OpenCodeInterpreter-DS-6.7B_En__components_size_252_epochs_10_2024-06-21_09-46-21_3556554

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: 3.5900
• Accuracy: 0.514
• Chrf: 0.06
• Bleu: 0.02
• Sacrebleu: 0.0
• Rouge1: 0.058
• Rouge2: 0.0
• Rougel: 0.058
• Rougelsum: 0.058
• Meteor: 0.155

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
0.0295	4.0	252	1.4806	0.466	0.566	0.372	0.4	0.497	0.251	0.418	0.481	0.474
0.0749	8.0	504	2.8206	0.501	0.187	0.134	0.1	0.305	0.115	0.274	0.297	0.244
0.1239	12.0	756	3.8491	0.489	0.043	0.0	0.0	0.01	0.0	0.007	0.01	0.093
1.2844	16.0	1008	3.9696	0.463	0.029	0.0	0.0	0.005	0.0	0.005	0.005	0.132
0.2183	20.0	1260	3.9671	0.462	0.031	0.0	0.0	0.003	0.0	0.003	0.003	0.132
0.5107	24.0	1512	3.8342	0.482	0.047	0.0	0.0	0.003	0.0	0.003	0.003	0.119
0.0534	28.0	1764	3.7173	0.507	0.059	0.0	0.0	0.056	0.0	0.053	0.056	0.146
0.2079	32.0	2016	3.6924	0.522	0.047	0.0	0.0	0.053	0.0	0.053	0.053	0.154
0.0699	36.0	2268	3.6272	0.523	0.046	0.019	0.0	0.048	0.0	0.048	0.048	0.162
0.0702	40.0	2520	3.5900	0.514	0.06	0.02	0.0	0.058	0.0	0.058	0.058	0.155

Framework versions

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