A question about model fine-tuning.

I encountered a problem while creating a model fine-tuning task: when the output of a certain data sample is too long, it prompts an error, indicating that it cannot exceed 5000 characters. Additionally, when I use the fine-tuned model for text generation, the output stops at a maximum of 1024 tokens. I wonder if these limitations are because I am using a free project? Would these restrictions still apply if I had a paid account?