Create README.md

README.md

@@ -0,0 +1,58 @@
+---
+license: mit
+tags:
+- generated_from_trainer
+model-index:
+- name: result
+  results: []
+---
+
+# result
+
+This model is a fine-tuned version of [neuralmind/bert-large-portuguese-cased](https://huggingface.co/neuralmind/bert-large-portuguese-cased) on an unknown dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.8031
+
+## 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: 5e-05
+- train_batch_size: 2
+- eval_batch_size: 8
+- seed: 42
+- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
+- lr_scheduler_type: linear
+- num_epochs: 3.0
+
+### Training results
+
+### Eval metrics
+
+  epoch                   =        3.0
+  eval_loss               =     0.8031
+  eval_runtime            = 0:02:17.77
+  eval_samples            =        134
+  eval_samples_per_second =      0.973
+  eval_steps_per_second   =      0.123
+  perplexity              =     2.2325
+
+### Framework versions
+
+- Transformers 4.13.0.dev0
+- Pytorch 1.10.0+cu102
+- Datasets 1.16.1
+- Tokenizers 0.10.3