venkataparswanadh/ qwen3.8-27b:iq1_s

55 1 week ago

... /
license
9e2d7afaad12 · 221B
Apache 2.0
Model weights: (c) Alibaba Group / Qwen team. Quantized with the IQ1_S
1-bit technique from unsloth / llama.cpp. YaRN rope scaling baked into
GGUF metadata provides 1M context (factor 4.0 over the 262144 base).