gemma3_1bInstructQ4 constant

SyniModelSpec const gemma3_1bInstructQ4

Gemma 3 1B Instruct, Q4_K_M quantization. ~770 MB. Hosted in the Synheart-owned, non-gated mirror so a single org-controlled source covers both GGUF and tokenizer.

Smaller than qwen25_15bInstructQ4 — useful when the platform is memory-constrained or first-install latency matters more than peak instruction quality.

Implementation

static const SyniModelSpec gemma3_1bInstructQ4 = SyniModelSpec(
  id: 'gemma-3-1b-it-q4_k_m',
  filename: 'gemma-3-1b-it-q4_k_m.gguf',
  downloadUrl:
      'https://huggingface.co/Synheart/syni-life-gguf-gemma-3-1b/resolve/main/google_gemma-3-1b-it-Q4_K_M.gguf',
  // Tokenizer is hosted in the same Synheart repo — `tokenizer.json`
  // must be uploaded alongside the GGUF for installs to succeed.
  tokenizerUrl:
      'https://huggingface.co/Synheart/syni-life-gguf-gemma-3-1b/resolve/main/tokenizer.json',
  sha256: '12bf0fff8815d5f73a3c9b586bd8fee8e7b248c935de70dec367679873d0f29d',
  approxBytes: 806058496, // ~770 MB
);