arabert_cross_relevance_task3_fold4

This model is a fine-tuned version of aubmindlab/bert-base-arabertv02 on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 0.2244
  • Qwk: 0.3147
  • Mse: 0.2244

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: 64
  • eval_batch_size: 64
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 10

Training results

Training Loss Epoch Step Validation Loss Qwk Mse
No log 0.1176 2 0.6247 0.1100 0.6247
No log 0.2353 4 0.4274 0.2657 0.4274
No log 0.3529 6 0.3933 0.3269 0.3933
No log 0.4706 8 0.3251 0.1665 0.3251
No log 0.5882 10 0.4511 0.2506 0.4511
No log 0.7059 12 0.3380 0.2405 0.3380
No log 0.8235 14 0.2594 0.1665 0.2594
No log 0.9412 16 0.2948 0.2149 0.2948
No log 1.0588 18 0.2980 0.2995 0.2980
No log 1.1765 20 0.2541 0.2638 0.2541
No log 1.2941 22 0.2285 0.2330 0.2285
No log 1.4118 24 0.2279 0.2837 0.2279
No log 1.5294 26 0.2347 0.2760 0.2347
No log 1.6471 28 0.2414 0.3095 0.2414
No log 1.7647 30 0.2534 0.3281 0.2534
No log 1.8824 32 0.2613 0.3439 0.2613
No log 2.0 34 0.2476 0.3483 0.2476
No log 2.1176 36 0.2444 0.3210 0.2444
No log 2.2353 38 0.2368 0.3126 0.2368
No log 2.3529 40 0.2271 0.3358 0.2271
No log 2.4706 42 0.2280 0.3354 0.2280
No log 2.5882 44 0.2317 0.3468 0.2317
No log 2.7059 46 0.2288 0.3795 0.2288
No log 2.8235 48 0.2284 0.3808 0.2284
No log 2.9412 50 0.2248 0.4119 0.2248
No log 3.0588 52 0.2179 0.3948 0.2179
No log 3.1765 54 0.2135 0.3797 0.2135
No log 3.2941 56 0.2038 0.3300 0.2038
No log 3.4118 58 0.2043 0.3232 0.2043
No log 3.5294 60 0.2143 0.3534 0.2143
No log 3.6471 62 0.2325 0.4158 0.2325
No log 3.7647 64 0.2488 0.4475 0.2488
No log 3.8824 66 0.2377 0.3842 0.2377
No log 4.0 68 0.2342 0.3300 0.2342
No log 4.1176 70 0.2380 0.4475 0.2380
No log 4.2353 72 0.2243 0.4095 0.2243
No log 4.3529 74 0.2117 0.3895 0.2117
No log 4.4706 76 0.2105 0.3541 0.2105
No log 4.5882 78 0.2109 0.3824 0.2109
No log 4.7059 80 0.2150 0.4058 0.2150
No log 4.8235 82 0.2189 0.4186 0.2189
No log 4.9412 84 0.2207 0.3858 0.2207
No log 5.0588 86 0.2206 0.3152 0.2206
No log 5.1765 88 0.2173 0.2918 0.2173
No log 5.2941 90 0.2159 0.3281 0.2159
No log 5.4118 92 0.2217 0.3375 0.2217
No log 5.5294 94 0.2240 0.4196 0.2240
No log 5.6471 96 0.2310 0.4755 0.2310
No log 5.7647 98 0.2348 0.4823 0.2348
No log 5.8824 100 0.2267 0.3877 0.2267
No log 6.0 102 0.2189 0.3233 0.2189
No log 6.1176 104 0.2198 0.3013 0.2198
No log 6.2353 106 0.2188 0.3233 0.2188
No log 6.3529 108 0.2193 0.3440 0.2193
No log 6.4706 110 0.2212 0.3392 0.2212
No log 6.5882 112 0.2240 0.3533 0.2240
No log 6.7059 114 0.2234 0.4012 0.2234
No log 6.8235 116 0.2253 0.3980 0.2253
No log 6.9412 118 0.2273 0.3906 0.2273
No log 7.0588 120 0.2269 0.4035 0.2269
No log 7.1765 122 0.2251 0.3784 0.2251
No log 7.2941 124 0.2253 0.4107 0.2253
No log 7.4118 126 0.2255 0.3891 0.2255
No log 7.5294 128 0.2218 0.3462 0.2218
No log 7.6471 130 0.2198 0.3255 0.2198
No log 7.7647 132 0.2206 0.3328 0.2206
No log 7.8824 134 0.2204 0.3286 0.2204
No log 8.0 136 0.2202 0.3416 0.2202
No log 8.1176 138 0.2237 0.3458 0.2237
No log 8.2353 140 0.2257 0.3529 0.2257
No log 8.3529 142 0.2246 0.3433 0.2246
No log 8.4706 144 0.2235 0.3464 0.2235
No log 8.5882 146 0.2231 0.3396 0.2231
No log 8.7059 148 0.2225 0.3328 0.2225
No log 8.8235 150 0.2224 0.3119 0.2224
No log 8.9412 152 0.2226 0.3159 0.2226
No log 9.0588 154 0.2231 0.3205 0.2231
No log 9.1765 156 0.2239 0.3275 0.2239
No log 9.2941 158 0.2243 0.3275 0.2243
No log 9.4118 160 0.2236 0.3147 0.2236
No log 9.5294 162 0.2234 0.3080 0.2234
No log 9.6471 164 0.2238 0.3080 0.2238
No log 9.7647 166 0.2243 0.3080 0.2243
No log 9.8824 168 0.2244 0.3147 0.2244
No log 10.0 170 0.2244 0.3147 0.2244

Framework versions

  • Transformers 4.44.0
  • Pytorch 2.4.0
  • Datasets 2.21.0
  • Tokenizers 0.19.1
Downloads last month
1
Safetensors
Model size
135M params
Tensor type
F32
·
Inference API
Unable to determine this model's library. Check the docs .

Model tree for salbatarni/arabert_cross_relevance_task3_fold4

Finetuned
(2940)
this model