From 7a0cada18ec705f13cdcf5fc709fae9c9906c811 Mon Sep 17 00:00:00 2001
From: Cardiff NLP <espinosa-ankel@cardiff.ac.uk>
Date: Sat, 17 Apr 2021 06:25:37 +0000
Subject: [PATCH] Update README.md

---
 README.md | 48 +++++++++++++++++++++++++++++-------------------
 1 file changed, 29 insertions(+), 19 deletions(-)

diff --git a/README.md b/README.md
index a79e3a0..c6c386c 100644
--- a/README.md
+++ b/README.md
@@ -1,3 +1,15 @@
+---
+language: multilingual
+widget:
+- text: "T'estimo!"
+- text: "I love you!"
+- text: "I hate you"
+- text: "Mahal kita!"
+- text: "사랑해!"
+- text: "난 너가 싫어"
+---
+
+
 # twitter-XLM-roBERTa-base for Sentiment Analysis
 
 This is a XLM-roBERTa-base model trained on ~198M tweets and finetuned for sentiment analysis in 
@@ -5,7 +17,16 @@ This is a XLM-roBERTa-base model trained on ~198M tweets and finetuned for senti
 - Paper: [XLM-T: A Multilingual Language Model Toolkit for Twitter](https://...). 
 - Git Repo: [Tweeteval official repository](https://github.com/cardiffnlp/xlm-t).
 
-## Example of classification
+## Example Pipeline
+```python
+from transformers import pipeline
+model_path = "cardiffnlp/twitter-xlm-roberta-base-sentiment"
+sentiment_task = pipeline("sentiment-analysis", model=model_path, tokenizer=model_path)
+
+sentiment_task("T'estimo!")
+```
+
+## Full classification example
 
 ```python
 from transformers import AutoModelForSequenceClassification
@@ -13,32 +34,20 @@ from transformers import TFAutoModelForSequenceClassification
 from transformers import AutoTokenizer
 import numpy as np
 from scipy.special import softmax
-import csv
-import urllib.request
 
 # Preprocess text (username and link placeholders)
 def preprocess(text):
     new_text = []
- 
- 
     for t in text.split(" "):
         t = '@user' if t.startswith('@') and len(t) > 1 else t
         t = 'http' if t.startswith('http') else t
         new_text.append(t)
     return " ".join(new_text)
 
+MODEL = f"/home/jupyter/misc/tweeteval/TweetEval_models/xlm-twitter/twitter-xlm-roberta-base-sentiment"
 
-MODEL = f"cardiffnlp/twitter-xlm-roberta-base-sentiment"
 tokenizer = AutoTokenizer.from_pretrained(MODEL)
-
-# download label mapping
-labels=[]
-mapping_link = f"https://raw.githubusercontent.com/cardiffnlp/tweeteval/main/datasets/sentiment/mapping.txt"
-with urllib.request.urlopen(mapping_link) as f:
-    html = f.read().decode('utf-8').split("\\
-")
-    csvreader = csv.reader(html, delimiter='\\\\t')
-labels = [row[1] for row in csvreader if len(row) > 1]
+config = AutoConfig.from_pretrained(MODEL)
 
 # PT
 model = AutoModelForSequenceClassification.from_pretrained(MODEL)
@@ -61,10 +70,11 @@ scores = softmax(scores)
 # scores = output[0][0].numpy()
 # scores = softmax(scores)
 
+# Print labels and scores
 ranking = np.argsort(scores)
 ranking = ranking[::-1]
 for i in range(scores.shape[0]):
-    l = labels[ranking[i]]
+    l = config.id2label[ranking[i]]
     s = scores[ranking[i]]
     print(f"{i+1}) {l} {np.round(float(s), 4)}")
 
@@ -73,8 +83,8 @@ for i in range(scores.shape[0]):
 Output: 
 
 ```
-1) positive 0.76726073
-2) neutral 0.201
-3) negative 0.0312
+1) Positive 0.7673
+2) Neutral 0.2015
+3) Negative 0.0313
 ```