vae / app.py
Humair332's picture
Update app.py
7140878 verified
Raw
History Blame Contribute Delete
11.6 kB
import gradio as gr
import torch
import numpy as np
import soundfile as sf
from scipy.signal import resample as scipy_resample
from dataclasses import dataclass, field
from huggingface_hub import hf_hub_download
import time
import json
# =============================
# DACVAE WRAPPER
# =============================
@dataclass
class SimpleDACCodec:
model: torch.nn.Module
sample_rate: int
hop_size: int # encoder stride in samples β€” probed at load time
device: torch.device
@classmethod
def load(cls, repo_id="Aratako/Semantic-DACVAE-Japanese-32dim", device="cpu"):
from dacvae import DACVAE
weights_path = hf_hub_download(repo_id=repo_id, filename="weights.pth")
model = DACVAE.load(weights_path).eval().to(device)
sr = int(model.sample_rate)
# ── Probe the real hop size ───────────────────────────────────────────
# We feed a known-length signal and measure how many frames come out.
# This is the only correct way β€” no magic constants needed.
# hop = input_samples / output_frames (for a signal long enough to
# avoid edge effects we use 1 second = sr samples)
probe_len = sr # exactly 1 second of silence
dummy = torch.zeros(1, 1, probe_len, device=device,
dtype=next(model.parameters()).dtype)
with torch.inference_mode():
z = model.encode(dummy) # (1, D, T_latent)
t_latent = z.shape[2]
hop = probe_len // t_latent # integer hop in samples
print(f"[codec] sample_rate={sr} probe_frames={t_latent} "
f"hop={hop} frame_rate={sr/hop:.4f} Hz", flush=True)
return cls(
model = model,
sample_rate = sr,
hop_size = hop,
device = torch.device(device),
)
@property
def frame_rate(self) -> float:
"""Latent frames per second."""
return self.sample_rate / self.hop_size
def frames_to_seconds(self, num_frames: int) -> float:
"""Convert latent frame count -> audio duration in seconds."""
return num_frames * self.hop_size / self.sample_rate
@torch.inference_mode()
def encode(self, audio: torch.Tensor) -> torch.Tensor:
"""audio: (1, 1, T) -> latent: (1, T_latent, D)"""
z = self.model.encode(audio) # (B, D, T)
return z.transpose(1, 2) # (B, T, D)
@torch.inference_mode()
def decode(self, latent: torch.Tensor) -> torch.Tensor:
"""latent: (B, T_latent, D) -> audio: (B, 1, T)"""
return self.model.decode(latent.transpose(1, 2))
# =============================
# INIT
# =============================
DEVICE = "cuda" if torch.cuda.is_available() else "cpu"
print(f"[init] Using device: {DEVICE}")
codec = SimpleDACCodec.load(device=DEVICE)
print(f"[init] Codec ready. Frame rate = {codec.frame_rate:.4f} Hz "
f"(hop={codec.hop_size}, sr={codec.sample_rate})")
# =============================
# AUDIO UTILS
# =============================
def load_audio(path: str) -> tuple[np.ndarray, int]:
audio, sr = sf.read(path, dtype="float32")
if audio.ndim > 1:
audio = np.mean(audio, axis=1)
return audio, sr
def resample_audio(audio: np.ndarray, orig_sr: int, target_sr: int) -> np.ndarray:
if orig_sr == target_sr:
return audio
num_samples = int(len(audio) * target_sr / orig_sr)
return scipy_resample(audio, num_samples)
def to_tensor(audio: np.ndarray) -> torch.Tensor:
return torch.from_numpy(audio).unsqueeze(0).unsqueeze(0) # (1, 1, T)
def format_stats(stats: dict) -> str:
"""Render stats dict as a clean markdown table for display."""
lines = ["| Property | Value |", "|---|---|"]
for k, v in stats.items():
lines.append(f"| {k} | `{v}` |")
return "\n".join(lines)
# =============================
# ENCODE
# =============================
def encode_audio(file):
if file is None:
return None, None, "⚠️ Please upload an audio file first."
