You can deploy this model and start earning money today!

kanishka/
opt-babylm2-rewritten-clean-spacy-earlystop_adj-num-swapped-bpe_seed-1024_1e-3

Name: opt-babylm2-rewritten-clean-spacy-earlystop_adj-num-swapped-bpe_seed-1024_1e-3
Availability: InStock

generated_from_trainer

kanishka/babylm2-rewritten-clean-spacy_adj-num-swapped

model-index

autotrain_compatible

text-generation-inference

endpoints_compatible

Views

No views yet

opt-babylm2-rewritten-clean-spacy-earlystop_adj-num-swapped-bpe_seed-1024_1e-3

This model was trained from scratch on the kanishka/babylm2-rewritten-clean-spacy_adj-num-swapped dataset. It achieves the following results on the evaluation set:

Loss: 2.6859
Accuracy: 0.4783

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: 0.001
train_batch_size: 32
eval_batch_size: 64
seed: 1024
gradient_accumulation_steps: 8
total_train_batch_size: 256
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: linear
lr_scheduler_warmup_steps: 32000
num_epochs: 20.0
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
4.0984	1.0	2257	3.8230	0.3595
3.4546	2.0	4514	3.3142	0.4079
3.1309	3.0	6771	3.1048	0.4285
2.9254	4.0	9028	2.9916	0.4395
2.8436	5.0	11285	2.9260	0.4466
2.788	6.0	13542	2.8870	0.4506
2.7429	7.0	15799	2.8624	0.4537
2.7117	8.0	18056	2.8465	0.4556
2.689	9.0	20313	2.8287	0.4579
2.6685	10.0	22570	2.8176	0.4587
2.6507	11.0	24827	2.8119	0.4600
2.6497	12.0	27084	2.8034	0.4603
2.6387	13.0	29341	2.8000	0.4610
2.6313	14.0	31598	2.7950	0.4611
2.6138	15.0	33855	2.7694	0.4646
2.5705	16.0	36112	2.7448	0.4679
2.522	17.0	38369	2.7236	0.4710
2.4658	18.0	40626	2.7025	0.4741
2.4016	19.0	42883	2.6879	0.4769
2.3272	19.9915	45120	2.6859	0.4783

Framework versions

Transformers 4.48.0
Pytorch 2.6.0+cu124
Datasets 3.2.0
Tokenizers 0.21.1

kanishka/opt-babylm2-rewritten-clean-spacy-earlystop_adj-num-swapped-bpe_seed-1024_1e-3

opt-babylm2-rewritten-clean-spacy-earlystop_adj-num-swapped-bpe_seed-1024_1e-3

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

kanishka/
opt-babylm2-rewritten-clean-spacy-earlystop_adj-num-swapped-bpe_seed-1024_1e-3