{ "model_type": "antalia-mini", "format_version": 1, "version": "1.0.1", "sample_rate": 48000, "acoustic": { "n_symbols": 47, "n_mels": 128, "d_text": 192, "text_convnext": 3, "text_hidden": 384, "text_attn": 1, "heads": 4, "d_dur": 128, "d_gen": 192, "gen_blocks": 5, "gen_hidden": 512, "d_time": 128, "stack": 2, "cond_drop": 0.1, "shortcut_frac": 0.25, "shortcut_levels": 7, "mel_mean": -5.9, "mel_std": 2.6, "max_dur_frames": 60 }, "vocoder": { "n_mels": 128, "dim": 256, "hidden": 768, "layers": 8, "n_fft": 2048, "hop": 512, "sample_rate": 48000 }, "symbols": [ "", "", "", " ", ".", ",", "?", "!", "'", "-", ":", ";", "a", "b", "c", "ç", "d", "e", "f", "g", "ğ", "h", "ı", "i", "j", "k", "l", "m", "n", "o", "ö", "p", "q", "r", "s", "ş", "t", "u", "ü", "v", "w", "x", "y", "z", "â", "î", "û" ], "defaults": { "steps": 8, "cfg": 2.0, "speed": 0.95, "temperature": 0.6 }, "options": { "suffix": " ee.", "prefix": "ee, ", "prefix_initials": "öı", "gem_stretch": 1.3, "curve": [ [ "hpf", 80, 0.707 ], [ "low", 200, -4, 0.707 ], [ "peak", 350, -2.5, 1.2 ], [ "peak", 3200, 3, 0.9 ], [ "high", 8000, 2, 0.707 ] ], "deess": { "f0": 5500, "threshold_db": -34, "ratio": 3 }, "makeup_db": 1.6 }, "params": { "acoustic": 3686977, "vocoder": 3933954, "total": 7620931 }, "source": { "acoustic": { "file": "ac-main-short.pt", "step": 51880, "weights": "ema" }, "vocoder": { "file": "voc-gta3-76860.pt", "step": 76860 } }, "files": { "acoustic.safetensors": { "bytes": 14757148, "sha256": "72c36a6855ce95dc1d35960a7b61515136813313406e5b9ae1c2ebe6cd2ad428" }, "vocoder.safetensors": { "bytes": 15742616, "sha256": "89ca374cc55864b38a36f9ea87c41e1b17376e5d6d68e3a0458332861509bbf6" }, "acoustic.fp16.safetensors": { "bytes": 7383130, "sha256": "c94d5cbe11df77dbc293a85782c28eac377a5f9b8027ca2270cc9b2b51f35d87" }, "vocoder.fp16.safetensors": { "bytes": 7874644, "sha256": "a5ba4b559daccf93c0ac888d50a198c810ab95554b3be6df937dac690d6fbcee" } } }