update model card README.md

README.md

@@ -0,0 +1,57 @@
+---
+tags:
+- summarization
+- en
+- ecnoder-decoder
+- xlmroberta
+- Abstractive Summarization
+- generated_from_trainer
+datasets:
+- cnn_dailymail
+model-index:
+- name: xlmroberta-finetune-en-cnn
+  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. -->
+
+# xlmroberta-finetune-en-cnn
+
+This model is a fine-tuned version of [](https://huggingface.co/) on the cnn_dailymail dataset.
+
+## 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: 0.0005
+- train_batch_size: 4
+- eval_batch_size: 4
+- seed: 42
+- gradient_accumulation_steps: 8
+- total_train_batch_size: 32
+- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
+- lr_scheduler_type: linear
+- lr_scheduler_warmup_steps: 250
+- num_epochs: 5
+- label_smoothing_factor: 0.1
+
+### Framework versions
+
+- Transformers 4.19.4
+- Pytorch 1.11.0+cu113
+- Datasets 2.3.0
+- Tokenizers 0.12.1