tinylm1M-vanilla-vanilla-2025-04-04-14-18

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: 2.2412

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
2.2869	0.8418	500	2.2561
2.3193	1.6835	1000	2.2529
2.3692	2.5253	1500	2.2459
2.2689	3.3670	2000	2.2464
2.1996	4.2088	2500	2.2444
2.21	5.0505	3000	2.2466
2.1869	5.8923	3500	2.2421
2.2192	6.7340	4000	2.2407
2.1762	7.5758	4500	2.2433
2.2334	8.4175	5000	2.2412

Framework versions

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