Links: repos moved to the Fritter-ASR org

index.html

@@ -4,7 +4,7 @@
 table{border-collapse:collapse;width:100%;margin:1rem 0 2rem}td,th{border-bottom:1px solid #ddd;padding:.5rem;vertical-align:top;text-align:left}
 audio{width:220px;height:32px}th{background:#f6f6f6}.t{max-width:420px}</style></head><body>
 <h1>Fritter-TTS v1: listening samples</h1>
-<p>Fritter-TTS reads technical and general English (model names, Linux commands, acronyms, version numbers) in Daniel Rosehill's voice. A fine-tune of Qwen3-TTS-12Hz-1.7B-Base. Model: <a href="https://huggingface.co/danielrosehill/Fritter-TTS">danielrosehill/Fritter-TTS</a> (access by request).</p>
+<p>Fritter-TTS reads technical and general English (model names, Linux commands, acronyms, version numbers) in Daniel Rosehill's voice. A fine-tune of Qwen3-TTS-12Hz-1.7B-Base. Model: <a href="https://huggingface.co/Fritter-ASR/Fritter-TTS">Fritter-ASR/Fritter-TTS</a> (access by request).</p>
 <p>No sample here was picked by ear: these are every line the release run generated, unedited. Real takes are Daniel's own
 recordings after the same cleaning chain as the training data. "Zero-shot" is the untuned base model cloning the voice from one reference take.</p>
 <h2>Before and after</h2><p><b>Before</b>: the untuned base model, Qwen3-TTS-12Hz-1.7B-Base, cloning Daniel's voice from one reference recording. <b>After</b>: Fritter-TTS v1. Same text, nothing picked by ear. Passages of more than three sentences are generated sentence by sentence and joined.</p><table><tr><th>Text</th><th>Before</th><th>After</th></tr>