Tag: opensource

Simple python script to fine-tune GPT 2.5

Fine-tune GPT-2.5 Create new file and save the contents below to train.py Change file path in the code below to where you want to save the gpt2.5-fine-tuned model and change the path to Hugging Face dataset InnerI/synCAI_144kda Start training by running: Data Card for synCAI-144k-gpt-2.5 (Hugging Face) # synCAI-144k-gpt-2.5 ## OverviewsynCAI-144k-gpt-2.5 is a large language…

Read More

synCAI-144k-llama3 (model) fine-tuned on synCAI_144kda (dataset)

synCAI144kda is a Synthetic Consciousness Artificial Intelligence dataset containing 144,001 data rows designed to advance AI and consciousness studies. It includes 10,000 original rows with diverse questions and responses, as well as 144,000 synthetic rows from Mostly AI, totaling 3,024,000 individual datapoints. This comprehensive dataset is ideal for training AI models, exploring consciousness topics, and…

Read More