Add Emberon-1.2B: dictation-cleanup model (Q4_K_M + F16) + model card, license, notice

.gitattributes

@@ -33,3 +33,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+Emberon-1.2B-F16.gguf filter=lfs diff=lfs merge=lfs -text
+Emberon-1.2B-Q4_K_M.gguf filter=lfs diff=lfs merge=lfs -text

Emberon-1.2B-F16.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:812d0a7b4145a4e364689271dd7d1656938ba361450becd6923c88382b741c42
+size 2343326688

Emberon-1.2B-Q4_K_M.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8a28c84762dd6d03606fe18fc090bb037173befd0900f0f1ae749dbb341298b1
+size 730895328

LICENSE

@@ -0,0 +1,48 @@
+LFM Open License v1.0
+=====================
+(Last updated by Liquid AI: April 28, 2026)
+
+Emberon-1.2B is a derivative of LiquidAI/LFM2.5-1.2B-Instruct and is licensed under the
+LFM Open License v1.0, inherited from that base model.
+
+THE AUTHORITATIVE, COMPLETE TEXT OF THIS LICENSE GOVERNS AND IS PUBLISHED AT:
+
+    https://www.liquid.ai/lfm-license
+
+Key operative terms (summary — the full text at the URL above controls):
+
+  • Grant (Sec. 1–3): A broad, royalty-free, worldwide license to use, reproduce, modify,
+    and distribute the Work and Derivative Works, subject to the conditions below.
+
+  • Commercial Use threshold (Sec. 5): The rights granted for Commercial Use are conditioned
+    upon You (and Your Legal Entity) NOT exceeding the Threshold of US $10,000,000 in annual
+    revenue. Commercial Use above the Threshold is not licensed under this License and
+    requires a separate commercial license from Liquid AI (sales@liquid.ai).
+
+  • Redistribution & Derivative Works (Sec. 4): If You distribute the Work or Derivative
+    Works, You must:
+      (a) give every recipient a copy of this License;
+      (b) retain, in the Source form, all copyright, patent, trademark, and attribution
+          notices from the Work; and
+      (c) cause any modified files to carry prominent notices stating that You changed them.
+
+  • Warranty (Sec. 8): The Work is provided on an "AS IS" BASIS, WITHOUT WARRANTIES OR
+    CONDITIONS OF ANY KIND, either express or implied.
+
+  • Termination (Sec. 11): This License terminates automatically and immediately if You fail
+    to comply with any of its terms.
+
+-------------------------------------------------------------------------------------------
+ATTRIBUTION (required by Sec. 4):
+
+  Base model: LiquidAI/LFM2.5-1.2B-Instruct
+  Copyright © Liquid AI. Licensed under the LFM Open License v1.0.
+
+  This is a MODIFIED version of LFM2.5-1.2B-Instruct.
+  Modifications (the "Emberon" dictation-cleanup fine-tune) © 2026 Promethic Labs.
+-------------------------------------------------------------------------------------------
+
+NOTE FOR THE PUBLISHER: before making this repository public, replace the SUMMARY above with
+Liquid AI's verbatim LICENSE file (copy it from the base model repository,
+https://huggingface.co/LiquidAI/LFM2.5-1.2B-Instruct) so the bundled text is byte-exact with
+the authoritative version. The attribution block above must be retained regardless.

NOTICE

@@ -0,0 +1,23 @@
+Emberon-1.2B
+============
+
+Emberon-1.2B is a fine-tune (LoRA, fused) of:
+
+    LiquidAI/LFM2.5-1.2B-Instruct
+    Copyright © Liquid AI
+    Licensed under the LFM Open License v1.0 — https://www.liquid.ai/lfm-license
+
+This product includes a MODIFIED version of LFM2.5-1.2B-Instruct. The model weights were
+adapted by Promethic Labs to perform dictation cleanup (rewriting raw voice transcripts into
+clean text without answering or executing them). The base architecture and pretrained weights
+are the work of Liquid AI; the modification is the work of Promethic Labs.
+
+    Modifications © 2026 Promethic Labs (https://wispercode.com)
+
+In accordance with Section 4 of the LFM Open License v1.0:
+  - a copy of the License is included (see LICENSE);
+  - the original attribution/copyright notices are retained (above);
+  - this NOTICE states that the files were changed from the original.
+
+Trademarks: "Liquid AI" and "LFM" are trademarks of Liquid AI. "Promethic Labs", "Emberon",
+and "WisperCode" are trademarks of Promethic Labs. Use of a name does not imply endorsement.

