update model card README.md

README.md

@@ -0,0 +1,57 @@
+---
+tags:
+- generated_from_trainer
+datasets:
+- squad
+model-index:
+- name: ft_xtremedistil_lr-2e-1_epochs-1
+  results: []
+---
+
+<!-- This model card has been generated automatically according to the information the Trainer had access to. You
+should probably proofread and complete it, then remove this comment. -->
+
+# ft_xtremedistil_lr-2e-1_epochs-1
+
+This model is a fine-tuned version of [loodos/electra-small-turkish-uncased-discriminator](https://huggingface.co/loodos/electra-small-turkish-uncased-discriminator) on the squad dataset.
+It achieves the following results on the evaluation set:
+- Loss: 5.9506
+
+## 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: 0.2
+- train_batch_size: 16
+- eval_batch_size: 16
+- seed: 42
+- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
+- lr_scheduler_type: linear
+- num_epochs: 1
+
+### Training results
+
+| Training Loss | Epoch | Step | Validation Loss |
+|:-------------:|:-----:|:----:|:---------------:|
+| 5.951         | 1.0   | 5818 | 5.9506          |
+
+
+### Framework versions
+
+- Transformers 4.15.0
+- Pytorch 1.10.0+cu111
+- Datasets 1.17.0
+- Tokenizers 0.10.3