update model card README.md

README.md

@@ -0,0 +1,54 @@
+---
+tags:
+- generated_from_trainer
+datasets:
+- cnn_dailymail
+model-index:
+- name: albert_bert_summarization_cnn_dailymail
+  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. -->
+
+# albert_bert_summarization_cnn_dailymail
+
+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: 5e-05
+- train_batch_size: 8
+- eval_batch_size: 8
+- seed: 42
+- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
+- lr_scheduler_type: linear
+- lr_scheduler_warmup_steps: 2000
+- num_epochs: 3.0
+- mixed_precision_training: Native AMP
+
+### Training results
+
+
+
+### Framework versions
+
+- Transformers 4.12.0.dev0
+- Pytorch 1.10.0+cu111
+- Datasets 1.18.3
+- Tokenizers 0.10.3