update model card README.md

README.md

@@ -0,0 +1,58 @@
+---
+tags:
+- generated_from_trainer
+model-index:
+- name: fine-tune-Pegasus-large
+  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. -->
+
+# fine-tune-Pegasus-large
+
+This model is a fine-tuned version of [google/pegasus-large](https://huggingface.co/google/pegasus-large) on an unknown dataset.
+It achieves the following results on the evaluation set:
+- Loss: 7.6408
+
+## 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: 6.35e-05
+- train_batch_size: 4
+- eval_batch_size: 4
+- seed: 42
+- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
+- lr_scheduler_type: cosine
+- lr_scheduler_warmup_steps: 500
+- num_epochs: 1
+- mixed_precision_training: Native AMP
+
+### Training results
+
+| Training Loss | Epoch | Step | Validation Loss |
+|:-------------:|:-----:|:----:|:---------------:|
+| 7.4563        | 0.4   | 100  | 7.5562          |
+| 6.5608        | 0.8   | 200  | 6.7437          |
+
+
+### Framework versions
+
+- Transformers 4.15.0
+- Pytorch 1.10.1
+- Datasets 1.17.0
+- Tokenizers 0.10.3