tinylm33M-stella-1sent_5clust-2025-04-04-23-41

This model is a fine-tuned version of roneneldan/TinyStories-1M on the None dataset. It achieves the following results on the evaluation set:

• Loss: 2.1470

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-06
• train_batch_size: 4
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 8
• optimizer: Use OptimizerNames.PAGED_ADAMW_8BIT with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss
2.7458	0.8418	500	2.7149
2.6715	1.6835	1000	2.6364
2.6808	2.5253	1500	2.5744
2.548	3.3670	2000	2.5256
2.4352	4.2088	2500	2.4820
2.3871	5.0505	3000	2.4435
2.3559	5.8923	3500	2.4035
2.3425	6.7340	4000	2.3704
2.2846	7.5758	4500	2.3412
2.3005	8.4175	5000	2.3104
2.2133	9.2593	5500	2.2850
2.2329	10.1010	6000	2.2611
2.2431	10.9428	6500	2.2379
2.2264	11.7845	7000	2.2192
2.2356	12.6263	7500	2.2030
2.0921	13.4680	8000	2.1905
2.1761	14.3098	8500	2.1791
2.2719	15.1515	9000	2.1696
2.1107	15.9933	9500	2.1625
2.229	16.8350	10000	2.1559
2.0957	17.6768	10500	2.1515
2.1215	18.5185	11000	2.1493
2.0368	19.3603	11500	2.1470

Framework versions

• Transformers 4.50.3
• Pytorch 2.6.0+cu124
• Datasets 3.5.0
• Tokenizers 0.21.1