fil_weak_checkpoint

This model is a fine-tuned version of mikhail-panzo/malay_full_checkpoint on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.5971

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: 1e-07
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 2000
• training_steps: 4000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.8731	6.1	500	0.7693
0.8254	12.2	1000	0.7411
0.7658	18.29	1500	0.7017
0.7163	24.39	2000	0.6637
0.6894	30.49	2500	0.6350
0.6636	36.59	3000	0.6143
0.6534	42.68	3500	0.6035
0.6485	48.78	4000	0.5971

Framework versions

• Transformers 4.40.0.dev0
• Pytorch 2.2.1+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2