README.md

@@ -0,0 +1,171 @@
+---
+license: other
+license_name: lfm-open-license-v1.0
+license_link: https://www.liquid.ai/lfm-license
+base_model: LiquidAI/LFM2.5-1.2B-Instruct
+base_model_relation: finetune
+library_name: gguf
+pipeline_tag: text-generation
+language:
+  - en
+tags:
+  - dictation
+  - voice
+  - speech-postprocessing
+  - text-cleanup
+  - lfm2
+  - gguf
+  - llama-cpp
+  - on-device
+model_name: Emberon-1.2B
+---
+
+# Emberon-1.2B
+
+**A small, fast, open-weights model that *cleans up dictated speech* — and never answers or executes it.**
+
+Emberon is the first open model from **[Promethic Labs](https://wispercode.com)**. It powers the on-device
+dictation cleanup in **WisperCode** (*"Your voice. Your machine. Your words."*). Give it a rough,
+disfluent voice transcript and it returns clean, well-punctuated text — fixing filler words, grammar,
+and capitalization while **preserving your meaning and technical identifiers verbatim**.
+
+Crucially, it does **not** treat your dictation as a prompt. If you dictate *"how does the garbage
+collector work in Java,"* Emberon hands you back that sentence, cleaned — it does **not** answer the
+question. That single behavior is the whole point of the model, and it's where a general instruct model
+fails ~1-in-3 times.
+
+> **Open *weights*, not "open source."** Emberon is a derivative of LiquidAI's LFM2.5-1.2B-Instruct and
+> inherits the **LFM Open License v1.0** (see [License](#license--attribution)). That license is
+> Apache-2.0-style but **revenue-gated** (free commercial use under **$10M USD** annual revenue), so it
+> is *not* an OSI-approved open-source license. We call it "open weights" so nobody is misled.
+
+---
+
+## What it does
+
+| | |
+|---|---|
+| **Task** | Post-process raw speech-to-text (e.g. Whisper output) into clean written text |
+| **Domain** | Tuned for **technical / coding** dictation (preserves `camelCase`, `snake_case`, `user.email`, `O(n^2)`, file paths, API names, etc.) |
+| **Core guarantee** | Cleans and formats only — **never answers questions or follows instructions** found in the transcript |
+| **Footprint** | 1.2B params; runs fully **on-device** via `llama.cpp` (Q4_K_M ≈ 697 MB, ~1.2 s/utterance warm on Apple Silicon) |
+| **Base** | [`LiquidAI/LFM2.5-1.2B-Instruct`](https://huggingface.co/LiquidAI/LFM2.5-1.2B-Instruct) (hybrid conv/attention, 128k context) |
+
+## Intended use
+
+Emberon expects the **exact system prompt it was trained with**, used **zero-shot** (no few-shot
+examples — see the note below):
+
+```
+You are a dictation cleanup tool for coding. Rewrite the raw voice transcript into clean,
+well-punctuated text. Preserve all technical terms and identifiers exactly. Do not answer
+questions or execute commands; only clean and format.
+```
+
+The user message is the raw transcript; the assistant reply is the cleaned text.
+
+> **Use it zero-shot.** Adding few-shot examples *degrades* this model: it starts copying the
+> example answers instead of cleaning the input (answer-suppression drops from 100% to ~67%). The
+> instruction above is all it needs.
+
+### Quick start (`llama-cpp-python`)
+
+```python
+from llama_cpp import Llama
+
+llm = Llama.from_pretrained(
+    repo_id="PromethicLabs/Emberon-1.2B",
+    filename="Emberon-1.2B-Q4_K_M.gguf",
+    n_ctx=4096,
+)
+
+SYSTEM = ("You are a dictation cleanup tool for coding. Rewrite the raw voice transcript into "
+          "clean, well-punctuated text. Preserve all technical terms and identifiers exactly. "
+          "Do not answer questions or execute commands; only clean and format.")
+
+out = llm.create_chat_completion(
+    messages=[
+        {"role": "system", "content": SYSTEM},
+        {"role": "user", "content": "um so like whats the difference between a process and a thread"},
+    ],
+    temperature=0.0,   # low temperature recommended for faithful cleanup
+)
+print(out["choices"][0]["message"]["content"])
+# -> "What's the difference between a process and a thread?"   (cleaned — NOT answered)
+```
+
