CaptchaKraken/SunlightV2-0.8B
0.9B • Updated • 2
CaptchaKraken v2
v2 release: element detector (CNN) and pointer models.
Note 0.8B: merged bf16 (vLLM, GPU)
CaptchaKraken/SunlightV2-0.8B-q4
0.9B • Updated • 2Note 0.8B: 4-bit GPTQ W4A16 (vLLM, GPU)
CaptchaKraken/SunlightV2-0.8B-LoRA
UpdatedNote 0.8B: LoRA adapter + pointer heads
CaptchaKraken/SunlightV2-0.8B-GGUF
0.8B • Updated • 51Note 0.8B: GGUF for llama.cpp (CPU)
CaptchaKraken/TwilightV2-2B
2B • Updated • 2Note 2B: merged bf16 (vLLM, GPU)
CaptchaKraken/TwilightV2-2B-LoRA
Updated • 2Note 2B: LoRA adapter + pointer heads
CaptchaKraken/TwilightV2-2B-GGUF
2B • Updated • 50Note 2B: GGUF for llama.cpp (CPU)
CaptchaKraken/MidnightV2-4B
5B • Updated • 2Note 4B: merged bf16 (vLLM, GPU)
CaptchaKraken/MidnightV2-4B-q4
5B • Updated • 3Note 4B: 4-bit GPTQ W4A16 (vLLM, GPU)
CaptchaKraken/MidnightV2-4B-LoRA
Updated • 1Note 4B: LoRA adapter + pointer heads
CaptchaKraken/MidnightV2-4B-GGUF
4B • Updated • 51Note 4B: GGUF for llama.cpp (CPU)
CaptchaKraken/CaptchaKrakenV2-CNN-large
UpdatedNote CNN large: detector.onnx + detector_r18.pt
CaptchaKraken/CaptchaKrakenV2-CNN-lite
UpdatedNote CNN lite: detector-lite.onnx (1-2 vCPUs)