judge_answer___33_deberta_large_enwiki-answerability-2411

This model is a fine-tuned version of microsoft/deberta-v3-large on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.2559
• Accuracy: 0.9392
• Precision: 0.9429
• Recall: 0.9326
• F1: 0.9377
• F0.5: 0.9409

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: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 2
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1	F0.5
0.2133	0.0797	2000	0.2326	0.9218	0.9375	0.9007	0.9187	0.9299
0.2011	0.1593	4000	0.2527	0.9231	0.9084	0.9378	0.9229	0.9141
0.2094	0.2390	6000	0.2083	0.9256	0.9130	0.9378	0.9253	0.9179
0.1941	0.3186	8000	0.2156	0.9282	0.9460	0.9054	0.9253	0.9376
0.1933	0.3983	10000	0.2356	0.9290	0.9495	0.9033	0.9258	0.9399
0.1874	0.4779	12000	0.2501	0.9267	0.9325	0.9169	0.9247	0.9294
0.1849	0.5576	14000	0.2294	0.9272	0.9120	0.9425	0.9270	0.9180
0.1886	0.6372	16000	0.2367	0.9277	0.9554	0.8945	0.9239	0.9425
0.1865	0.7169	18000	0.1955	0.9356	0.9360	0.9326	0.9343	0.9353
0.1677	0.7966	20000	0.2023	0.9362	0.9398	0.9295	0.9346	0.9377
0.1662	0.8762	22000	0.2184	0.9341	0.9295	0.9368	0.9331	0.9309
0.163	0.9559	24000	0.2025	0.9408	0.9422	0.9368	0.9395	0.9411
0.1384	1.0355	26000	0.2516	0.9395	0.9463	0.9295	0.9378	0.9429
0.139	1.1152	28000	0.2647	0.9390	0.9397	0.9357	0.9377	0.9389
0.136	1.1948	30000	0.2608	0.9392	0.9458	0.9295	0.9375	0.9425
0.1431	1.2745	32000	0.2793	0.9351	0.9496	0.9164	0.9327	0.9428
0.1393	1.3542	34000	0.2370	0.9397	0.9454	0.9310	0.9381	0.9425
0.1325	1.4338	36000	0.2606	0.9369	0.9413	0.9295	0.9353	0.9389
0.1465	1.5135	38000	0.2371	0.9369	0.9450	0.9253	0.9351	0.9410
0.1254	1.5931	40000	0.2831	0.9367	0.9398	0.9305	0.9352	0.9380
0.1383	1.6728	42000	0.2655	0.9397	0.9458	0.9305	0.9381	0.9427
0.1386	1.7524	44000	0.2582	0.9385	0.9476	0.9258	0.9366	0.9432
0.1405	1.8321	46000	0.2535	0.9382	0.9400	0.9336	0.9368	0.9387
0.1428	1.9117	48000	0.2554	0.9392	0.9467	0.9284	0.9375	0.9430
0.1321	1.9914	50000	0.2559	0.9392	0.9429	0.9326	0.9377	0.9409

Framework versions

• Transformers 4.46.2
• Pytorch 2.4.1+cu124
• Datasets 3.1.0
• Tokenizers 0.20.3