ambrosfitz/opt-history-v2

opt-history-gen2

This model is a fine-tuned version of facebook/opt-350m on the None dataset. It achieves the following results on the evaluation set:

• Loss: 2.8025

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: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 6
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
3.1245	0.1693	100	3.0013
2.9473	0.3386	200	2.9180
2.9014	0.5078	300	2.8787
2.873	0.6771	400	2.8646
2.8631	0.8464	500	2.8631
2.8366	1.0157	600	2.8451
2.64	1.1849	700	2.8380
2.6295	1.3542	800	2.8202
2.6272	1.5235	900	2.7987
2.6327	1.6928	1000	2.7971
2.626	1.8620	1100	2.7739
2.5237	2.0313	1200	2.7829
2.3242	2.2006	1300	2.7812
2.319	2.3699	1400	2.7727
2.3314	2.5391	1500	2.7668
2.3579	2.7084	1600	2.7561
2.307	2.8777	1700	2.7586
2.2612	3.0470	1800	2.7795
2.056	3.2163	1900	2.7801
2.0802	3.3855	2000	2.7670
2.1104	3.5548	2100	2.7708
2.1115	3.7241	2200	2.7629
2.0828	3.8934	2300	2.7606
1.996	4.0626	2400	2.7849
1.8701	4.2319	2500	2.7938
1.92	4.4012	2600	2.7928
1.8844	4.5705	2700	2.7846
1.9058	4.7397	2800	2.7840
1.901	4.9090	2900	2.7821
1.8418	5.0783	3000	2.8017
1.757	5.2476	3100	2.8055
1.7503	5.4168	3200	2.8095
1.7606	5.5861	3300	2.8065
1.7316	5.7554	3400	2.8043
1.7632	5.9247	3500	2.8025

Framework versions

• Transformers 4.42.4
• Pytorch 2.3.1+cu121
• Datasets 2.21.0
• Tokenizers 0.19.1