camembert-pcg-annotation

This model is a fine-tuned version of camembert-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 6.4139
• Accuracy: 0.1360
• Top3 Accuracy: 0.3559
• Top5 Accuracy: 0.4498
• Precision: 0.0977
• Recall: 0.1360
• F1 Weighted: 0.0718
• F1 Macro: 0.0015

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: 2e-05
• train_batch_size: 32
• eval_batch_size: 64
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 64
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 0.1
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Top3 Accuracy	Top5 Accuracy	Precision	Recall	F1 Weighted	F1 Macro
14.3746	0.0460	500	7.1874	0.0001	0.0004	0.0011	0.0000	0.0001	0.0000	0.0000
14.3475	0.0919	1000	7.1713	0.0002	0.0025	0.0065	0.0005	0.0002	0.0001	0.0000
14.2752	0.1379	1500	7.1294	0.0203	0.0284	0.0316	0.0005	0.0203	0.0008	0.0000
14.1118	0.1838	2000	7.0500	0.0203	0.0422	0.0553	0.0004	0.0203	0.0008	0.0000
13.9231	0.2298	2500	6.9662	0.0203	0.0424	0.0553	0.0004	0.0203	0.0008	0.0000
13.7866	0.2757	3000	6.9052	0.0207	0.0425	0.1148	0.0007	0.0207	0.0012	0.0001
13.5777	0.3217	3500	6.8596	0.0203	0.0948	0.2596	0.0037	0.0203	0.0008	0.0000
13.5243	0.3677	4000	6.8313	0.0318	0.1927	0.3050	0.0110	0.0318	0.0148	0.0002
13.4701	0.4136	4500	6.8140	0.0414	0.1453	0.3289	0.0017	0.0414	0.0033	0.0001
13.3691	0.4596	5000	6.8055	0.0469	0.0912	0.2763	0.0033	0.0469	0.0058	0.0002
13.4375	0.5055	5500	6.7851	0.0548	0.2899	0.4550	0.0301	0.0548	0.0224	0.0003
13.2271	0.5515	6000	6.7735	0.0422	0.2361	0.3570	0.0081	0.0422	0.0132	0.0003
13.1913	0.5975	6500	6.7579	0.1073	0.2300	0.3041	0.0181	0.1073	0.0307	0.0007
13.2021	0.6434	7000	6.7508	0.1209	0.2456	0.3223	0.0426	0.1209	0.0606	0.0008
13.0196	0.6894	7500	6.6944	0.1899	0.4371	0.5526	0.0781	0.1899	0.1057	0.0014
13.0377	0.7353	8000	6.6679	0.2436	0.4310	0.5533	0.1251	0.2436	0.1564	0.0024
12.7614	0.7813	8500	6.6127	0.1679	0.3864	0.5012	0.0542	0.1679	0.0799	0.0013
12.6990	0.8272	9000	6.5721	0.2370	0.4391	0.5454	0.1489	0.2370	0.1534	0.0023
12.5293	0.8732	9500	6.5496	0.1423	0.3454	0.4803	0.1231	0.1423	0.0857	0.0016
12.4603	0.9192	10000	6.5019	0.1150	0.2962	0.4191	0.1250	0.1150	0.0548	0.0013
12.5206	0.9651	10500	6.4139	0.1360	0.3559	0.4498	0.0977	0.1360	0.0718	0.0015

Framework versions

• Transformers 5.2.0
• Pytorch 2.10.0+cu128
• Datasets 4.5.0
• Tokenizers 0.22.2