electra-adapter

ucla-nb-project/electra-adapter

electra-adapter is a machine learning model from ucla-nb-project. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as apache-2.0.

should probably proofread and complete it, then remove this comment. --

Downloads · 30 days

Access

Public

Updated May 18, 2024

Repo size

95.1 MB

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.bin190 MB · 99%

At a glance

License: apache-2.0
Access: Public
Created: May 18, 2024
Updated: May 18, 2024
SHA: 5ccf9544

Base models

google/electra-base-generator

Datasets

datasets/all_binary_and_xe_ey_fae_counterfactual

Model card

From the Hugging Face model README

electra-adapter-finetuned-xe_ey_fae

This model is a fine-tuned version of google/electra-base-generator on the datasets/all_binary_and_xe_ey_fae_counterfactual dataset. It achieves the following results on the evaluation set:

Loss: 2.0392
Accuracy: 0.6258

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: 1e-05
train_batch_size: 8
eval_batch_size: 8
seed: 100
gradient_accumulation_steps: 2
total_train_batch_size: 16
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 3.0
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
3.9488	0.06	500	3.1500	0.5509
2.942	0.13	1000	2.5844	0.5680
2.6751	0.19	1500	2.4443	0.5790
2.582	0.26	2000	2.3701	0.5869
2.5267	0.32	2500	2.3097	0.5937
2.4722	0.39	3000	2.2695	0.5986
2.4289	0.45	3500	2.2329	0.6024
2.404	0.52	4000	2.2063	0.6055
2.3826	0.58	4500	2.1840	0.6087
2.3633	0.64	5000	2.1646	0.6109
2.3425	0.71	5500	2.1557	0.6121
2.333	0.77	6000	2.1350	0.6141
2.311	0.84	6500	2.1292	0.6152
2.3014	0.9	7000	2.1182	0.6166
2.2974	0.97	7500	2.1121	0.6170
2.2866	1.03	8000	2.1079	0.6173
2.2675	1.1	8500	2.0940	0.6192
2.2789	1.16	9000	2.0882	0.6201
2.2684	1.22	9500	2.0873	0.6200
2.2608	1.29	10000	2.0796	0.6209
2.2478	1.35	10500	2.0827	0.6204
2.2524	1.42	11000	2.0741	0.6215
2.2502	1.48	11500	2.0685	0.6220
2.243	1.55	12000	2.0665	0.6228
2.2417	1.61	12500	2.0632	0.6229
2.2398	1.68	13000	2.0593	0.6232
2.2233	1.74	13500	2.0600	0.6232
2.2277	1.8	14000	2.0535	0.6236
2.2344	1.87	14500	2.0485	0.6248
2.2274	1.93	15000	2.0507	0.6245
2.2212	2.0	15500	2.0428	0.6256
2.214	2.06	16000	2.0464	0.6244
2.2104	2.13	16500	2.0477	0.6250
2.2185	2.19	17000	2.0397	0.6257
2.2157	2.26	17500	2.0419	0.6257
2.2128	2.32	18000	2.0439	0.6255
2.2154	2.38	18500	2.0372	0.6259
2.2099	2.45	19000	2.0337	0.6263
2.2045	2.51	19500	2.0396	0.6259
2.2138	2.58	20000	2.0390	0.6262
2.2103	2.64	20500	2.0339	0.6263

Framework versions

Transformers 4.36.2
Pytorch 2.2.0+cu121
Datasets 2.17.0
Tokenizers 0.15.2

License: apache-2.0
Created: May 18, 2024
Updated: May 18, 2024