Update README.md

README.md

@@ -21,14 +21,19 @@ tags:
 Mistral 7b-based model fine-tuned in Spanish to add high quality Spanish text generation.
 
 * Base model Mistral-7b
+  
+* Two GGUF versions, int4 and int8 for fast inference in consumer hardware
+  
+* Quantized using llama.cpp in int4 Q4_0 and int8 Q8_0
 
-* Based on the excelent job of fine-tuning base mistral from udkai/Turdus
+* Based on the excelent udkai/Turdus fine-tuning mistral
 
-* Fine-tuned in Spanish with a collection of poetry, books, wikipedia articles, phylosophy texts and dolly and alpaca-es datasets.
+* Fine-tuned in Spanish with a collection of texts: poetry, books, phylosophy, wikipedia articles cleaned and prepared by author.
+  
+* Added some instruction dolly and alpaca-es datasets.
 
 * Trained using Lora and PEFT and INT8 quantization on 2 GPUs for several days.
 
-* Quantized using llama.cpp in int4 Q4_0 and int8 Q8_0
 
 ## Usage: 
 * Any framework that uses GGUF format.