OpenCodeInterpreter-DS-6.7B_Fi__components_size_252_epochs_10_2024-06-21_20-34-40_3557380

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

• Loss: 3.5549
• Accuracy: 0.467
• Chrf: 0.034
• Bleu: 0.014
• Sacrebleu: 0.0
• Rouge1: 0.003
• Rouge2: 0.0
• Rougel: 0.003
• Rougelsum: 0.003
• Meteor: 0.15

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
0.034	4.0	252	1.3887	0.448	0.574	0.393	0.4	0.49	0.268	0.435	0.474	0.516
0.0767	8.0	504	2.4467	0.478	0.29	0.19	0.2	0.368	0.154	0.332	0.362	0.333
0.1187	12.0	756	3.6036	0.492	0.053	0.062	0.1	0.139	0.069	0.137	0.139	0.148
1.342	16.0	1008	3.9231	0.467	0.041	0.0	0.0	0.076	0.05	0.076	0.076	0.138
0.2705	20.0	1260	4.0131	0.465	0.032	0.011	0.0	0.0	0.0	0.0	0.0	0.113
0.6023	24.0	1512	3.8735	0.499	0.024	0.0	0.0	0.001	0.0	0.001	0.001	0.097
0.0648	28.0	1764	3.7202	0.466	0.02	0.0	0.0	0.0	0.0	0.0	0.0	0.1
0.2405	32.0	2016	3.6423	0.484	0.027	0.0	0.0	0.0	0.0	0.0	0.0	0.125
0.0757	36.0	2268	3.5933	0.497	0.03	0.0	0.0	0.0	0.0	0.0	0.0	0.146
0.0805	40.0	2520	3.5549	0.467	0.034	0.014	0.0	0.003	0.0	0.003	0.003	0.15

Framework versions

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