readme

README.md

@@ -4,4 +4,19 @@ language:
 - cs
 tags:
 - text-generation-inference
----
+---
+# Barticzech
+
+This repository contains a new model based on the BART (Bidirectional and Auto-Regressive Transformers) architecture, specifically pretrained on the Czech language. The model has been fine-tuned using 214GB of plaintext data sourced from the Common Crawl project, providing a comprehensive language understanding and generation capability.
+
+BART, an encoder-decoder transformer model, is known for its effectiveness in various natural language processing (NLP) tasks such as text summarization, text generation, and machine translation. The Czech BART model, trained on a significant amount of Czech textual data, offers valuable insights and enhanced performance for Czech language-specific tasks.
+
+This repository provides easy access to the pretrained Czech BART model, enabling researchers and developers to leverage its powerful capabilities for a wide range of NLP applications in the Czech language. Users can utilize the model's fine-tuned weights and take advantage of the Huggingface library's functionality to integrate the Czech BART model into their own projects or pipelines seamlessly.
+
+Whether you're working on Czech text summarization, machine translation, or any other NLP task involving the Czech language, the Huggingface Czech BART repository serves as a valuable resource for state-of-the-art language understanding and generation in Czech.
+
+Metacentrum, a national e-infrastructure for scientific and research purposes in the Czech Republic, provided the necessary computational resources to train the Czech BART model effectively. The training process spanned of 130 days, utilizing the power of four A40 GPUs. By leveraging the computational capabilities of Metacentrum, our team could effectively harness the massive amount of plaintext data from the Common Crawl project and train the Czech BART model to achieve its high-performance levels.
+
+We are also working on finetuned models for specific tasks to demonstrate high quality of this model. Please help us to track consequent research and link this repository to your projects. In the future we are going to add experiment results to this repository. 
+
+For additional info please contact [email protected].