"""eval_2x2.py — canonical eval harness: stock vs SOAR head on identical 9-dialogue scripts. All fixes baked in (v10): - demo-exact generate pattern (return_tensors="pt", voice paths, tokenizer pass-through) - sf.read tuple unpacks everywhere - audio written via proc.save_audio (fork demo's own path), read back for metrics Env: ARMS (comma list, default "stock,soar"). Pushes wavs + results to ACloudCenter/vibevoice-eval-2x2 under {arm}/. Prints EVAL DONE at the end. """ import json, os, sys, time sys.path.insert(0, "/VibeVoice-classic") import numpy as np import soundfile as sf import torch import jiwer import statistics as st from huggingface_hub import hf_hub_download, HfApi from resemblyzer import VoiceEncoder, preprocess_wav from vibevoice.modular.modeling_vibevoice_inference import VibeVoiceForConditionalGenerationInference from vibevoice.processor.vibevoice_processor import VibeVoiceProcessor MODEL = "microsoft/VibeVoice-1.5B" SOAR_REPO = "ACloudCenter/vibevoice-soar-head-1.5B" OUT_REPO = "ACloudCenter/vibevoice-eval-2x2" SR = 24000 device = "cuda" ARMS = [a.strip() for a in os.environ.get("ARMS", "stock,soar").split(",")] SHORTS = [ "Speaker 1: Did you see the meteor shower last night?\nSpeaker 2: I missed it! Was it worth staying up for?\nSpeaker 1: Absolutely. The sky lit up green for a full minute.\nSpeaker 2: Okay, now I'm actually upset I slept through it.", "Speaker 1: The bread machine finally arrived.\nSpeaker 2: So you're a baker now?\nSpeaker 1: I'm a disaster with a bread machine. The first loaf came out like a brick.\nSpeaker 2: Send me a photo. Immediately.", "Speaker 1: My flight lands at six, so don't wait for dinner.\nSpeaker 2: I'll keep something warm anyway.\nSpeaker 1: You really don't have to.\nSpeaker 2: I know. That's why I'm doing it.", ] MEDIUMS = [ "Speaker 1: I've been reading about how old lighthouse keepers kept journals.\nSpeaker 2: What did they write about?\nSpeaker 1: Mostly weather, ships passing, and how lonely it was.\nSpeaker 2: That sounds peaceful and terrible at the same time.\nSpeaker 1: Exactly. One keeper wrote that the light was his only conversation for months.\nSpeaker 2: Honestly, that might fix my screen time problem.\nSpeaker 1: Or break you completely. There was no off switch on the ocean.", "Speaker 1: The city is redoing the bike lane on Fifth Street.\nSpeaker 2: Finally. That stretch is dangerous.\nSpeaker 1: They're adding a physical barrier between cars and cyclists.\nSpeaker 2: About time. I've almost been clipped twice this month.\nSpeaker 1: Construction starts Monday, so expect detours.\nSpeaker 2: Worth it if it means I stop white-knuckling my handlebars.", "Speaker 1: My grandmother used to make bread every Sunday morning.\nSpeaker 2: Sourdough?\nSpeaker 1: A dense rye. She said the recipe came from her grandmother.\nSpeaker 2: Four generations of the same loaf.\nSpeaker 1: I still have her mixing bowl. It's older than the house.\nSpeaker 2: That bowl has earned a museum display.", ] LONGS = [ "Speaker 1: Let's talk about sleep. Everyone says eight hours, but nobody agrees on anything else.\nSpeaker 2: Right, and half the advice contradicts the other half.\nSpeaker 1: Some people swear by cold rooms. Others say the temperature barely matters.\nSpeaker 2: The one thing that's consistent is the schedule. Same bedtime, same wake time.\nSpeaker 1: Even on weekends, which is where everyone cheats.\nSpeaker 2: That's exactly where I cheat.\nSpeaker 1: The researchers say your body doesn't know it's Saturday.\nSpeaker 2: My body absolutely knows it's Saturday.\nSpeaker 1: Apparently not, biologically speaking.\nSpeaker 2: Fine, I'll set an alarm for Sunday too. This conversation is costing me my weekend.\nSpeaker 1: Sleep hygiene comes with sacrifices.\nSpeaker 2: Don't lecture me at nine on a Sunday morning.", "Speaker 1: The bakery on the corner closed last week.\nSpeaker 2: No! That place had the best almond croissants in the city.\nSpeaker 1: Rent tripled. The owner said she couldn't make the numbers work anymore.\nSpeaker 2: Where does a neighborhood bakery even go from there?\nSpeaker 1: She's moving production to a shared kitchen across the river and doing delivery only.\nSpeaker 2: So no more counter, no more coffee, just a website?\nSpeaker 1: Just a website. She seemed okay about it, which surprised me.\nSpeaker 2: Thirty years behind that counter and she seemed okay?\nSpeaker 1: Maybe she's more practical than sentimental.\nSpeaker 2: I'm going to miss walking past that smell every morning.\nSpeaker 1: Delivery still smells. It just doesn't linger.", "Speaker 1: I finally fixed the shed door.\nSpeaker 2: The one that's been hanging crooked since spring?\nSpeaker 1: Since spring. The hinges were stripped, not the frame, which saved me a full rebuild.