Name the model accurately: Holotron-12B (NVIDIA Nemotron-derived)

README.md

@@ -46,7 +46,7 @@ Puck is a small, mischievous creature that lives on your screen. He **roams**, *
 
 ## Built Small
 Puck is a constellation of small, laptop-sized models — **nothing over 32B**, and it fits in your pocket:
-- 👁️ **Vision + voice-of-the-fairy:** Nemotron-Nano-**12B**-VL (a single VLM that both *sees* and *speaks* the quip).
+- 👁️ **Vision + voice-of-the-fairy:** **Holotron-12B** — H Company's computer-use VLM post-trained from NVIDIA's Nemotron-Nano-12B-VL (one model that both *sees* and *speaks* the quip).
 - 🔤 **Tool/site recognition:** on-device **Vision OCR** + an **ONNX CLIP ViT-B/32** (~88M) fingerprinter.
 - 🗣️ **Neural voice:** **Kokoro-82M**, running in-browser (WebGPU) — a fairy voice in 82M params.
 

submission/FIELD_NOTES.md

@@ -19,7 +19,7 @@ quip.** ~95% ambient, near-zero interruptions. He's "marginally useful" on purpo
 
 The instinct is a pipeline: vision model describes the patch → text model writes a quip in
 character. Two calls, two models, more latency, more ways to drift. Instead we give a single
-12B VLM (NVIDIA's Nemotron-Nano-VL) a system prompt that *is* Puck, and ask it to react to the
+12B VLM — Holotron-12B, post-trained from NVIDIA's Nemotron-Nano-VL — a system prompt that *is* Puck, and ask it to react to the
 image directly. The model that sees is the model that speaks. Fewer moving parts, and the quip
 is grounded in pixels rather than in a lossy description of them.
 
@@ -93,4 +93,4 @@ act, learned relevance to decide, nightly fine-tuning to improve — the differe
 gimmick and a familiar that genuinely grows up alongside you. He starts marginally useful on
 purpose, so that by the time he's useful for real, you already trust him.
 
-— Built with Hugging Face · Modal · NVIDIA Nemotron.
+— Built with Hugging Face · Modal · Holotron-12B (post-trained from NVIDIA Nemotron).