t0 = time.perf_counter()
# Load + resample
audio_orig, sr_orig = load_audio(file)
orig_samples = len(audio_orig)
orig_duration = orig_samples / sr_orig
audio_resampled = resample_audio(audio_orig, sr_orig, codec.sample_rate)
resampled_samples = len(audio_resampled)
wav = to_tensor(audio_resampled).to(DEVICE)
# Encode
latent = codec.encode(wav) # (1, T_latent, D)
t_enc = time.perf_counter() - t0
num_frames = latent.shape[1]
latent_dim = latent.shape[2]
calc_dur = codec.frames_to_seconds(num_frames)
latent_np = latent.squeeze(0).detach().cpu().numpy() # (T, D)
latent_list = latent_np.tolist()
# Stats
stats = {
"πŸ“ Original sample rate": f"{sr_orig} Hz",
"🎡 Codec sample rate": f"{codec.sample_rate} Hz",
"⏱ Original duration": f"{orig_duration:.4f} s ({orig_samples:,} samples)",
"⏱ Resampled duration": f"{resampled_samples / codec.sample_rate:.4f} s ({resampled_samples:,} samples)",
"πŸ”’ Latent frames (T)": f"{num_frames}",
"πŸ“ Latent dim (D)": f"{latent_dim}",
"πŸ“ Encoder hop size": f"{codec.hop_size} samples",
"πŸ”„ Latent frame rate": f"{codec.frame_rate:.4f} Hz",
"⏳ Duration from latent": f"{calc_dur:.4f} s (T Γ— hop / sr = {num_frames} Γ— {codec.hop_size} / {codec.sample_rate})",
"βœ… Duration match": f"{'βœ“ exact' if abs(calc_dur - resampled_samples / codec.sample_rate) < 0.05 else '⚠ mismatch'}",
"⚑ Encode time": f"{t_enc*1000:.1f} ms",
"πŸ’Ύ Latent tensor size": f"{latent_np.nbytes / 1024:.1f} KB (float32)",
"πŸ“Š Latent value range": f"[{latent_np.min():.4f}, {latent_np.max():.4f}]",
"πŸ“Š Latent mean / std": f"{latent_np.mean():.4f} / {latent_np.std():.4f}",
}
stats_md = format_stats(stats)
return latent_list, latent_list, stats_md
# =============================
# DECODE
# =============================
def decode_audio(latent_list, stats_md_current):
if latent_list is None:
return None, (stats_md_current or "") + "\n\n⚠️ No latent found. Encode first."
t0 = time.perf_counter()
try:
latent = torch.tensor(latent_list, dtype=torch.float32, device=DEVICE)
except Exception as e:
return None, f"⚠️ Invalid latent: {e}"
if latent.ndim == 2:
latent = latent.unsqueeze(0) # (1, T, D)
audio = codec.decode(latent) # (B, 1, T_out)
t_dec = time.perf_counter() - t0
audio_np = audio.squeeze().detach().cpu().numpy()
audio_np = np.nan_to_num(audio_np)
audio_np = np.clip(audio_np, -1.0, 1.0)
num_frames = latent.shape[1]
out_samples = len(audio_np)
actual_dur = out_samples / codec.sample_rate
calc_dur = codec.frames_to_seconds(num_frames)
actual_hop = out_samples // num_frames
decode_stats = {
"πŸ”’ Latent frames decoded": f"{num_frames}",
"πŸ”Š Output samples": f"{out_samples:,}",
"⏱ Reconstructed duration": f"{actual_dur:.4f} s",
"⏳ Duration from latent": f"{calc_dur:.4f} s",
"πŸ” Actual output hop": f"{actual_hop} samples/frame (expected {codec.hop_size})",
"βœ… Formula confirmation": f"T={num_frames} Γ— hop={actual_hop} / sr={codec.sample_rate} = {num_frames * actual_hop / codec.sample_rate:.4f} s",
"⚑ Decode time": f"{t_dec*1000:.1f} ms",
"πŸ“Š Output value range": f"[{audio_np.min():.4f}, {audio_np.max():.4f}]",
