.5470	9.8966	26.6442	89.7022	1.0
1.2589	0.4337	300	0.5498	10.0924	26.5891	88.7918	1.0
1.1056	0.4626	320	0.5419	10.852	26.8782	88.4185	1.0
1.1813	0.4916	340	0.5414	11.7649	27.9435	87.1194	1.0
1.1506	0.5205	360	0.5326	11.125	28.5236	86.8577	1.0
1.0654	0.5494	380	0.5303	11.4262	28.5943	88.5115	1.0
1.1014	0.5783	400	0.5307	10.66	28.1805	88.1961	1.0
1.1259	0.6072	420	0.5356	10.6158	27.8987	99.1666	1.0
1.0986	0.6361	440	0.5291	12.0846	28.7401	87.9881	1.0
1.1226	0.6650	460	0.5338	12.0306	28.6747	86.7028	1.0
1.0617	0.6940	480	0.5278	12.736	29.7926	85.4065	1.0
1.0705	0.7229	500	0.5371	12.1319	29.6755	87.3459	1.0
1.0779	0.7518	520	0.5308	12.3201	29.5072	85.42	1.0
1.1431	0.7807	540	0.5236	12.0074	30.2736	86.5512	1.0
1.161	0.8096	560	0.5211	11.7915	29.3134	86.8655	1.0
1.0006	0.8385	580	0.5207	11.8458	29.334	87.0373	1.0
1.0935	0.8674	600	0.5227	11.9221	30.197	84.8232	1.0
1.0323	0.8964	620	0.5187	11.3342	29.6238	87.1094	1.0
1.0505	0.9253	640	0.5144	12.1703	30.24	85.4506	1.0
1.0925	0.9542	660	0.5182	11.8802	29.3723	85.3179	1.0
1.1714	0.9831	680	0.5183	12.985	30.9415	84.9945	1.0
0.6904	1.0130	700	0.5487	13.6607	31.9255	83.6157	1.0
0.745	1.0419	720	0.5377	13.9662	31.646	84.1042	1.0
0.7148	1.0708	740	0.5452	13.2872	31.5421	83.9215	1.0
0.643	1.0998	760	0.5499	13.63	31.8782	84.903	1.0
0.6744	1.1287	780	0.5575	14.2596	32.304	84.681	1.0
0.7373	1.1576	800	0.5456	13.5862	31.8253	84.393	1.0
0.6757	1.1865	820	0.5403	14.0562	32.1402	84.7114	1.0
0.6486	1.2154	840	0.5475	13.5419	32.0602	85.0341	1.0
0.708	1.2443	860	0.5417	13.4943	31.5649	84.5704	1.0
0.7236	1.2732	880	0.5395	12.9227	31.2859	85.3346	1.0
0.7797	1.3022	900	0.5476	13.436	31.7015	84.6371	1.0
0.6754	1.3311	920	0.5375	13.2487	31.5434	84.021	1.0
0.6798	1.3600	940	0.5428	14.0067	32.8167	83.7605	1.0
0.6646	1.3889	960	0.5398	14.19	32.5846	84.297	1.0
0.6552	1.4178	980	0.5380	14.2965	33.2382	83.9637	1.0
0.6775	1.4467	1000	0.5430	14.1695	32.9092	84.0506	1.0
0.684	1.4756	1020	0.5372	14.7039	33.3028	83.5982	1.0
0.6972	1.5046	1040	0.5379	14.3924	32.8448	83.1456	1.0
0.6508	1.5335	1060	0.5352	14.583	33.067	83.287	1.0
0.632	1.5624	1080	0.5468	14.7325	33.3281	83.8706	1.0
0.727	1.5913	1100	0.5331	14.3633	32.9322	83.4413	1.0
0.7348	1.6202	1120	0.5333	14.2374	32.6143	82.9146	1.0
0.6656	1.6491	1140	0.5429	14.201	32.6961	83.9726	1.0
0.6256	1.6781	1160	0.5388	14.5453	33.0475	83.4882	1.0
0.6924	1.7070	1180	0.5365	14.6304	33.1562	83.2688	1.0
0.6858	1.7359	1200	0.5393	15.0899	33.5447	82.996	1.0
0.5881	1.7648	1220	0.5392	14.667	33.3438	83.1356	1.0
0.5651	1.7937	1240	0.5402	15.4378	33.6541	81.7708	1.0
