llama-7b-finnish-instruct-v0.2_En__components_size_252_epochs_10_2024-06-21_16-27-39_3556556

This model is a fine-tuned version of Finnish-NLP/llama-7b-finnish-instruct-v0.2 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.5638
• Accuracy: 0.767
• Chrf: 0.008
• Bleu: 0.0
• Sacrebleu: 0.0
• Rouge1: 0.0
• Rouge2: 0.0
• Rougel: 0.0
• Rougelsum: 0.0
• Meteor: 0.013

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: 1
• eval_batch_size: 1
• seed: 3407
• distributed_type: multi-GPU
• num_devices: 4
• total_train_batch_size: 4
• total_eval_batch_size: 4
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-06
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 252
• training_steps: 2520

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Chrf	Bleu	Sacrebleu	Rouge1	Rouge2	Rougel	Rougelsum	Meteor
0.032	4.0	252	2.7586	0.549	0.153	0.049	0.0	0.329	0.231	0.326	0.328	0.388
0.0499	8.0	504	1.1096	0.778	0.248	0.159	0.2	0.351	0.226	0.346	0.342	0.487
0.0564	12.0	756	1.0501	0.77	0.282	0.181	0.2	0.394	0.21	0.386	0.382	0.48
1.3688	16.0	1008	1.2661	0.769	0.334	0.228	0.2	0.432	0.244	0.421	0.422	0.475
0.2421	20.0	1260	2.1179	0.729	0.002	0.0	0.0	0.009	0.0	0.009	0.009	0.001
0.5986	24.0	1512	2.1818	0.545	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
0.0524	28.0	1764	1.6688	0.766	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
0.2247	32.0	2016	1.6031	0.765	0.001	0.0	0.0	0.0	0.0	0.0	0.0	0.0
0.0649	36.0	2268	1.5699	0.765	0.01	0.0	0.0	0.0	0.0	0.0	0.0	0.107
0.065	40.0	2520	1.5638	0.767	0.008	0.0	0.0	0.0	0.0	0.0	0.0	0.013

Framework versions

• Transformers 4.37.0
• Pytorch 2.2.1+cu121
• Datasets 2.20.0
• Tokenizers 0.15.2