mtzig/v4_mistral_lora_bugged

v4_mistral_lora

This model is a fine-tuned version of peiyi9979/math-shepherd-mistral-7b-prm on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.2886
• Accuracy: 0.8631
• Precision: 0.8457
• Recall: 0.6260
• F1: 0.7195

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: 6
• eval_batch_size: 8
• seed: 89234
• distributed_type: multi-GPU
• num_devices: 4
• gradient_accumulation_steps: 3
• total_train_batch_size: 72
• total_eval_batch_size: 32
• 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: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 1

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
No log	0	0	0.5995	0.7340	0.6	0.1535	0.2445
0.7266	0.0251	20	0.5909	0.7384	0.6232	0.1693	0.2663
0.674	0.0502	40	0.5497	0.7517	0.6526	0.2441	0.3553
0.5187	0.0753	60	0.4896	0.7759	0.6409	0.4567	0.5333
0.4811	0.1004	80	0.4410	0.7958	0.7066	0.4646	0.5606
0.2811	0.1255	100	0.4249	0.8102	0.7595	0.4724	0.5825
0.2959	0.1506	120	0.3751	0.8212	0.7674	0.5197	0.6197
0.336	0.1757	140	0.3764	0.8278	0.8451	0.4724	0.6061
0.3239	0.2008	160	0.3608	0.8201	0.8421	0.4409	0.5788
0.2767	0.2259	180	0.3362	0.8543	0.8112	0.6260	0.7067
0.276	0.2510	200	0.3406	0.8389	0.8462	0.5197	0.6439
0.2715	0.2762	220	0.3223	0.8411	0.8274	0.5472	0.6588
0.2737	0.3013	240	0.3202	0.8521	0.8125	0.6142	0.6996
0.3245	0.3264	260	0.3098	0.8466	0.8249	0.5748	0.6775
0.2868	0.3515	280	0.3159	0.8433	0.7887	0.6024	0.6830
0.2601	0.3766	300	0.3105	0.8587	0.7669	0.7126	0.7388
0.2597	0.4017	320	0.3162	0.8510	0.8362	0.5827	0.6868
0.287	0.4268	340	0.2997	0.8532	0.8071	0.6260	0.7051
0.3115	0.4519	360	0.3028	0.8543	0.8315	0.6024	0.6986
0.2654	0.4770	380	0.3008	0.8543	0.8245	0.6102	0.7014
0.2443	0.5021	400	0.2955	0.8565	0.8039	0.6457	0.7162
0.2743	0.5272	420	0.3011	0.8543	0.8389	0.5945	0.6959
0.2248	0.5523	440	0.3031	0.8532	0.8380	0.5906	0.6928
0.2149	0.5774	460	0.2868	0.8609	0.7991	0.6732	0.7308
0.1998	0.6025	480	0.2975	0.8587	0.8316	0.6220	0.7117
0.2459	0.6276	500	0.2978	0.8510	0.8324	0.5866	0.6882
0.1953	0.6527	520	0.2989	0.8576	0.8492	0.5984	0.7021
0.3153	0.6778	540	0.2864	0.8642	0.8359	0.6417	0.7261
0.2172	0.7029	560	0.3190	0.8444	0.8844	0.5118	0.6484
0.2604	0.7280	580	0.2830	0.8687	0.8358	0.6614	0.7385
0.2671	0.7531	600	0.2970	0.8565	0.8523	0.5906	0.6977
0.2049	0.7782	620	0.2862	0.8587	0.8316	0.6220	0.7117
0.2972	0.8033	640	0.2890	0.8609	0.8404	0.6220	0.7149
0.1953	0.8285	660	0.2911	0.8609	0.8441	0.6181	0.7136
0.24	0.8536	680	0.2824	0.8653	0.8367	0.6457	0.7289
0.282	0.8787	700	0.2860	0.8631	0.8385	0.6339	0.7220
0.1931	0.9038	720	0.2885	0.8620	0.8413	0.6260	0.7178
0.2251	0.9289	740	0.2898	0.8631	0.8457	0.6260	0.7195
0.178	0.9540	760	0.2889	0.8631	0.8457	0.6260	0.7195
0.2431	0.9791	780	0.2886	0.8631	0.8457	0.6260	0.7195

Framework versions

• PEFT 0.13.2
• Transformers 4.46.0
• Pytorch 2.5.1+cu124
• Datasets 3.1.0
• Tokenizers 0.20.3