|
Download implementation-notes.md from SpeechAntiSpoofingBenchmarks/Vocoded-VoxCeleb2-SSL-P2: direct link, hf CLI and curl.
- Browser
- Download file 8.26 kB
-
https://huggingface.co/SpeechAntiSpoofingBenchmarks/Vocoded-VoxCeleb2-SSL-P2/resolve/main/implementation-notes.md
- Command line
-
hf download hf://SpeechAntiSpoofingBenchmarks/Vocoded-VoxCeleb2-SSL-P2/implementation-notes.md
-
curl -L -o implementation-notes.md https://huggingface.co/SpeechAntiSpoofingBenchmarks/Vocoded-VoxCeleb2-SSL-P2/resolve/main/implementation-notes.md
8.26 kB
| # Vocoded-VoxCeleb2-SSL-P1 / -P2 / -P3 β implementation notes | |
| Plan: `speech-spoof-bench/docs/plans/2026-11-01-nine-model-batch-arena-submission.md` (models #5, #6, #10). | |
| The three dirs share `_net.py` (byte-identical), the wrapper/test/sweep/TRT scripts differ only in the | |
| model ID (and `_TWO_SSL = True` for P2). This file is copied into each dir. | |
| ## Source | |
| - Code: github.com/nii-yamagishilab/project-NN-Pytorch-scripts @ `b529705b65588d62db17b56f98636ee734ff725e` | |
| (BSD-3-Clause, `LICENSE.source`), `project/10-asvspoof-vocoded-trn-ssl/model-ID-{P1,P2,P3}/model.py`, | |
| `sandbox/block_nn.py` (DropoutForMC), `sandbox/util_bayesian.py` (score), `core_scripts/nn_manager` | |
| (inference loop: `model.eval()`, `torch.no_grad()`, batch size 1, full utterance). | |
| - Weights: Zenodo record 8337778, `project10-cm-ssl-model-ID-{P1,P2,P3}.tar` β | |
| `model-ID-*/trained-for-paper/01/trained_network.pt` (+ the authors' score logs). Tars listed before | |
| extraction (only the .pt and 8 score logs, no odd paths); extracted to | |
| `benchmarks/_ckpt_dl/zenodo__8337778/<ID>/`. The .pt is a plain `OrderedDict` of tensors: read with | |
| `torch.load(weights_only=True)`, strict-loaded, re-saved **unchanged** (same keys, bit-equal tensors) as | |
| `model.safetensors` by `convert_ckpt.py`. P1/P3: 443 tensors, 317,571,590 elements; P2: 872 tensors, | |
| 634,962,182 elements (two SSLs). Keys: `m_front_end.ssl_model[2].model.upstream.model.*` (s3prl | |
| nesting, kept), `m_back_end.*`, and four unused scalars `input_mean/std, output_mean/std` (= 0/1/0/1). | |
| - What the IDs are (paper Β§3.3): v.vox = XLSR-53 continually pre-trained on vocoded VoxCeleb2. | |
| P1 = single SSL from v.vox; P2 = two SSLs, back-end input = `ssl_model(x) β ssl_model2(x)` with | |
| `ssl_model` = wav2vec2_custom (v.vox) and `ssl_model2` = xlsr_53 (verified in model-ID-P2/model.py); | |
| P3 = single student SSL initialised from xlsr_53 and distilled from the (xlsr, v.vox) difference. All | |
| fine-tuned end-to-end on voc.LA (vocoded ASVspoof 2019 LA train bona fide) with RawBoost; 19LA dev for | |
| early stopping. β `trained_on`: none. | |
| ## Front end (s3prl β fairseq, no hub downloads) | |
| - XLSR-53 architecture config: read from the `cfg.model` of fairseq's `xlsr_53_56k.pt` by HTTP-range | |
| fetching only the first 2 MB (zip entry `data.pkl` comes first; unpickled with stubbed | |
| persistent_load, no tensors). The vocoded SSL `wav2vec_ft2_vox_vocoded.pt` (Zenodo 8336949, s3prl | |
| converted format) was checked the same way (first 4 MB): identical architecture (only the | |
| conv-layer string is spelled differently; quantizer_depth/factor = defaults). Both task cfgs have | |
| `normalize: true`. Not the XLS-R 300M cfg (`xlsr2_300m.pt`), though the encoder shapes coincide. | |
| - s3prl wav2vec2 expert semantics reproduced: `F.layer_norm(wav, wav.shape)` per clip (numpy_wav_normalize | |
| False), `extract_features(mask=False)`, padding mask all-False (equal-length windows), `all_hs[-1]` | |
| = hook on `model.encoder` output = last block after the final encoder LayerNorm = fairseq | |
