6a832f856c445c20d219f242e2d9848c

This model is a fine-tuned version of facebook/opt-1.3b on the dair-ai/emotion [split] dataset. It achieves the following results on the evaluation set:

• Loss: 0.2407
• Data Size: 1.0
• Epoch Runtime: 87.3843
• Accuracy: 0.9259
• F1 Macro: 0.8753

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: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• total_train_batch_size: 32
• total_eval_batch_size: 32
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: constant
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Data Size	Epoch Runtime	Accuracy	F1 Macro
No log	0	0	2.1107	0	4.0041	0.1462	0.1163
No log	1	500	1.8818	0.0078	4.7074	0.2692	0.1646
No log	2	1000	1.5066	0.0156	6.8027	0.4531	0.2233
No log	3	1500	0.6801	0.0312	10.0577	0.7954	0.7368
No log	4	2000	0.5085	0.0625	13.6549	0.8458	0.7957
0.0423	5	2500	0.3931	0.125	19.9926	0.8800	0.8397
0.3405	6	3000	0.3624	0.25	29.8949	0.8765	0.7703
0.047	7	3500	0.3248	0.5	52.7267	0.9078	0.8582
0.2396	8.0	4000	0.1941	1.0	90.2748	0.9239	0.8875
0.2017	9.0	4500	0.2168	1.0	87.1754	0.9158	0.8812
0.1396	10.0	5000	0.2289	1.0	87.2644	0.9209	0.8830
0.1139	11.0	5500	0.2489	1.0	85.4137	0.9199	0.8784
0.1146	12.0	6000	0.2407	1.0	87.3843	0.9259	0.8753

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.3.0
• Tokenizers 0.22.1