Thor Lin
coolthor
AI & ML interests
On-device & edge LLM inference on NVIDIA GB10 / DGX Spark.
Quantization (NVFP4 W4A4/W4A16, FP8), vLLM serving, speculative
decoding (EAGLE-3), and multimodal/omni models. Measure-first
benchmarking — I publish the numbers, including the ones that fail.
Recent Activity
updated a model 4 days ago
coolthor/Kandinsky-6.0-Pro-distill-5s-NVFP4 published a model 4 days ago
coolthor/Kandinsky-6.0-Pro-distill-5s-NVFP4 updated a model 13 days ago
coolthor/Qwen3.8-27B-Uncensored-NVFP4-FastLLMOrganizations
None yet
update chat template
1
#2 opened 3 months ago
by 
Sidd065
Update chat template
1
#1 opened 3 months ago
by Sidd065
Model hallucinates
2
#1 opened 5 months ago
by 
johnlockejrr