\nSpeaker 2: How long did that take?\nSpeaker 1: Four hours, two of which were finding screws that actually fit.\nSpeaker 2: The hardware store run is always the real project.\nSpeaker 1: Two trips. One for the wrong screws, one for the right ones.\nSpeaker 2: Classic. Did you paint it?\nSpeaker 1: Not yet. The wood needs to dry out first, apparently.\nSpeaker 2: So the project continues.\nSpeaker 1: The project continues. But the door closes now, which is more than March could say.", ] def load_model_with_head(load_soar): model = VibeVoiceForConditionalGenerationInference.from_pretrained( MODEL, torch_dtype=torch.bfloat16, low_cpu_mem_usage=True, attn_implementation="sdpa").to(device).eval() if load_soar: sd_path = hf_hub_download(SOAR_REPO, "soar_head_state.pt") sd = torch.load(sd_path, map_location="cpu", weights_only=True) model.model.prediction_head.load_state_dict(sd, strict=True) print(f"loaded SOAR head: {len(sd)} tensors", flush=True) return model def gt_text(script): return " ".join(line.split(": ", 1)[1].strip() for line in script.strip().splitlines()) def main(): api = HfApi() scripts = ([("short", i, t) for i, t in enumerate(SHORTS)] + [("medium", i, t) for i, t in enumerate(MEDIUMS)] + [("long", i, t) for i, t in enumerate(LONGS)]) os.makedirs("/ref", exist_ok=True) ref_wavs = {} for spk, name in ((1, "en-Carter_man.wav"), (2, "en-Maya_woman.wav")): wav, sr = sf.read(f"/VibeVoice-classic/demo/voices/{name}") assert sr == SR ref_wavs[spk] = wav.astype(np.float32) sf.write(f"/ref/s{spk}.wav", ref_wavs[spk], SR) ref_paths = {spk: f"/ref/s{spk}.wav" for spk in (1, 2)} proc = VibeVoiceProcessor.from_pretrained(MODEL) from transformers import pipeline transcriber = pipeline("automatic-speech-recognition", model="openai/whisper-large-v3", device=device, torch_dtype=torch.float16) encoder = VoiceEncoder(device) ref_wav = np.concatenate([ref_wavs[1], np.zeros(SR // 4, dtype=np.float32), ref_wavs[2]]) emb_ref = encoder.embed_utterance(preprocess_wav(ref_wav, SR)) results = [] os.makedirs("/out", exist_ok=True) for arm in ARMS: print(f"=== ARM {arm} ===", flush=True) model = load_model_with_head(load_soar=(arm == "soar")) model.set_ddpm_inference_steps(num_steps=10) for tier, idx, script in scripts: t0 = time.time() with torch.no_grad(): inputs = proc(text=[script], voice_samples=[[ref_paths[1], ref_paths[2]]], padding=True, return_tensors="pt", return_attention_mask=True) out = model.generate( **inputs, max_new_tokens=None, cfg_scale=1.3, tokenizer=proc.tokenizer, generation_config={"do_sample": False}, ) gen_s = time.time() - t0 fn = f"{arm}_{tier}_{idx}.wav" audio = out.speech_outputs[0] proc.save_audio(audio, output_path=f"/out/{fn}") wav = sf.read(f"/out/{fn}", dtype="float32")[0] wav = np.asarray(wav, dtype=np.float32) audio_s = len(wav) / SR rtf = gen_s / max(audio_s, 1e-6) hyp = transcriber(f"/out/{fn}", generate_kwargs={"language": "english"})["text"].strip() gt = gt_text(script) wer = float(jiwer.wer(gt, hyp)) emb_gen = encoder.embed_utterance(preprocess_wav(f"/out/{fn}")) sim = float(np.dot(emb_gen, emb_ref) / (np.linalg.norm(emb_gen) * np.linalg.norm(emb_ref))) drift = None if tier == "long" and audio_s > 60: e1 = encoder.embed_utterance(preprocess_wav(wav[: 30 * SR], SR)) e2 = encoder.embed_utterance(preprocess_wav(wav[-30 * SR:], SR)) drift = float(np.dot(e1, e2) / (np.linalg.norm(e1) * np.linalg.norm(e2))) results.append({"arm": arm, "tier": tier, "idx": idx, "audio_s": audio_s, "gen_s": gen_s, "rtf": rtf, "wer": wer, "sim_ref": sim, "drift": drift, "hyp": hyp}) d = "n/a" if drift is None else f"{drift:.3f}" print(f"{arm} {tier}{idx}: rtf {rtf:.3f} wer {wer:.3f} sim {sim:.3f} drift {d}", flush=True) trackio.log({"rtf": rtf, "wer": wer, "sim": sim}, step=len(results)) api.upload_file(path_or_fileobj=f"/out/{fn}", path_in_repo=f"{arm}/{fn}", repo_id=OUT_REPO, repo_type="model") torch.cuda.empty_cache() del model torch.cuda.empty_cache() with open("/out/results.json", "w") as f: json.dump(results, f, indent=2) summary = {} for arm in ARMS: rows = [r for r in results if r["arm"] == arm] summary[arm] = {"mean_wer": st.mean(r["wer"] for r in rows), "mean_sim": st.mean(r["sim_ref"] for r in rows), "mean_rtf": st.mean(r["rtf"] for r in rows), "long_drift": st.mean(r["drift"] for r in rows if r["drift"] is not None)} with open("/out/summary.json", "w") as f: json.dump(summary, f, indent=2) print("SUMMARY " + json.dumps(summary), flush=True) api.upload_file(path_or_fileobj="/out/results.json", path_in_repo="results.json", repo_id=OUT_REPO, repo_type="model") api.upload_file(path_or_fileobj="/out/summary.json", path_in_repo="summary.json", repo_id=OUT_REPO, repo_type="model") print("EVAL DONE", flush=True) import trackio trackio.init(project="vibevoice-soar-head", space_id="ACloudCenter/vibevoice-soar-head-trackio") main() trackio.finish()