Project goals were to develop and fine-tune LLM model to generate text which CHAT GPT cannot generate due to censorship restrictions.
Project Details
AI generated content gained immense popularity. Specificially textual data opens up new prospectives for business operartion optimizations and cutting costs.
Client needed to automate paraphrase of specific content for specific resources and in order to achieve that I have fine-tuned open-source Mistral 7B model using Python’s HuggingFace and Langchain frameworks.
Results were pretty impressive and I have cotinued with development on 2 additional contracts where I have created REST API to serve and deploy with Python Fast framework and cloud providers like PaperSpace, Digital Ocean,Kaggle,GCP GPU VMs to rent GPUs.