CodeLlama-13b-Instruct-hf_Fi__CMP_TR_size_304_epochs_10_2024-06-22_21-11-23_3558624

This model is a fine-tuned version of codellama/CodeLlama-13b-Instruct-hf on the None dataset. It achieves the following results on the evaluation set:

• Loss: 2.9516
• Accuracy: 0.486
• Chrf: 0.064
• Bleu: 0.0
• Sacrebleu: 0.0
• Rouge1: 0.113
• Rouge2: 0.0
• Rougel: 0.101
• Rougelsum: 0.113
• Meteor: 0.125

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
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-06
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 304
• training_steps: 3040

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Chrf	Bleu	Sacrebleu	Rouge1	Rouge2	Rougel	Rougelsum	Meteor
0.8913	1.0	304	3.5216	0.458	0.071	0.0	0.0	0.062	0.0	0.061	0.062	0.186
0.0623	2.0	608	4.0963	0.456	0.01	0.0	0.0	0.003	0.0	0.003	0.003	0.004
0.0664	3.0	912	3.4598	0.496	0.013	0.0	0.0	0.02	0.001	0.02	0.02	0.027
0.0251	4.0	1216	3.3636	0.459	0.046	0.0	0.0	0.012	0.0	0.012	0.012	0.165
1.117	5.0	1520	3.4675	0.455	0.014	0.0	0.0	0.004	0.0	0.004	0.004	0.084
0.0704	6.0	1824	3.7849	0.459	0.016	0.0	0.0	0.024	0.0	0.024	0.024	0.068
0.0489	7.0	2128	3.3296	0.486	0.041	0.0	0.0	0.066	0.0	0.066	0.066	0.125
0.0588	8.0	2432	3.0692	0.486	0.06	0.0	0.0	0.059	0.0	0.056	0.057	0.145
0.0243	9.0	2736	2.9771	0.464	0.064	0.0	0.0	0.091	0.0	0.089	0.091	0.124
0.0546	10.0	3040	2.9516	0.486	0.064	0.0	0.0	0.113	0.0	0.101	0.113	0.125

Framework versions

• PEFT 0.7.1
• Transformers 4.37.0
• Pytorch 2.2.1+cu121
• Datasets 2.20.0
• Tokenizers 0.15.2