eugenehp commited on
Commit
e0bf164
·
verified ·
1 Parent(s): 9330047

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +28 -9
README.md CHANGED
@@ -13,21 +13,40 @@ tags:
13
 
14
  ZipVoice ONNX text encoder + flow decoder for RLX.
15
 
16
- **Kind:** Converted / re-laid-out for RLX from an upstream checkpoint.
 
 
 
 
 
17
 
18
- **Suggested Hub repo:** `eugenehp/zipvoice`
19
 
20
- **Upstream:** https://huggingface.co/k2-fsa/ZipVoice
 
 
 
 
 
 
 
 
 
 
 
 
 
 
21
 
22
- **RLX crate:** [`rlx-zipvoice`](https://github.com/MIT-RLX/rlx-models/tree/main/crates/rlx-zipvoice)
23
 
24
- ## Contents
25
 
26
- Weight tensors and configs in this directory (see file listing on the Hub).
27
 
28
  ## Run with RLX
29
 
30
- From [rlx-models](https://github.com/MIT-RLX/rlx-models) with this directory as the model path:
31
 
32
  ```bash
33
  cargo run -p rlx-zipvoice --release -- --model-dir .
@@ -39,6 +58,6 @@ Apache License 2.0 — see `LICENSE`. Inherit upstream terms when redistributing
39
 
40
  Original weights and authorship: https://huggingface.co/k2-fsa/ZipVoice
41
 
42
- ## Staging note
43
 
44
- 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.
 
13
 
14
  ZipVoice ONNX text encoder + flow decoder for RLX.
15
 
16
+ | Field | Value |
17
+ |---|---|
18
+ | **Hub id** | [`eugenehp/zipvoice`](https://huggingface.co/eugenehp/zipvoice) |
19
+ | **Kind** | Converted / re-laid-out for RLX from an upstream checkpoint. |
20
+ | **RLX crate** | [`rlx-zipvoice`](https://github.com/MIT-RLX/rlx-models/tree/main/crates/rlx-zipvoice) |
21
+ | **Upstream** | https://huggingface.co/k2-fsa/ZipVoice |
22
 
23
+ ## Quick start
24
 
25
+ ```bash
26
+ hf download eugenehp/zipvoice --local-dir .
27
+ cargo run -p rlx-zipvoice --release -- --model-dir .
28
+ ```
29
+
30
+ ## File highlights
31
+
32
+ - `fm_decoder.onnx` (455.3 MiB)
33
+ - `zipvoice.f16.gguf` (267.9 MiB)
34
+ - `onnx/vocoder_spec.onnx` (51.6 MiB)
35
+ - `text_encoder.onnx` (16.8 MiB)
36
+ - `encoder_body.onnx` (16.8 MiB)
37
+ - `tokens.txt` (2.5 KiB)
38
+
39
+ ## Do not use as the main runtime pack
40
 
41
+ These files may be present for historical / staging reasons but are **not** the supported RLX load path:
42
 
43
+ - `zipvoice.f16.gguf`
44
 
45
+ Community `*.f16.gguf` / LM-only Q4 packs are format wraps they are not drop-in replacements for the RLX primary file above.
46
 
47
  ## Run with RLX
48
 
49
+ Clone [rlx-models](https://github.com/MIT-RLX/rlx-models), place this repo under `weights/tts/zipvoice` (or pass the path explicitly), then:
50
 
51
  ```bash
52
  cargo run -p rlx-zipvoice --release -- --model-dir .
 
58
 
59
  Original weights and authorship: https://huggingface.co/k2-fsa/ZipVoice
60
 
61
+ ## Maintenance
62
 
63
+ 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`.