Feed back pls 😭
Tell me what and how is my model doing with your current setup
Bro, I really liked your model, it's really high-quality. There are minor flaws in the roleplay scenarios (sometimes the model develops events very quickly within the stated scenarios). But I think this can be fixed if there are more parameters in the initial model.
Regarding the technical implementation. I tested it on my Macbook Pro (m4 pro processor). The model is stable, I did not notice any lags in the system and the load on the RAM is also negligible.
It is generally quite flexible. I adjusted it with the help of ollama and the result for my project is very impressive.
Can you tell me which dataset you used?
Overall, this is a very high-quality job. I would like to repeat the same quality, but with more parameters. Can you tell me what you used as the source?
really appreciate man , try this model . i used 1k+ high quality story and rp example dataset its the best thing I created . its may be little unstable sometimes
samunder12/llama-3.1-8b-Rp-tadashinu-gguf
give the feedback also
Ok thanks. Today I'll try to check it and give you a feedback.
Now I'm preparing for one my project and that's why need to check really many models. If you also help me a bit with some of your recommendations
umm see my following list maybe you find some good models
I've reviewed your model samunder12/llama-3.1-8b-Rp-tadashinu-gguf. It's really very good. I would even say that better then this one. It is particularly good at dealing with the dynamics of scenarios and is able to build interesting relationships with the character being played.
Can you reveal which dataset you used? And do you plan to build a model using 13B or more with the llama3.1 as base in the future?
Ahh, I'm using my own synthetic dataset that I made for RP. You can find it on my profile 😗. I'll also continue making more fine-tunes from January; I'll be focusing on Qwen3 and GPT OSS 20B.
Also, for now, I've only uploaded 500 examples. I'm too lazy to update the other 500, but I will in the future.
here
samunder12/Roleplay_Ennui
Hey, bro! I tested your dataset, it's also good. When do you plan to download the full version you mentioned?
thz man ill upload the dataset i cant give you the date tho , but i will
hey buddy just created new finetune model
samunder12/Llama-3.2-3B-small_Shiro_roleplay-gguf