TestMeanFraction2

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

Loss: 1.3967
Matthews Correlation: 0.2537

Model description

More information needed

Intended uses & limitations

"La panique totale" Cette femme trouve une énorme araignée suspendue à sa douche.

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
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 8

Training results

Training Loss	Epoch	Step	Validation Loss	Matthews Correlation
No log	0.13	50	1.1126	0.1589
No log	0.25	100	1.0540	0.1884
No log	0.38	150	1.1533	0.0818
No log	0.51	200	1.0676	0.1586
No log	0.64	250	0.9949	0.2280
No log	0.76	300	1.0343	0.2629
No log	0.89	350	1.0203	0.2478
No log	1.02	400	1.0041	0.2752
No log	1.15	450	1.0808	0.2256
1.023	1.27	500	1.0029	0.2532
1.023	1.4	550	1.0204	0.2508
1.023	1.53	600	1.1377	0.1689
1.023	1.65	650	1.0499	0.2926
1.023	1.78	700	1.0441	0.2474
1.023	1.91	750	1.0279	0.2611
1.023	2.04	800	1.1511	0.2804
1.023	2.16	850	1.2381	0.2512
1.023	2.29	900	1.3340	0.2385
1.023	2.42	950	1.4372	0.2842
0.7325	2.54	1000	1.3967	0.2537
0.7325	2.67	1050	1.4272	0.2624
0.7325	2.8	1100	1.3869	0.1941
0.7325	2.93	1150	1.4983	0.2063
0.7325	3.05	1200	1.4959	0.2409

Framework versions

Transformers 4.18.0
Pytorch 1.10.0a0+0aef44c
Datasets 2.0.0
Tokenizers 0.11.6