Add detail to README

README.md

@@ -27,42 +27,37 @@ model-index:
       value: 6.5785713084850626
 ---
 
-<!-- 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. -->
+# Whisper Medium Portuguese 🇧🇷🇵🇹
 
-# Whisper Medium Portuguese
+Bem-vindo ao whisper medium para transcrição em português 👋🏻
 
-This model is a fine-tuned version of [openai/whisper-medium](https://huggingface.co/openai/whisper-medium) on the mozilla-foundation/common_voice_11_0 pt dataset.
-It achieves the following results on the evaluation set:
-- Loss: 0.3205
-- Wer: 6.5786
+If you are looking to **quickly**, and **reliably**, transcribe portuguese audio to text, you are in the right place!
 
-## Model description
+With a state-of-the-art [Word Error Rate](https://huggingface.co/spaces/evaluate-metric/wer) (WER) of just **6.58** in Common Voice 11, this model shows increases in precision of more than **x2** compared to past state of the art [wav2vec2](https://huggingface.co/Edresson/wav2vec2-large-xlsr-coraa-portuguese) models. When compared to the original [whisper-medium](https://huggingface.co/openai/whisper-medium) model it shows a **x1.2** improvement 🚀. 
 
-More information needed
+This model is a fine-tuned version of [openai/whisper-medium](https://huggingface.co/openai/whisper-medium) on the [mozilla-foundation/common_voice_11](https://huggingface.co/datasets/mozilla-foundation/common_voice_11_0) dataset. 
 
-## Intended uses & limitations
+The following table shows a **comparison** between the results of our model and those achieved by the most downloaded models in the hub for portuguese Automatic Speech Recognition:
 
-More information needed
+| Model                                            | WER    | Parameters |
+|--------------------------------------------------|:--------:|:------------:|
+| [openai/whisper-medium](https://huggingface.co/openai/whisper-medium)                            | 8.10   | 769M       |
+| [jlondonobo/whisper-medium-pt](https://huggingface.co/jlondonobo/whisper-medium-pt)                     | **6.58** 🤗  | 769M       |
+| [jonatasgrosman/wav2vec2-large-xlsr-53-portuguese](https://huggingface.co/jonatasgrosman/wav2vec2-large-xlsr-53-portuguese) | 11.31  | 317M       |
+| [Edresson/wav2vec2-large-xlsr-coraa-portuguese](https://huggingface.co/Edresson/wav2vec2-large-xlsr-coraa-portuguese)    | 20.08 | 317M       |
 
-## Training and evaluation data
-
-More information needed
-
-## Training procedure
 
 ### Training hyperparameters
-
-The following hyperparameters were used during training:
-- learning_rate: 1e-05
-- train_batch_size: 32
-- eval_batch_size: 16
-- seed: 42
-- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
-- lr_scheduler_type: linear
-- lr_scheduler_warmup_steps: 500
-- training_steps: 5000
-- mixed_precision_training: Native AMP
+We used the following hyperparameters for training:
+- `learning_rate`: 1e-05
+- `train_batch_size`: 32
+- `eval_batch_size`: 16
+- `seed`: 42
+- `optimizer`: Adam with betas=(0.9,0.999) and epsilon=1e-08
+- `lr_scheduler_type`: linear
+- `lr_scheduler_warmup_steps`: 500
+- `training_steps`: 5000
+- `mixed_precision_training`: Native AMP
 
 ### Training results
 
@@ -72,7 +67,7 @@ The following hyperparameters were used during training:
 | 0.0218        | 3.07  | 2000 | 0.2254          | 7.1098 |
 | 0.0053        | 5.06  | 3000 | 0.2711          | 6.9686 |
 | 0.0017        | 7.04  | 4000 | 0.3030          | 6.6862 |
-| 0.0005        | 9.02  | 5000 | 0.3205          | 6.5786 |
+| 0.0005        | 9.02  | 5000 | 0.3205          | **6.5786** 🤗 |
 
 
 ### Framework versions
@@ -80,4 +75,4 @@ The following hyperparameters were used during training:
 - Transformers 4.26.0.dev0
 - Pytorch 1.13.0+cu117
 - Datasets 2.7.1.dev0
-- Tokenizers 0.13.2
+- Tokenizers 0.13.2