update model card README.md

README.md

@@ -0,0 +1,57 @@
+---
+tags:
+- generated_from_trainer
+model-index:
+- name: nezha-chinese-base-finetuned-product
+  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. -->
+
+# nezha-chinese-base-finetuned-product
+
+This model is a fine-tuned version of [peterchou/nezha-chinese-base](https://huggingface.co/peterchou/nezha-chinese-base) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.0004
+
+## 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: 2e-05
+- train_batch_size: 64
+- eval_batch_size: 64
+- seed: 42
+- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
+- lr_scheduler_type: linear
+- num_epochs: 3.0
+
+### Training results
+
+| Training Loss | Epoch | Step  | Validation Loss |
+|:-------------:|:-----:|:-----:|:---------------:|
+| 0.0309        | 1.0   | 6473  | 0.0037          |
+| 0.0033        | 2.0   | 12946 | 0.0006          |
+| 0.0017        | 3.0   | 19419 | 0.0004          |
+
+
+### Framework versions
+
+- Transformers 4.17.0
+- Pytorch 1.6.0
+- Datasets 2.0.0
+- Tokenizers 0.11.6