v3d_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: 0.3848
• Accuracy: 0.8329
• Precision: 0.7398
• Recall: 0.4527
• F1: 0.5617

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: 8
• eval_batch_size: 8
• seed: 765837
• distributed_type: multi-GPU
• num_devices: 4
• gradient_accumulation_steps: 2
• total_train_batch_size: 64
• 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	0.6112	0.7506	0.3659	0.0746	0.1240
0.5265	0.0532	20	0.6081	0.7647	0.5455	0.0299	0.0566
0.4127	0.1064	40	0.5412	0.7682	0.7	0.0348	0.0664
0.3347	0.1596	60	0.5020	0.7741	0.7368	0.0697	0.1273
0.3077	0.2128	80	0.4463	0.7965	0.6944	0.2488	0.3663
0.307	0.2660	100	0.4498	0.8	0.6782	0.2935	0.4097
0.2413	0.3191	120	0.4316	0.8141	0.7087	0.3632	0.4803
0.326	0.3723	140	0.4107	0.8235	0.7257	0.4080	0.5223
0.2404	0.4255	160	0.4615	0.8094	0.7671	0.2786	0.4088
0.2962	0.4787	180	0.4205	0.8282	0.7619	0.3980	0.5229
0.2727	0.5319	200	0.4830	0.8	0.7627	0.2239	0.3462
0.2844	0.5851	220	0.4187	0.8259	0.7524	0.3930	0.5163
0.2632	0.6383	240	0.4037	0.8235	0.7339	0.3980	0.5161
0.2499	0.6915	260	0.3885	0.8247	0.7241	0.4179	0.5300
0.2121	0.7447	280	0.3953	0.8224	0.7232	0.4030	0.5176
0.2704	0.7979	300	0.3849	0.8329	0.736	0.4577	0.5644
0.2333	0.8511	320	0.3878	0.8318	0.7417	0.4428	0.5545
0.2896	0.9043	340	0.3886	0.8306	0.7395	0.4378	0.55
0.2398	0.9574	360	0.3848	0.8329	0.7398	0.4527	0.5617

Framework versions

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