Install from WinGet (Windows)
winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf BadXStudio/pigeon-gemma-4-dictation:Q4_K_M# Run inference directly in the terminal:
llama cli -hf BadXStudio/pigeon-gemma-4-dictation:Q4_K_MUse pre-built binary
# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf BadXStudio/pigeon-gemma-4-dictation:Q4_K_M# Run inference directly in the terminal:
./llama-cli -hf BadXStudio/pigeon-gemma-4-dictation:Q4_K_MBuild from source code
git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf BadXStudio/pigeon-gemma-4-dictation:Q4_K_M# Run inference directly in the terminal:
./build/bin/llama-cli -hf BadXStudio/pigeon-gemma-4-dictation:Q4_K_MUse Docker
docker model run hf.co/BadXStudio/pigeon-gemma-4-dictation:Q4_K_MPigeon Gemma 4 โ Dictation Refinement v2 (E2B + E4B, Q4_K_M GGUF)
Fine-tuned Gemma 4 E2B and E4B models that rewrite rough spoken dictation into clean written text โ the AI refinement layer of Pigeon, a local Windows dictation app. Everything runs locally via llama.cpp; no cloud.
What they do: fix self-corrections ("tell mark no actually johnโฆ" โ "Tell Johnโฆ"), remove disfluencies, restore punctuation, preserve names/facts/politeness, and resist instructions embedded in the dictation (dictating "write me a poem" outputs the refined request โ it does not write the poem).
This is the v2 (round-3) release, superseding the earlier v1 fine-tunes.
Files
| File | Size | What |
|---|---|---|
gemma4-e4b-pigeon-tuned-v2.Q4_K_M.gguf |
5.3 GB | E4B fine-tune โ best quality, ~0.24 s/refine on GPU |
gemma4-e2b-pigeon-tuned-v2.Q4_K_M.gguf |
3.4 GB | E2B fine-tune โ fastest, ~0.18 s/refine on GPU, ~2 s on CPU |
adapters/e4b/ |
~0.3 GB | E4B round-3 LoRA adapter (r=32, ฮฑ=64, all-linear) โ retraining/re-merge source |
adapters/e2b/ |
~0.3 GB | E2B round-3 LoRA adapter (r=32, ฮฑ=64, all-linear) โ retraining/re-merge source |
Training
Sequence-level knowledge distillation: a 3,654-pair dictation-cleanup dataset was
generated by a frontier teacher model, Qwen3.5-122B-A10B (Apache 2.0), then Gemma 4
was SFT'd on the (raw โ refined) pairs. LoRA r=32/ฮฑ=64 on all linear projections, 3
epochs, completion-only masking on Gemma-4 <|turn> markers, bf16. The identical system
prompt is used for training and inference (embedded expectation โ see Usage).
Measured evaluation (internal 825-case deterministic Gauntlet eval)
| Model | Accuracy | Notes |
|---|---|---|
| E4B-v2 | 87.9% | beats the reference teacher-class model at ~6x the speed |
| Qwen 3.5 9B (reference) | 81.1% | frontier-class reference model, ~6x slower |
| E2B-v2 | 83.2% | smallest/fastest tier |
Full methodology in the Pigeon repo (docs/EVAL_REPORT.md).
Intended use
These weights are the refine layer inside the Pigeon local dictation app: raw speech-to-text output is passed through the model to produce clean, punctuated, self-correction-resolved text before it lands in the user's document. They are not general-purpose chat or instruction-following models, and injection resistance is trained specifically for this refine-layer role.
Usage (llama.cpp / llama-cpp-python)
The GGUFs carry the correct Gemma-4 <|turn> chat template in their metadata โ pass no
chat_format override. Inference uses the byte-exact system prompt the models were
trained on (shipped in the Pigeon app itself; not reproduced here โ the models expect
ONLY that system message plus the raw dictation as the user turn). Recommended sampling:
temperature=0.2, top_p=0.9.
Limitations
- No tone conditioning โ trained on a single neutral style; tone instructions are ignored.
- Self-correction handling is strong but not perfect; some "โฆno actually Xโฆ" phrasings with trailing clauses can resolve to the wrong item.
- English only. Q4_K_M quantization only (LoRA adapters included for re-quantization).
License & use restrictions
These are derivatives of Google Gemma 4 (via unsloth/gemma-4-E2B-it /
unsloth/gemma-4-E4B-it) and are distributed under the
Gemma Terms of Use. Use of these models is subject to
the Gemma Prohibited Use Policy, which
downstream users must also comply with and pass on. "Gemma is provided under and subject to
the Gemma Terms of Use found at ai.google.dev/gemma/terms."
- Downloads last month
- 78
4-bit
Install (macOS, Linux)
# Start a local OpenAI-compatible server with a web UI: llama serve -hf BadXStudio/pigeon-gemma-4-dictation:Q4_K_M# Run inference directly in the terminal: llama cli -hf BadXStudio/pigeon-gemma-4-dictation:Q4_K_M