Update README.md

README.md

@@ -14,13 +14,12 @@ datasets:
 # Description model
 
 Chocolatine-Admin-3B version specialized in French administrative language, supervised fine-tuning of [jpacifico/Chocolatine-3B-Instruct-DPO-v1.2](https://huggingface.co/jpacifico/Chocolatine-3B-Instruct-DPO-v1.2) based on [microsoft/Phi-3.5-mini-instruct](https://huggingface.co/microsoft/Phi-3.5-mini-instruct)  
-based on the official [lexicon](https://www.modernisation.gouv.fr/outils-et-formations/lexique-administratif) published by the French Ministère de la Fonction Publique et de la Réforme de l'Etat.  
 Developed in collaboration with Microsoft.
 
 
 # Data & Training
 
-The [dataset](jpacifico/merged-admin-def-dataset-16k) gathers 2362 administrative terms constituting the basis of the simulation of prompt-answer pairs.  
+The [dataset](jpacifico/merged-admin-def-dataset-16k) based on the official [lexicon](https://www.modernisation.gouv.fr/outils-et-formations/lexique-administratif) published by the French DITP, gathers 2362 administrative terms constituting the basis of the simulation of prompt-answer pairs.  
 The GPT-4o model deployed on Azure OpenAI was used to carry out the building of the dataset in several phases:  
 
 - Extraction of the lexicon pages (previously converted into jpg format)  
@@ -28,7 +27,7 @@ The GPT-4o model deployed on Azure OpenAI was used to carry out the building of
 - Generation of questions from the terms and definitions  
 - Generation of answers in three successive rounds taking into account the previous generations to ensure variety.  
 
-For this version the Fine Tuning (SFT) was performed on 11 epochs with an A100 GPU instance on Azure Machine Learning.
+For this 0.3b version, the Fine Tuning (SFT) was performed on 11 epochs with an A100 GPU instance on Azure Machine Learning.
 
 # Usage