Upload README.md with huggingface_hub

README.md

@@ -12,21 +12,36 @@ tags:
 
 EnCodec 24 kHz safetensors for MetaVoice / RLX encodec path.
 
-**Kind:** Converted / re-laid-out for RLX from an upstream checkpoint.
+| Field | Value |
+|---|---|
+| **Hub id** | [`eugenehp/encodec24`](https://huggingface.co/eugenehp/encodec24) |
+| **Kind** | Converted / re-laid-out for RLX from an upstream checkpoint. |
+| **RLX crate** | [`rlx-encodec`](https://github.com/MIT-RLX/rlx-models/tree/main/crates/rlx-encodec) |
+| **Upstream** | https://huggingface.co/facebook/encodec_24khz |
 
-**Suggested Hub repo:** `eugenehp/encodec24`
+## Quick start
 
-**Upstream:** https://huggingface.co/facebook/encodec_24khz
+```bash
+hf download eugenehp/encodec24 --local-dir .
+cargo test -p rlx-encodec --release
+```
+
+## File highlights
+
+- `model.safetensors` (88.8 MiB)
+- `encodec24.f16.gguf` (44.4 MiB)
+
+## Do not use as the main runtime pack
 
-**RLX crate:** [`rlx-encodec`](https://github.com/MIT-RLX/rlx-models/tree/main/crates/rlx-encodec)
+These files may be present for historical / staging reasons but are **not** the supported RLX load path:
 
-## Contents
+- `encodec24.f16.gguf`
 
-Weight tensors and configs in this directory (see file listing on the Hub).
+Community `*.f16.gguf` / LM-only Q4 packs are format wraps — they are not drop-in replacements for the RLX primary file above.
 
 ## Run with RLX
 
-From [rlx-models](https://github.com/MIT-RLX/rlx-models) with this directory as the model path:
+Clone [rlx-models](https://github.com/MIT-RLX/rlx-models), place this repo under `weights/tts/encodec24` (or pass the path explicitly), then:
 
 ```bash
 cargo test -p rlx-encodec --release
@@ -38,6 +53,6 @@ MIT — see `LICENSE`. Inherit upstream terms when redistributing.
 
 Original weights and authorship: https://huggingface.co/facebook/encodec_24khz
 
-## Staging note
+## Maintenance
 
-Prepared from the local `weights/` tree in [rlx-models](https://github.com/MIT-RLX/rlx-models) for a future standalone Hub upload. Prefer fetching upstream when this is only a redistrib.
+Cards and LFS attrs are regenerated from the local `weights/` tree in [rlx-models](https://github.com/MIT-RLX/rlx-models) via `python3 scripts/prepare_weights_hf.py`.