deepseek-coder-6.7b-instruct_En__components_size_252_epochs_10_2024-06-21_16-27-40_3556557

This model is a fine-tuned version of deepseek-ai/deepseek-coder-6.7b-instruct on the None dataset. It achieves the following results on the evaluation set:

• Loss: 3.6387
• Accuracy: 0.5
• Chrf: 0.033
• Bleu: 0.0
• Sacrebleu: 0.0
• Rouge1: 0.0
• Rouge2: 0.0
• Rougel: 0.0
• Rougelsum: 0.0
• Meteor: 0.106

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
0.023	4.0	252	1.4114	0.478	0.568	0.392	0.4	0.479	0.269	0.432	0.474	0.5
0.0622	8.0	504	2.7415	0.514	0.179	0.134	0.1	0.282	0.15	0.264	0.28	0.231
0.1198	12.0	756	4.0117	0.468	0.033	0.0	0.0	0.0	0.0	0.0	0.0	0.061
1.3468	16.0	1008	4.0140	0.483	0.043	0.0	0.0	0.003	0.0	0.003	0.003	0.151
0.2086	20.0	1260	3.9480	0.505	0.029	0.0	0.0	0.01	0.0	0.01	0.01	0.114
0.515	24.0	1512	3.8490	0.505	0.045	0.0	0.0	0.032	0.0	0.032	0.032	0.118
0.039	28.0	1764	3.7452	0.503	0.04	0.0	0.0	0.0	0.0	0.0	0.0	0.104
0.1975	32.0	2016	3.6962	0.501	0.035	0.0	0.0	0.003	0.0	0.003	0.003	0.105
0.0561	36.0	2268	3.6596	0.499	0.036	0.0	0.0	0.0	0.0	0.0	0.0	0.104
0.0567	40.0	2520	3.6387	0.5	0.033	0.0	0.0	0.0	0.0	0.0	0.0	0.106

Framework versions

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