fine-tuning-Phi2-with-webglm-qa-with-lora_9

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

• Loss: 0.1781

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: 5
• total_train_batch_size: 10
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 60
• training_steps: 700
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
1.9045	0.16	10	1.6827
1.6798	0.32	20	1.5996
1.5222	0.48	30	1.4548
1.286	0.64	40	1.2269
1.0827	0.8	50	1.1089
0.9913	0.96	60	1.0021
0.881	1.13	70	0.8796
0.7673	1.29	80	0.7637
0.6315	1.45	90	0.6618
0.554	1.61	100	0.5964
0.5132	1.77	110	0.5487
0.4915	1.93	120	0.5030
0.4787	2.09	130	0.4705
0.4298	2.25	140	0.4451
0.4009	2.41	150	0.4099
0.3886	2.57	160	0.3889
0.3729	2.73	170	0.3674
0.3236	2.89	180	0.3527
0.3377	3.05	190	0.3407
0.3356	3.22	200	0.3261
0.3083	3.38	210	0.3121
0.2794	3.54	220	0.2992
0.2917	3.7	230	0.2926
0.2895	3.86	240	0.2879
0.2764	4.02	250	0.2782
0.2585	4.18	260	0.2732
0.2489	4.34	270	0.2678
0.2401	4.5	280	0.2591
0.2489	4.66	290	0.2573
0.2529	4.82	300	0.2501
0.2637	4.98	310	0.2455
0.255	5.14	320	0.2411
0.2266	5.31	330	0.2370
0.2209	5.47	340	0.2326
0.2311	5.63	350	0.2276
0.2203	5.79	360	0.2275
0.2048	5.95	370	0.2210
0.2133	6.11	380	0.2179
0.2045	6.27	390	0.2142
0.2053	6.43	400	0.2137
0.1898	6.59	410	0.2102
0.1897	6.75	420	0.2073
0.2141	6.91	430	0.2040
0.1872	7.07	440	0.2028
0.1938	7.23	450	0.1998
0.187	7.4	460	0.2004
0.1782	7.56	470	0.1973
0.1908	7.72	480	0.1967
0.1899	7.88	490	0.1912
0.1823	8.04	500	0.1912
0.1769	8.2	510	0.1915
0.1774	8.36	520	0.1909
0.1793	8.52	530	0.1890
0.1853	8.68	540	0.1880
0.1785	8.84	550	0.1861
0.1515	9.0	560	0.1845
0.1689	9.16	570	0.1845
0.1552	9.32	580	0.1836
0.1712	9.49	590	0.1828
0.1642	9.65	600	0.1818
0.1703	9.81	610	0.1806
0.1772	9.97	620	0.1804
0.1615	10.13	630	0.1796
0.1494	10.29	640	0.1801
0.1702	10.45	650	0.1798
0.1656	10.61	660	0.1787
0.1688	10.77	670	0.1782
0.1452	10.93	680	0.1780
0.1732	11.09	690	0.1782
0.1719	11.25	700	0.1781

Framework versions

• PEFT 0.7.1
• Transformers 4.36.2
• Pytorch 2.0.0
• Datasets 2.15.0
• Tokenizers 0.15.0