gemma-4-E4B-it-dspark-GGUF

GGUF build of rasyosef/gemma-4-E4B-it-dspark for llama.cpp (DSpark speculative decoding is in mainline, ggml-org/llama.cpp #25173).

This is a standalone draft sidecar: it carries only the drafter (4 attention layers, rank-256 Markov head, confidence head, block size 8, 32,000-token draft vocabulary). Token embeddings and the LM head are shared from the target model at load time, so it must be paired with a gemma-4-E4B-it GGUF target file, for example one from unsloth/gemma-4-E4B-it-GGUF.

📦 Files

Choose the draft quant by memory budget. The target model quant is the main speed/quality lever and is independent of this file.

🏃 How to run (llama.cpp)

llama-server -hf unsloth/gemma-4-E4B-it-GGUF:Q8_0 \
  -hfd rasyosef/gemma-4-E4B-it-dspark-GGUF:Q8_0 \
  --spec-type draft-dspark --spec-draft-n-max 8 --spec-draft-n-min 0 \
  -fa on -ngl 99

The block size is read from the sidecar metadata (n-max is clamped to it). Speculative decoding is exact: the target verifies every proposed token, so greedy output equals the target alone; per-response timings report draft_n / draft_n_accepted.

Other models in the DSpark GGUF family:

📊 Acceptance and benchmarks

See rasyosef/gemma-4-E4B-it-dspark for acceptance-length metrics and training details.