Gemma 4 E2B text-only GGUF models compressed by TurinTech with Artemis. Recommended: gemma-4-E2B-it-Artemis-Q3_K_L.gguf. Model files, checksums and measured trade-offs: https://huggingface.co/turintech/gemma-4-E2B-it-Artemis-GGUF. Download size is not runtime memory. Measured Q3_K_L: Mean KLD 0.016, Same top p 95.8% vs BF16 on 20 fixed documents (the selection tier); protocol ryzen9950x-2026-10-12.
Gemma 4 E2B text-only GGUF models compressed by TurinTech with Artemis. Recommended: gemma-4-E2B-it-Artemis-Q3_K_L.gguf. Model files, checksums and measured trade-offs: https://huggingface.co/turintech/gemma-4-E2B-it-Artemis-GGUF. Download size is not runtime memory. Measured Q3_K_L: Mean KLD 0.016, Same top p 95.8% vs BF16 on 20 fixed documents (the selection tier); protocol ryzen9950x-2026-10-12.
Gemma 4 E2B text-only GGUF models compressed by TurinTech with Artemis. Recommended: gemma-4-E2B-it-Artemis-Q3_K_L.gguf. Model files, checksums and measured trade-offs: https://huggingface.co/turintech/gemma-4-E2B-it-Artemis-GGUF. Download size is not runtime memory. Measured Q3_K_L: Mean KLD 0.016, Same top p 95.8% vs BF16 on 20 fixed documents (the selection tier); protocol ryzen9950x-2026-10-12.
Gemma 4 E2B text-only GGUF models compressed by TurinTech with Artemis. Recommended: gemma-4-E2B-it-Artemis-Q3_K_L.gguf. Model files, checksums and measured trade-offs: https://huggingface.co/turintech/gemma-4-E2B-it-Artemis-GGUF. Download size is not runtime memory. Measured Q3_K_L: Mean KLD 0.016, Same top p 95.8% vs BF16 on 20 fixed documents (the selection tier); protocol ryzen9950x-2026-10-12.