CoolFace
Modelpublic

TorpedoSoftware/R1-Distill-Qwen-14B-Roblox-Luau

sourceHugging Facemitupdated 10mo agoView on Hugging Face
1likes31downloads
Model Card

R1-Distill-Qwen-14B-Roblox-Luau

A fine tune of deepseek-ai/DeepSeek-R1-Distill-Qwen-14B using boatbomber/roblox-info-dump and boatbomber/the-luau-stack for Roblox domain knowledge.

Recommended inference settings:

ParameterValueNotes
System PromptYou are an expert Roblox developer and Luau software engineer.Model was fine tuned with this prompt.
temperature0.5-0.7Underlying R1 Distill uses this. I've found best results with 0.55.
top_p0.95Underlying R1 Distill uses this.

Quantization done using Unsloth.

Available quants:

QuantSizeNotes
F1629.55GBRetains 100% accuracy. Slow and memory hungry.
Q8_O15.70GBHigh resource use, but generally acceptable. Use when accuracy is crucial.
Q6_K12.12GBUses Q6_K for all tensors. Good for high end GPUs.
Q5KM10.51GBRecommended. Uses Q6K for half of the attention.wv and feedforward.w2 tensors, else Q5_K
Q4KM8.99GBRecommended. Uses Q6K for half of the attention.wv and feedforward.w2 tensors, else Q4_K
Q3KM7.34GBUses Q4K for the attention.wv, attention.wo, and feedforward.w2 tensors, else Q3_K. Quality is noticeably degraded.