Hi guys, please send me all of the names of download links of known eaglercraft clients so I can make a repo for it. I do not care if it is AI generated or completely vibecoded. If not, I'll just make a reddit crawler.
​
I am going to fine-tune a satire model, with the base model being `Qwen3.5-14B-Base`. The dataset has ~18000 examples, most of which are pretty long and do not align with the model's internal knowledge (they have incorrect answers to facts), so I needed to do a full fine tune rather than use LoRA.
Today, I rented a cloud computer on vast.ai. It had 4x12GB GPUs and 48GB of system RAM, but I feel stupid because it kept running out of memory when I tried training, and I kept trying to change settings rather than just renting a different computer, so I basically wasted $3. I underestimated how much RAM would be needed for this task.
So, I would like to know if any of you have some recommendations for proper hardware to use.
I'm making a satire AI model that gives fake, onion-like responses. Would it be better to just make a completely new model rather than fine tuning an existing one for this? Most of the info in my dataset completely conflicts with the info that almost all models have.
Hi, I'm not sure if this is the right community for this, but I am currently trying to train an AI on satire responses which are all mostly incorrect, but have an onion-like style (the Onion is a popular satire news parody company).
My plan is to fine-tune a model which has less than 30b parameters. My dataset has around 1.4k examples, most of them manually written, and there should be no conflicting information in the dataset, but the dataset will need to overwrite the base model's knowledge while still retaining grammar knowledge.
Also, since my examples have information that conflicts with already learned knowledge, will the base model learn the new info? (Eg. If one of the examples says that a duck is a tool brand, but also that they were founded in 1957, would the fine tuned model be able to tell me that when I ask it what tool companies were founded in 1957)?
What model would be good for fine-tuning in this case?
Hi, I'm not sure if this is the right community for this, but I am currently trying to train an AI on satire responses which are all mostly incorrect, but have an onion-like style (the Onion is a popular satire news parody company).
My plan is to fine-tune a model which has less than 30b parameters. My dataset has around 1.4k examples, most of them manually written, and there should be no conflicting information in the dataset, but the dataset will need to overwrite the base model's knowledge while still retaining grammar knowledge.
What model would be good for fine-tuning in this case?