}
decode_md = format_stats(decode_stats)
combined = (stats_md_current or "") + "\n\n### Decode Stats\n" + decode_md
return (codec.sample_rate, audio_np), combined
# =============================
# UI
# =============================
css = """
body, .gradio-container {
background: #0d0d0d !important;
font-family: 'IBM Plex Mono', monospace !important;
color: #e0e0e0 !important;
}
h1, h2, h3 { color: #00e5a0 !important; letter-spacing: 0.08em; }
.gr-button {
background: #00e5a0 !important;
color: #000 !important;
font-weight: 700 !important;
border-radius: 2px !important;
border: none !important;
font-family: 'IBM Plex Mono', monospace !important;
letter-spacing: 0.05em;
}
.gr-button:hover { background: #00ffa8 !important; }
.gr-box, .gr-panel { background: #151515 !important; border: 1px solid #2a2a2a !important; }
table { width: 100%; border-collapse: collapse; font-size: 0.82em; }
th { color: #00e5a0; border-bottom: 1px solid #2a2a2a; padding: 4px 8px; text-align: left; }
td { padding: 4px 8px; border-bottom: 1px solid #1a1a1a; }
td code { background: #1e1e1e; padding: 2px 6px; border-radius: 2px; color: #a8ff78; }
"""
with gr.Blocks(css=css, title="DACVAE Inspector") as demo:
gr.HTML("""
<link href="https://fonts.googleapis.com/css2?family=IBM+Plex+Mono:wght@400;700&display=swap" rel="stylesheet">
<div style="padding: 24px 0 8px 0;">
<h1 style="font-size:1.6em; margin:0; letter-spacing:0.12em;">
β—ˆ DACVAE CODEC INSPECTOR
</h1>
<p style="color:#666; margin:4px 0 0 0; font-size:0.78em; letter-spacing:0.06em;">
Aratako/Semantic-DACVAE-Japanese-32dim &nbsp;Β·&nbsp;
sr={sr} Hz &nbsp;Β·&nbsp; hop={hop} &nbsp;Β·&nbsp; frame_rate={fr:.4f} Hz
</p>
</div>
""".format(sr=codec.sample_rate, hop=codec.hop_size, fr=codec.frame_rate))
latent_state = gr.State()
with gr.Row():
# ── Left column ───────────────────────────────
with gr.Column(scale=1):
audio_in = gr.Audio(type="filepath", label="Input Audio")
with gr.Row():
encode_btn = gr.Button("β–Ά ENCODE", variant="primary")
decode_btn = gr.Button("β—€ DECODE", variant="primary")
audio_out = gr.Audio(label="Reconstructed Audio", interactive=False)
# ── Right column ──────────────────────────────
with gr.Column(scale=1):
stats_out = gr.Markdown(
value="*Stats will appear here after encoding.*",
label="Stats"
)
with gr.Accordion("Raw Latent JSON (first 3 frames)", open=False):
latent_preview = gr.JSON(label="Latent preview")
# ── Wire up ───────────────────────────────────────
def encode_and_preview(file):
latent_list, _, stats_md = encode_audio(file)
if latent_list is None:
return None, None, stats_md
preview = latent_list[:3] if latent_list else []
return latent_list, preview, stats_md
encode_btn.click(
fn=encode_and_preview,
inputs=audio_in,
outputs=[latent_state, latent_preview, stats_out],
)
decode_btn.click(
fn=decode_audio,
inputs=[latent_state, stats_out],
outputs=[audio_out, stats_out],
)
# =============================
# RUN
# =============================
if __name__ == "__main__":
demo.launch(share=True)