0.7284	1.8226	1260	0.5346	14.8227	33.4247	83.4535	1.0
0.6306	1.8515	1280	0.5319	14.8269	33.6046	83.1934	1.0
0.6572	1.8805	1300	0.5320	15.1053	34.0897	82.3907	1.0
0.6535	1.9094	1320	0.5357	14.8839	33.4102	82.4797	1.0
0.6261	1.9383	1340	0.5380	14.9109	33.6909	83.3605	1.0
0.6464	1.9672	1360	0.5285	14.9013	33.5382	83.2153	1.0
0.6824	1.9961	1380	0.5267	15.0163	33.6645	82.6823	1.0
0.2613	2.0260	1400	0.6124	14.48	33.1279	84.4074	1.0
0.3124	2.0549	1420	0.6413	14.1054	32.5095	85.0131	1.0
0.3049	2.0839	1440	0.6374	14.2445	32.4733	85.507	1.0
0.2714	2.1128	1460	0.6361	14.2699	32.3035	85.755	1.0
0.3301	2.1417	1480	0.6309	13.8199	32.526	85.4411	1.0
0.2641	2.1706	1500	0.6380	14.3405	32.8969	85.4122	1.0
0.262	2.1995	1520	0.6387	14.2671	33.0908	85.1801	1.0
0.2673	2.2284	1540	0.6384	14.396	32.982	85.1313	1.0
0.3174	2.2573	1560	0.6401	14.5154	33.0967	85.131	1.0
0.26	2.2863	1580	0.6357	14.393	33.2324	85.1789	1.0
0.2839	2.3152	1600	0.6403	14.4803	32.8922	85.4798	1.0
0.2653	2.3441	1620	0.6384	14.7389	33.293	84.9899	1.0
0.35	2.3730	1640	0.6381	14.598	33.0187	84.9201	1.0
0.3045	2.4019	1660	0.6416	14.458	32.6372	85.1604	1.0
0.2818	2.4308	1680	0.6440	14.5244	32.856	85.571	1.0
0.2652	2.4597	1700	0.6450	14.6749	32.9037	85.3041	1.0
0.2693	2.4887	1720	0.6454	14.5879	32.8362	85.3871	1.0
0.3171	2.5176	1740	0.6439	14.7316	32.953	85.1034	1.0
0.2609	2.5465	1760	0.6454	14.4935	32.9601	84.997	1.0
0.2822	2.5754	1780	0.6446	14.3962	32.9424	85.2155	1.0
0.3131	2.6043	1800	0.6427	14.4512	32.8568	85.0961	1.0
0.2947	2.6332	1820	0.6434	14.372	32.7709	85.1213	1.0
0.2673	2.6621	1840	0.6432	14.4881	32.8784	85.1516	1.0
0.2909	2.6911	1860	0.6423	14.6748	33.1419	85.0915	1.0
0.2783	2.7200	1880	0.6416	14.8233	33.1044	85.0892	1.0
0.2306	2.7489	1900	0.6416	14.9317	33.182	84.7472	1.0
0.2913	2.7778	1920	0.6431	14.8385	33.1646	84.9655	1.0
0.322	2.8067	1940	0.6429	14.7481	33.1404	84.955	1.0
0.2826	2.8356	1960	0.6421	14.7345	33.0948	84.8726	1.0
0.2529	2.8646	1980	0.6424	14.721	33.0926	84.9997	1.0
0.352	2.8935	2000	0.6426	14.6022	32.9792	85.1056	1.0
0.2622	2.9224	2020	0.6425	14.8069	33.1242	84.8546	1.0
0.3045	2.9513	2040	0.6432	14.7134	33.0836	84.9625	1.0
0.2893	2.9802	2060	0.6431	14.6503	32.9918	84.9561	1.0

Framework versions

Transformers 4.49.0
Pytorch 2.6.0+cu124
Datasets 2.21.0
Tokenizers 0.21.0

BounharAbdelaziz/gemma-2-2b-it-bs-2-lr-5e-05-ep-3-wp-0.1-gacc-16-gnm-1.0-FP16-mx-2048-v2.3

gemma-2-2b-it-bs-2-lr-5e-05-ep-3-wp-0.1-gacc-16-gnm-1.0-FP16-mx-2048-v2.3

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

为此模型做种