Update README.md

README.md

@@ -9,36 +9,68 @@ pipeline_tag: zero-shot-classification
 ---
 
 # Presentation
-We introduce the Bloomz-3b-NLI model, fine-tuned on the [Bloomz-3b-chat-dpo](https://huggingface.co/cmarkea/bloomz-3b-dpo-chat) foundation model.
+We introduce the Bloomz-3b-NLI model, fine-tuned from the [Bloomz-3b-chat-dpo](https://huggingface.co/cmarkea/bloomz-3b-dpo-chat) foundation model.
 This model is trained on a Natural Language Inference (NLI) task in a language-agnostic manner. The NLI task involves determining the semantic relationship
 between a hypothesis and a set of premises, often expressed as pairs of sentences.
 
-The goal is to predict textual entailment (does sentence A imply/contradict/neither sentence B?) and is a classification task (given two sentences, predict one of
+The goal is to predict textual entailment (does sentence A imply/contradict/neither sentence B?) and is a classification task (given two sentences, predict one of the
 three labels).
-Sentence A is called *premise*, and sentence B is called *hypothesis*, then the goal of modelization is determined as follows:
+If sentence A is called *premise*, and sentence B is called *hypothesis*, then the goal of the modelization is to estimate the following:
 $$P(premise=c\in\{contradiction, entailment, neutral\}\vert hypothesis)$$
 
 ### Language-agnostic approach
 It should be noted that hypotheses and premises are randomly chosen between English and French, with each language combination representing a probability of 25%.
 
-### Detaset
-
 ### Performance
 
+| **class**          | **precision (%)** | **f1-score (%)** | **support** |
+| :----------------: | :---------------: | :--------------: | :---------: |
+| **global**         | 81.96             | 81.07            | 5,010       |
+| **contradiction**  | 81.80             | 84.04            | 1,670       | 
+| **entailment**     | 84.82             | 81.96            | 1,670       |
+| **neutral**        | 76.85             | 77.20            | 1,670       |
+
+### Benchmark
+
+Here are the performances for both the hypothesis and premise in French:
+
+| **model**          | **accuracy (%)** | **MCC (x100)** |
+| :--------------: | :--------------: | :------------: |
+| [cmarkea/distilcamembert-base-nli](https://huggingface.co/cmarkea/distilcamembert-base-nli) | 77.45     | 66.24         |
+| [BaptisteDoyen/camembert-base-xnli](https://huggingface.co/BaptisteDoyen/camembert-base-xnli) | 81.72     | 72.67         |
+| [MoritzLaurer/mDeBERTa-v3-base-mnli-xnli](https://huggingface.co/MoritzLaurer/mDeBERTa-v3-base-mnli-xnli) | 83.43 | 75.15     |
+| [cmarkea/bloomz-560m-nli](https://huggingface.co/cmarkea/bloomz-560m-nli) | 68.70 | 53.57     |
+| [cmarkea/bloomz-3b-nli](https://huggingface.co/cmarkea/bloomz-3b-nli) | 81.08 | 71.66     |
+| [cmarkea/bloomz-7b1-mt-nli](https://huggingface.co/cmarkea/bloomz-7b1-mt-nli) | 83.13 | 74.89     |
+
+And now the hypothesis in French and the premise in English (cross-language context):
+
+| **model**          | **accuracy (%)** | **MCC (x100)** |
+| :--------------: | :--------------: | :------------: |
+| [cmarkea/distilcamembert-base-nli](https://huggingface.co/cmarkea/distilcamembert-base-nli) | 16.89     | -26.82      |
+| [BaptisteDoyen/camembert-base-xnli](https://huggingface.co/BaptisteDoyen/camembert-base-xnli) | 74.59     | 61.97         |
+| [MoritzLaurer/mDeBERTa-v3-base-mnli-xnli](https://huggingface.co/MoritzLaurer/mDeBERTa-v3-base-mnli-xnli) | 85.15  | 77.74     |
+| [cmarkea/bloomz-560m-nli](https://huggingface.co/cmarkea/bloomz-560m-nli) | 68.84 | 53.55     |
+| [cmarkea/bloomz-3b-nli](https://huggingface.co/cmarkea/bloomz-3b-nli) | 82.12 | 73.22     |
+| [cmarkea/bloomz-7b1-mt-nli](https://huggingface.co/cmarkea/bloomz-7b1-mt-nli) | 85.43 | 78.25   |
+
 # Zero-shot Classification
-The primary appeal of training such models lies in their zero-shot classification performance. This means the model is capable of classifying any text with any label
-without specific training. What sets the Bloomz-3b-NLI LLMs apart in this realm is their ability to model and extract information from significantly more complex
+The primary interest of training such models lies in their zero-shot classification performance. This means that the model is able to classify any text with any label
+without a specific training. What sets the Bloomz-3b-NLI LLMs apart in this domain is their ability to model and extract information from significantly more complex
 and lengthy test structures compared to models like BERT, RoBERTa, or CamemBERT.
 
 The zero-shot classification task can be summarized by:
 $$P(hypothesis=i\in\mathcal{C}|premise)=\frac{e^{P(premise=entailment\vert hypothesis=i)}}{\sum_{j\in\mathcal{C}}e^{P(premise=entailment\vert hypothesis=j)}}$$
 With *i* representing a hypothesis composed of a template (for example, "This text is about {}.") and *#C* candidate labels ("cinema", "politics", etc.), the set
-of hypotheses comprises {"This text is about cinema.", "This text is about politics.", ...}. It is these hypotheses that we will measure against the premise, which
+of hypotheses is composed of {"This text is about cinema.", "This text is about politics.", ...}. It is these hypotheses that we will measure against the premise, which
 is the sentence we aim to classify.
 
 ### Performance
 
-# How to use Bloomz-560m-NLI
+The model is evaluated based on sentiment analysis evaluation on the French film review site [Allociné](https://huggingface.co/datasets/allocine). The dataset is labeled 
+into 2 classes, positive comments and negative comments. We then use the hypothesis template "Ce commentaire est {}. and the candidate classes "positif" and "negatif".
+
+# How to use Bloomz-3b-NLI
 
 ```python
 from transformers import pipeline