nllb-chechen

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

• Loss: 2.9375
• Model Preparation Time: 0.0036
• Bleu: 2.8
• Chrf: 27.34
• Chrfpp: 24.48

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-05
• train_batch_size: 128
• eval_batch_size: 256
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 256
• optimizer: Use OptimizerNames.ADAFACTOR and the args are: No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 400
• num_epochs: 4

Training results

Training Loss	Epoch	Step	Validation Loss	Model Preparation Time	Bleu	Chrf	Chrfpp
4.3699	0.0688	100	2.9961	0.0036	2.09	24.96	22.34
4.0284	0.6885	1000	2.9683	0.0036	2.23	25.43	22.75
3.7497	1.3766	2000	2.9708	0.0036	2.31	26.44	23.66
3.5211	2.0647	3000	2.9588	0.0036	2.93	27.02	24.2
3.4787	2.7532	4000	2.9509	0.0036	2.71	27.14	24.36
3.4302	3.4413	5000	2.9375	0.0036	2.8	27.34	24.48

Framework versions

• Transformers 5.3.0
• Pytorch 2.10.0+cu128
• Datasets 4.6.1
• Tokenizers 0.22.2