metadata

language:
  - en
  - ar
  - bg
  - de
  - el
  - fr
  - hi
  - ru
  - es
  - sw
  - th
  - tr
  - ur
  - vi
  - zh
tags:
  - generated_from_trainer
datasets:
  - xnli
metrics:
  - accuracy
model-index:
  - name: pixel-base-finetuned-xnli-translate-train-all
    results:
      - task:
          name: Text Classification
          type: text-classification
        dataset:
          name: XNLI
          type: xnli
          args: xnli
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.6254886211512718

pixel-base-finetuned-xnli-translate-train-all

This model is a fine-tuned version of Team-PIXEL/pixel-base on the XNLI dataset. It achieves the following results on the evaluation set:

Loss: 0.8312
Accuracy: 0.6255

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: 256
eval_batch_size: 8
seed: 555
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 1000
training_steps: 50000
mixed_precision_training: Apex, opt level O1

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.0422	0.04	1000	1.0647	0.4250
0.9622	0.09	2000	1.0015	0.5051
0.93	0.13	3000	0.9750	0.5285
0.9126	0.17	4000	0.9396	0.5488
0.9033	0.22	5000	0.9353	0.5603
0.8861	0.26	6000	0.9369	0.5606
0.8799	0.3	7000	0.9407	0.5575
0.8627	0.35	8000	0.9079	0.5774
0.8658	0.39	9000	0.9110	0.5711
0.8521	0.43	10000	0.8945	0.5837
0.8562	0.48	11000	0.8818	0.5871
0.8479	0.52	12000	0.8771	0.5938
0.8451	0.56	13000	0.8965	0.5844
0.8433	0.61	14000	0.8814	0.5937
0.8331	0.65	15000	0.8721	0.5983
0.8267	0.7	16000	0.8691	0.5978
0.8254	0.74	17000	0.8646	0.5999
0.8214	0.78	18000	0.8700	0.6004
0.815	0.83	19000	0.8621	0.6016
0.8145	0.87	20000	0.8482	0.6119
0.8067	0.91	21000	0.8601	0.6053
0.8063	0.96	22000	0.8535	0.6093
0.8008	1.0	23000	0.8455	0.6123
0.7863	1.04	24000	0.8524	0.6107
0.7918	1.09	25000	0.8450	0.6142
0.7746	1.13	26000	0.8531	0.6095
0.7855	1.17	27000	0.8442	0.6150
0.7903	1.22	28000	0.8386	0.6162
0.7808	1.26	29000	0.8403	0.6178
0.7847	1.3	30000	0.8421	0.6145
0.7822	1.35	31000	0.8427	0.6157
0.769	1.39	32000	0.8397	0.6187
0.7822	1.43	33000	0.8315	0.6213
0.771	1.48	34000	0.8505	0.6141
0.7713	1.52	35000	0.8482	0.6142
0.7663	1.56	36000	0.8490	0.6169
0.7653	1.61	37000	0.8295	0.6229
0.7669	1.65	38000	0.8313	0.6217
0.77	1.69	39000	0.8309	0.6234
0.763	1.74	40000	0.8310	0.6256
0.7609	1.78	41000	0.8302	0.6228
0.7627	1.83	42000	0.8242	0.6269
0.7617	1.87	43000	0.8232	0.6264
0.7636	1.91	44000	0.8265	0.6261
0.7585	1.96	45000	0.8258	0.6268
0.7572	2.0	46000	0.8223	0.6278
0.7396	2.04	47000	0.8348	0.6242
0.7344	2.09	48000	0.8299	0.6270
0.7385	2.13	49000	0.8314	0.6240
0.7275	2.17	50000	0.8312	0.6255

Framework versions

Transformers 4.17.0
Pytorch 1.11.0
Datasets 2.0.0
Tokenizers 0.12.1