Sanitize public model card metadata

README.md

@@ -1,60 +1,18 @@
 ---
-license: llama3.2
+license: "llama3.2"
 tags:
-  - qhexrt
-  - llama
-  - hexagon
-  - npu
-  - qnn
-base_model: meta-llama/Llama-3.2-1B
-pipeline_tag: text-generation
-language:
-  - en
+- "hnpu"
+- "hexagon"
+- "npu"
+- "llm"
+base_model: "meta-llama/Llama-3.2-1B"
+pipeline_tag: "text-generation"
 ---
 
-# Llama-3.2-1B — QHexRT NPU bundle (Hexagon v79)
+# llama3 2 1b HNPU
 
-Precompiled **Llama-3.2-1B** for the **QHexRT** runtime on Qualcomm Hexagon **v79** (Snapdragon 8 Elite /
-SM8750, e.g. Galaxy S25). Runs **entirely on the NPU**: the full 16-layer transformer **and** the lm-head
-execute on HTP/HMX via the `llama_generate` host-op (batched prefill + resident-KV decode); the host does
-only embed lookup, llama3-scaled RoPE, the cache mask, and sampling. Device-validated, coherent output.
+Prebuilt HNPU artifacts for [meta-llama/Llama-3.2-1B](https://huggingface.co/meta-llama/Llama-3.2-1B), a public text-generation model.
 
-**Beats Genie.** On the same device, QHexRT decode is **~16 tok/s vs Genie's 14.69** (fp16), and prefill
-**1983 tok/s vs 1538** — because QHexRT runs a prepared QNN context binary with an **in-graph NPU lm-head**
-(the host lm-head was the old bottleneck) and **W8A16** weights (half fp16's memory bandwidth).
+For model behavior, license, intended use, and limitations, see the [upstream model card](https://huggingface.co/meta-llama/Llama-3.2-1B).
 
-## Contents (`v79/`)
-| file | what | size |
-|------|------|------|
-| `llama-3.2-1b.json` | QHexRT manifest (llm family, `llama_generate` plan) | — |
-| `llama_full_wqo_o3.bin` | **W8A16 batched prefill** graph (AR=128) | 982 MB |
-| `llama_dec_wqo_o3.bin` | **W8A16 decode** graph (MAXCTX=512, GQA 32q/8kv, head_dim 64) | 981 MB |
-| `llama_lmhead.bin` | **NPU lm-head** graph — `hidden[1,2048] → logits[1,128256]`, **fp16** on HMX | 526 MB |
-| `llama_embed_f16.bin` | token embedding table `[128256,2048]` f16 (tied; host lookup) | 525 MB |
-| `tokenizer.json` | Llama-3 tokenizer (vocab 128256) | — |
-
-≈ **3 GB** on disk; ~2.5 GB peak device RSS.
-
-## Run (QHexRT CLI)
-```bash
-hf download runanywhere/llama3_2_1b_HNPU --local-dir llama3_2_1b_HNPU
-adb push llama3_2_1b_HNPU/v79 /data/local/tmp/wq/llama
-adb shell "cd /data/local/tmp/wq && export ADSP_LIBRARY_PATH='/data/local/tmp/wq/dsp;/data/local/tmp/wq;/vendor/dsp/cdsp'; \
-  LD_LIBRARY_PATH=. ./qhx_generate llama/llama-3.2-1b.json libQnnHtp.so libQnnSystem.so llama 64 'The capital of France is'"
-# -> "Paris. The Eiffel Tower is located in Paris. ..."  (decode ~16 tok/s, prefill ~1983 tok/s)
-```
-
-## Performance (measured, v79 / S25)
-| metric | QHexRT | Genie (fp16) |
-|--------|--------|--------------|
-| decode | **~16 tok/s** | 14.69 |
-| prefill | **1983 tok/s** | 1538 |
-
-## Notes
-- Arch: **v79 only** — context binaries are dsp-arch-pinned.
-- No custom op-package — pure-native HTP graphs (W8A16 matmuls on HMX).
-- **Prompt ≤128 tokens** (the batched-prefill graph is AR=128); generation extends to MAXCTX=512.
-- Precision: **W8A16** decode/prefill + **fp16** lm-head. W4A16 was tried but collapses a 1B model's
-  coherence (uniform 4-bit is too coarse here); W8A16 is the coherent sweet spot and still beats Genie.
-- Source: [`meta-llama/Llama-3.2-1B`](https://huggingface.co/meta-llama/Llama-3.2-1B), compiled with
-  QAIRT 2.45 for `qualcomm-snapdragon-8-elite-for-galaxy`.
+Artifacts are architecture-pinned. Available artifact directories: `v79/`, `v81/`.