mcdc-test-mistral7b_1

This model is a fine-tuned version of mistralai/Mistral-7B-Instruct-v0.2 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.3051
• Perplexity: 1.3567

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: 2e-05
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• optimizer: Use OptimizerNames.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: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Perplexity
1.304	0.1316	10	1.2325	3.4299
1.2036	0.2632	20	1.1727	3.2307
0.9835	0.3947	30	1.1148	3.0489
1.0769	0.5263	40	1.0580	2.8805
0.7655	0.6579	50	1.0016	2.7226
1.0971	0.7895	60	0.9483	2.5813
0.9665	0.9211	70	0.8926	2.4415
0.8997	1.0526	80	0.8345	2.3036
0.796	1.1842	90	0.7684	2.1564
0.5042	1.3158	100	0.6941	2.0019
0.7255	1.4474	110	0.6208	1.8605
0.3803	1.5789	120	0.5753	1.7777
0.667	1.7105	130	0.5440	1.7229
0.4937	1.8421	140	0.5159	1.6751
0.4055	1.9737	150	0.4944	1.6395
0.6353	2.1053	160	0.4794	1.6151
0.3476	2.2368	170	0.4687	1.5980
0.516	2.3684	180	0.4535	1.5738
0.3863	2.5	190	0.4433	1.5578
0.4376	2.6316	200	0.4343	1.5438
0.3465	2.7632	210	0.4267	1.5322
0.4466	2.8947	220	0.4182	1.5192
0.3201	3.0263	230	0.4120	1.5099
0.3491	3.1579	240	0.4063	1.5012
0.4887	3.2895	250	0.4010	1.4934
0.165	3.4211	260	0.3936	1.4823
0.4533	3.5526	270	0.3856	1.4705
0.3659	3.6842	280	0.3791	1.4609
0.3143	3.8158	290	0.3739	1.4533
0.3062	3.9474	300	0.3706	1.4485
0.4429	4.0789	310	0.3651	1.4406
0.2818	4.2105	320	0.3619	1.4360
0.4472	4.3421	330	0.3578	1.4302
0.3428	4.4737	340	0.3558	1.4273
0.4027	4.6053	350	0.3529	1.4232
0.3663	4.7368	360	0.3493	1.4181
0.2822	4.8684	370	0.3482	1.4165
0.4389	5.0	380	0.3465	1.4142
0.301	5.1316	390	0.3440	1.4106
0.4157	5.2632	400	0.3420	1.4078
0.2803	5.3947	410	0.3406	1.4057
0.2755	5.5263	420	0.3374	1.4014
0.5045	5.6579	430	0.3341	1.3966
0.3563	5.7895	440	0.3317	1.3934
0.3095	5.9211	450	0.3302	1.3913
0.2289	6.0526	460	0.3287	1.3891
0.3098	6.1842	470	0.3282	1.3884
0.2931	6.3158	480	0.3281	1.3884
0.2549	6.4474	490	0.3262	1.3858
0.3111	6.5789	500	0.3221	1.3800
0.2282	6.7105	510	0.3180	1.3744
0.3353	6.8421	520	0.3142	1.3692
0.2276	6.9737	530	0.3119	1.3660
0.3579	7.1053	540	0.3098	1.3631
0.3156	7.2368	550	0.3089	1.3620
0.3629	7.3684	560	0.3088	1.3618
0.2834	7.5	570	0.3094	1.3626
0.2912	7.6316	580	0.3086	1.3615
0.3692	7.7632	590	0.3083	1.3611
0.2412	7.8947	600	0.3081	1.3609
0.3029	8.0263	610	0.3077	1.3603
0.1588	8.1579	620	0.3071	1.3595
0.2616	8.2895	630	0.3068	1.3591
0.2893	8.4211	640	0.3073	1.3597
0.2552	8.5526	650	0.3069	1.3592
0.2677	8.6842	660	0.3057	1.3576
0.3908	8.8158	670	0.3057	1.3576
0.3204	8.9474	680	0.3051	1.3567
0.3005	9.0789	690	0.3050	1.3567
0.2001	9.2105	700	0.3054	1.3571
0.3873	9.3421	710	0.3053	1.3570
0.3552	9.4737	720	0.3051	1.3567

Framework versions

• PEFT 0.14.0
• Transformers 4.51.1
• Pytorch 2.5.1+cu124
• Datasets 3.5.0
• Tokenizers 0.21.0