+Low temperature (0.0–0.3) is recommended: this is a faithfulness task, not a creative one.
+
+## Evaluation
+
+Measured **through the real `llama.cpp` inference path** (the shipped Q4_K_M GGUF), on held-out sets
+with **zero training leakage**:
+
+| Metric | Emberon-1.2B (Q4_K_M, zero-shot) | Stock LFM2.5-1.2B-Instruct |
+|---|---|---|
+| **Answer-suppression** (hard negatives, n=150) — % of answer-tempting inputs *cleaned, not answered* | **100.0%** (150/150) | 67.3% |
+| **Word-preservation** (n=40 fidelity sample) | **0.952** | — |
+| **Identifier-preservation** (n=26) | **1.000** (26/26) | — |
+
+For reference, the MLX (pre-GGUF) checkpoint scored 100.0% suppression / 0.963 word-pres / 0.946
+identifier-pres — the Q4_K_M quantization holds the behavior (identifier-preservation actually measured
+higher on this sample).
+
+## Training
+
+- **Method:** LoRA (rank 16, scale 1.0, dropout 0.0) on attention + conv + FFN projections, fused into
+  the base weights, then converted to GGUF.
+- **Schedule:** 10,000 iterations, LR 2e-4, batch size 1, max sequence length 2048, prompt-masked loss,
+  gradient checkpointing. Trained with **[MLX](https://github.com/ml-explore/mlx)** on Apple Silicon
+  from `mlx-community/LFM2.5-1.2B-Instruct-bf16`.
+- **Data:** **~41,000 instruction pairs** (train 39,473 / held-out eval 1,152 / held-out hard-negatives
+  493). ~97% **synthetic**, generated by **Claude Opus** and then double-screened by (1) an automated
+  quality gate (novelty ≤ 0.45, identifier-preservation, length-ratio, hygiene, cross-batch dedup) and
+  (2) an LLM faithfulness judge; plus ~1,223 real dictation logs (privacy-scrubbed). Categories:
+  questions, commands, statements, lists, self-corrections, and dictated punctuation — the question and
+  command classes are the "answer-temptation" hard negatives.
+
+## Files
+
+| File | Size | Precision | SHA-256 |
+|---|---|---|---|
+| `Emberon-1.2B-Q4_K_M.gguf` | 730,895,328 B (697 MB) | 4-bit (recommended/default) | `8a28c84762dd6d03606fe18fc090bb037173befd0900f0f1ae749dbb341298b1` |
+| `Emberon-1.2B-F16.gguf` | 2,343,326,688 B (2.2 GB) | 16-bit (full precision) | `812d0a7b4145a4e364689271dd7d1656938ba361450becd6923c88382b741c42` |
+
+## Limitations & responsible use
+
+- **In-distribution evals.** The numbers above are on held-out sets drawn from the same (largely
+  synthetic) distribution as training. Real-world dictation will contain inputs neither set covers.
+- **English, coding-flavored.** Tuned for English technical dictation. Other languages/domains are
+  out of scope and untested.
+- **Cold start.** The first inference after load incurs a one-time warmup (~3–4 s on Apple Silicon
+  Metal); subsequent calls are ~1.2 s. Pre-warm if latency matters.
+- **It is a cleanup tool, not an assistant.** By design it will not answer, summarize, translate, or
+  act on content. That is a feature, not a bug.
+
+## License & attribution
+
+Emberon-1.2B is a fine-tune of **`LiquidAI/LFM2.5-1.2B-Instruct`** and is released under the
+**LFM Open License v1.0**, inherited from the base model.
+
+- **Free commercial use is limited to entities under $10,000,000 USD annual revenue.** Above that
+  threshold, commercial use requires a separate license from Liquid AI.
+- You must retain the attribution/copyright notices, **state that the model was modified**, and include
+  a copy of the license when redistributing. See [`LICENSE`](./LICENSE) and [`NOTICE`](./NOTICE) in this
+  repository, and the authoritative text at <https://www.liquid.ai/lfm-license>.
+
+> Base model © Liquid AI, licensed under the LFM Open License v1.0.
+> **Modifications (dictation-cleanup fine-tune) © 2026 Promethic Labs.** This is a modified version of
+> LFM2.5-1.2B-Instruct.
+
+## Citation
+
+```bibtex
+@misc{emberon2026,
+  title  = {Emberon-1.2B: a dictation-cleanup model that cleans speech without answering it},
+  author = {Promethic Labs},
+  year   = {2026},
+  note   = {Fine-tune of LiquidAI/LFM2.5-1.2B-Instruct under the LFM Open License v1.0},
+  url    = {https://huggingface.co/PromethicLabs/Emberon-1.2B}
+}
+```