metadata
license: apache-2.0
base_model: Jinx-org/Jinx-Qwen3-4B
base_model_relation: quantized
Recommended
Jinx-Qwen3-4B-gguf-q6_k-q-8 (mixed-precision): selected weights (output, token embeddings, attention/FFN layers in first and last blocks) quantized to Q8_0, remaining tensors Q6_k, reducing memory footprint while preserving inference fidelity.