debiased_disambiguated

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

• Loss: 1.2453
• Bleu: 43.5861
• Meteor: 0.6865
• Chrf++: 62.3743

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: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 1000
• training_steps: 20000

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Meteor	Chrf++
1.4455	4.0	2500	1.1810	43.8822	0.6877	62.5322
1.2959	8.0	5000	1.1951	43.9075	0.6891	62.7025
1.2075	12.0	7500	1.2133	43.6813	0.6888	62.5626
1.1445	16.0	10000	1.2240	43.5134	0.6885	62.4241
1.1008	20.0	12500	1.2336	43.5913	0.6874	62.444
1.0732	24.0	15000	1.2432	43.5071	0.6865	62.3624
1.058	28.0	17500	1.2448	43.5427	0.6867	62.3699
1.0526	32.0	20000	1.2453	43.5861	0.6865	62.3743

Framework versions

• Transformers 4.38.0
• Pytorch 2.2.1+cu121
• Datasets 2.19.1
• Tokenizers 0.15.2