Skip to content

The open-source AI voice studio. Clone, dictate, create.

fix(mlx): point Qwen 0.6B at the published mlx-community repo (#501)

Jamie Pine8929947c7aae
diff --git a/backend/backends/__init__.py b/backend/backends/__init__.py
index db19b140..e90311c4 100644
--- a/backend/backends/__init__.py
+++ b/backend/backends/__init__.py
@@ -177,7 +177,7 @@ def _get_qwen_model_configs() -> list[ModelConfig]:
backend_type = get_backend_type()
if backend_type == "mlx":
repo_1_7b = "mlx-community/Qwen3-TTS-12Hz-1.7B-Base-bf16"
- repo_0_6b = "mlx-community/Qwen3-TTS-12Hz-1.7B-Base-bf16" # 0.6B not available in MLX, falls back
+ repo_0_6b = "mlx-community/Qwen3-TTS-12Hz-0.6B-Base-bf16"
else:
repo_1_7b = "Qwen/Qwen3-TTS-12Hz-1.7B-Base"
repo_0_6b = "Qwen/Qwen3-TTS-12Hz-0.6B-Base"
diff --git a/backend/backends/mlx_backend.py b/backend/backends/mlx_backend.py
index c6157a40..5691cd50 100644
--- a/backend/backends/mlx_backend.py
+++ b/backend/backends/mlx_backend.py
@@ -45,11 +45,9 @@ class MLXTTSBackend:
Returns:
HuggingFace Hub model ID for MLX
"""
- # MLX model mapping
mlx_model_map = {
"1.7B": "mlx-community/Qwen3-TTS-12Hz-1.7B-Base-bf16",
- # 0.6B not yet converted to MLX format
- "0.6B": "mlx-community/Qwen3-TTS-12Hz-1.7B-Base-bf16", # Fallback to 1.7B
+ "0.6B": "mlx-community/Qwen3-TTS-12Hz-0.6B-Base-bf16",
}
if model_size not in mlx_model_map:
diff --git a/backend/tests/test_qwen_download.py b/backend/tests/test_qwen_download.py
index b6c42e17..764dc90e 100644
--- a/backend/tests/test_qwen_download.py
+++ b/backend/tests/test_qwen_download.py
@@ -175,7 +175,7 @@ async def main():
print(" βœ… Server is running")
# Test model
- model_name = "qwen-tts-0.6B" # Note: 0.6B currently maps to 1.7B on MLX
+ model_name = "qwen-tts-0.6B"
# Check current status
print(f"\nπŸ“Š Checking status of {model_name}...")
diff --git a/docs/PROJECT_STATUS.md b/docs/PROJECT_STATUS.md
index cb12ef2e..5f0cb2d7 100644
--- a/docs/PROJECT_STATUS.md
+++ b/docs/PROJECT_STATUS.md
@@ -309,11 +309,7 @@ Still reported. Users get stuck downloads, can't resume, offline mode edge cases
**Fix path:** PR #443 addresses infinite offline retry. CustomVoice-specific download failures (#475, #445) need triage β€” likely related to frozen-binary import fixes in PR #438. TADA cluster (#336, #348) and macOS ARM import regressions (#287, #275, #304) need a dedicated triage pass.
-**Qwen 0.6B-downloads-1.7B reports:** **#485** (2026-04-19), **#423** (macOS M1), **#329**. Platform-dependent:
-
-- **On MLX (Apple Silicon) β€” not a bug.** `mlx-community` only publishes 1.7B-Base-bf16 weights, so the 0.6B Base option intentionally resolves to the same repo (`backend/backends/__init__.py:180` β€” `# 0.6B not available in MLX, falls back`). UX gap: the selector offers a size that doesn't exist on the active backend. Fix: (a) hide the 0.6B option on MLX, or (b) label it "0.6B (uses 1.7B on Apple Silicon)".
-- **On PyTorch (Windows/Linux/CUDA/ROCm/XPU/CPU) β€” real bug if reported.** Both 0.6B and 1.7B have distinct repos (`Qwen/Qwen3-TTS-12Hz-0.6B-Base` vs `-1.7B-Base`). Triage each report by platform before merging into the MLX cluster.
-- **Qwen CustomVoice (either platform)** β€” no fallback, both sizes always have dedicated repos.
+**Qwen 0.6B-downloads-1.7B reports:** **#485** (2026-04-19), **#423** (macOS M1), **#329**. Originally a stale-fallback bug: `mlx-community/Qwen3-TTS-12Hz-0.6B-Base-bf16` wasn't published when MLX support shipped, so the 0.6B slot was aliased to the 1.7B repo. The 0.6B bf16 conversion is live now and both `backend/backends/mlx_backend.py` and `backend/backends/__init__.py` point at their correct repos. Qwen CustomVoice is unaffected β€” it runs via PyTorch on all platforms, both sizes always have dedicated repos.
### Language Requests (ongoing)
@@ -393,7 +389,7 @@ Notable:
| **#306** ("voice model"), **#389** ("New model"), **#473** ("New functionality") | Title-only issues, no content. Request details or close. |
| **#309** | Uninstall/cleanup question. Answer and close. |
| **#241** | "How to use in Colab" β€” support question, not a bug. |
-| **#423** / **#485** / **#329** | Platform-dependent. On MLX: not a bug (0.6B weights don't exist upstream, fallback is intentional β€” fix UX). On PyTorch: real bug if reproducible. Classify each by reporter's platform before deduping. |
+| **#423** / **#485** / **#329** | Stale MLX fallback to 1.7B repo β€” fixed; 0.6B bf16 conversion now live on `mlx-community`, registry points at correct repo on both backends. |
| **#336** / **#348** | TADA download/registration cluster β€” triage together. |
| **#287** / **#275** / **#304** | macOS ARM import regressions on new version β€” likely one root cause. |
| **#292**, **#349** | Possibly already fixed by merged PRs (#321/#412 and #345). Verify + close. |
diff --git a/docs/content/docs/overview/troubleshooting.mdx b/docs/content/docs/overview/troubleshooting.mdx
index 59e759fb..f615b3c6 100644
--- a/docs/content/docs/overview/troubleshooting.mdx
+++ b/docs/content/docs/overview/troubleshooting.mdx
@@ -446,20 +446,6 @@ Restart the app to create a fresh database.
huggingface-cli download Qwen/Qwen3-TTS-12Hz-1.7B-Base
```
-### Qwen 0.6B Downloads the Same Files as 1.7B on Apple Silicon
-
-**Symptoms:**
-- You select Qwen 0.6B on an Apple Silicon Mac and the download is the same size as 1.7B
-- Generation speed and VRAM usage match 1.7B, not the expected smaller model
-
-**Explanation:**
-This is intentional, not a bug. The MLX community only publishes `mlx-community/Qwen3-TTS-12Hz-1.7B-Base-bf16` β€” there is no 0.6B MLX build. Voicebox's model registry falls back to the 1.7B weights when 0.6B is selected on MLX (see `backend/backends/__init__.py`).
-
-**Solution:**
-- On Apple Silicon, both size options use the 1.7B model β€” pick either.
-- If you specifically need a smaller model, switch to **Kokoro 82M** (~350 MB) or **LuxTTS** (~300 MB) β€” both CPU-realtime.
-- On Windows/Linux with PyTorch, 0.6B and 1.7B are distinct repos and behave differently.
-
### Wrong Model Version
**Symptoms:**