Artigenz-Coder-DS-6.7B_Fi__components_size_252_epochs_10_2024-06-21_09-34-51_3556543

This model is a fine-tuned version of Artigenz/Artigenz-Coder-DS-6.7B on the None dataset. It achieves the following results on the evaluation set:

• Loss: 3.6318
• Accuracy: 0.48
• Chrf: 0.022
• Bleu: 0.0
• Sacrebleu: 0.0
• Rouge1: 0.0
• Rouge2: 0.0
• Rougel: 0.0
• Rougelsum: 0.0
• Meteor: 0.099

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
0.0255	4.0	252	1.2068	0.468	0.574	0.429	0.4	0.507	0.325	0.458	0.504	0.56
0.0989	8.0	504	3.8823	0.479	0.026	0.0	0.0	0.0	0.0	0.0	0.0	0.143
0.1071	12.0	756	3.8516	0.477	0.036	0.036	0.0	0.158	0.079	0.155	0.158	0.135
1.382	16.0	1008	3.7440	0.485	0.046	0.016	0.0	0.159	0.094	0.159	0.159	0.16
0.2463	20.0	1260	3.8049	0.48	0.04	0.0	0.0	0.066	0.033	0.066	0.066	0.139
0.6094	24.0	1512	3.9803	0.446	0.021	0.0	0.0	0.0	0.0	0.0	0.0	0.069
0.0514	28.0	1764	3.7417	0.48	0.023	0.0	0.0	0.0	0.0	0.0	0.0	0.104
0.2303	32.0	2016	3.6727	0.48	0.015	0.0	0.0	0.0	0.0	0.0	0.0	0.033
0.064	36.0	2268	3.6537	0.48	0.02	0.0	0.0	0.0	0.0	0.0	0.0	0.084
0.0697	40.0	2520	3.6318	0.48	0.022	0.0	0.0	0.0	0.0	0.0	0.0	0.099

Framework versions

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