Mistral-7B-v0.3_pct_default_r16

This model is a fine-tuned version of unsloth/mistral-7b-v0.3 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 2.0180

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: 0.0001
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 32
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.02
• num_epochs: 1

Training results

Training Loss	Epoch	Step	Validation Loss
1.9646	0.0206	8	2.0348
2.0531	0.0413	16	2.0345
2.1168	0.0619	24	2.0464
2.0712	0.0825	32	2.0462
2.0779	0.1032	40	2.0510
2.0905	0.1238	48	2.0476
2.0624	0.1445	56	2.0475
2.0793	0.1651	64	2.0428
2.0559	0.1857	72	2.0521
2.0597	0.2064	80	2.0680
2.1235	0.2270	88	2.0849
2.14	0.2476	96	2.0772
2.1586	0.2683	104	2.0880
2.0974	0.2889	112	2.0837
2.1577	0.3096	120	2.0838
2.0998	0.3302	128	2.0899
2.1069	0.3508	136	2.0882
2.1621	0.3715	144	2.0846
2.1441	0.3921	152	2.0949
2.1355	0.4127	160	2.0859
2.084	0.4334	168	2.0871
2.1649	0.4540	176	2.0845
2.0651	0.4746	184	2.0719
2.1708	0.4953	192	2.0722
2.1311	0.5159	200	2.0677
2.1038	0.5366	208	2.0627
2.0804	0.5572	216	2.0757
2.0695	0.5778	224	2.0649
2.0961	0.5985	232	2.0643
2.0808	0.6191	240	2.0567
2.1337	0.6397	248	2.0557
2.0565	0.6604	256	2.0555
2.1184	0.6810	264	2.0497
2.0604	0.7017	272	2.0412
2.1099	0.7223	280	2.0384
2.1048	0.7429	288	2.0415
2.0692	0.7636	296	2.0340
2.0489	0.7842	304	2.0331
2.057	0.8048	312	2.0275
2.0485	0.8255	320	2.0224
2.0364	0.8461	328	2.0202
2.014	0.8667	336	2.0240
2.0656	0.8874	344	2.0236
2.0473	0.9080	352	2.0197
2.0279	0.9287	360	2.0180
2.0415	0.9493	368	2.0178
2.0419	0.9699	376	2.0178
2.0597	0.9906	384	2.0180

Framework versions

• PEFT 0.12.0
• Transformers 4.44.0
• Pytorch 2.4.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1