| `features_only=True` `'x'` (dim 1024, T = 201 frames for 64600 samples). | |
| ## Back end + score (kept as the authors') | |
| - Linear 1024β128 (`m_feat_compress`), 3 Γ [Linear 128β128, LeakyReLU, DropoutForMC(p=0.5)] (no residual | |
| connection despite the class name), mean over frames, Linear 128β2. Dropout is per element of the | |
| [frames Γ 128] activations. | |
| - In the authors' inference only these three DropoutForMC layers are stochastic (`F.dropout(training=True)` | |
| regardless of mode); everything else is in eval mode (SSL dropouts are 0 in the cfg anyway). | |
| - Score = `compute_logit_from_prob(mean over 144 samples of softmax[:, 1])` (source | |
| `compute_llr_eps_ale`, verbatim incl. the 1e-7 floors β finite, |score| β€ ~16.1). Bona fide = class 1. | |
| - Reproducibility: the source draws masks from the global RNG (depends on order/batch). We draw explicit | |
| keep-masks (Γ2) per clip from a CUDA `torch.Generator` seeded with `blake2b(key="1", data=clip window | |
| float32 bytes)` (seed 1 = source default `--seed`). So masks depend only on the fixed seed and the clip β | |
| deterministic and batched == single (tested). Two identical clips get identical masks (harmless). | |
| - The SSL runs once per clip; the 144 samples are expanded only in the 128-d back end (cheap). | |
| - `< 3000 samples β score 0` (source `__inference`): **not applied**. It only guards the full-utterance | |
| path against too-short inputs; with the fixed 64600-sample repeat-tiled window it can never trigger | |
| (the debug full-utterance script `parity_fullutt.py` does apply it). Empty clips β zeros window. | |
| ## Window (maintainer decision 2026-11-01, Q2) | |
| Fixed first 64600 samples, shorter clips repeat-tiled; the authors evaluate the entire utterance at batch 1. | |
| FP32 with cuDNN/matmul TF32 off. | |
| ## Reproduction check vs the authors' released score logs (In-the-Wild) | |
| `parity_fullutt.py <ID> <full|win> 400`: 400 random ITW clips (50 per parquet shard, rng seed 0); uid | |
| mapping `Wild_%05d` β `ITW_N` (verified: the released logs give EER P1 6.14 / P2 4.22 / P3 5.86 % with our | |
| labels, = the paper/plan targets). Released-log EERs with our labels (19LA / 21LA / 21DF / ITW): | |
| P1 2.76 / 17.10 / 6.14 / 6.14; P2 1.43 / 15.03 / 7.37 / 4.22; P3 1.73 / 14.88 / 6.02 / 5.86. | |
| | model | mode | Pearson | Spearman | mean |Ξ| | max |Ξ| | EER ours / theirs (same 400) | | |
| |---|---|---|---|---|---|---| | |
| | P3 | full utterance (authors' protocol) | 0.99989 | 0.99945 | 0.048 | 0.42 | 6.35 / 5.20 | | |
| | P3 | Arena window (4 s) | 0.9719 | 0.9698 | 0.64 | 5.99 | 4.16 / 5.20 | | |
| | P2 | full utterance | 0.99993 | 0.99964 | 0.036 | 0.25 | 3.44 / 3.44 | | |
| | P2 | Arena window | 0.9805 | 0.9715 | 0.59 | 5.33 | 2.50 / 3.44 | | |
| | P1 | full utterance | 0.99992 | 0.99856 | 0.039 | 0.28 | 6.45 / 5.93 | | |
| | P1 | Arena window | 0.9742 | 0.9678 | 0.51 | 9.05 | 8.01 / 5.93 | | |
| Full-utterance agreement is at the MC-dropout noise level (different random masks), i.e. the port is | |
| exact; the 400-clip EER difference is 1β2 near-threshold clips. The 4-s window changes scores materially | |
| (expected). Raw: `parity_itw_{full,win}*.npz` / `.log` in each dir. | |
| ## Batch sweep (sweep.py, 64600-sample random input, GPUs shared with other agents, 2026-11-02) | |
| | model | RTX 4070 Ti SUPER (idx 2) | RTX 5060 Ti (idx 3) | | |
| |---|---|---| | |
| | P1/P3 | 61.5 utt/s @24 (plateau 4β32) | 20.0 @4 (~14.5 at β₯8) | | |
| | P2 | 32.7 @24 | (not swept; β 4) | | |
| ## TensorRT (`trt_vocoded_voxceleb2_ssl_<id>.py`, sidecar `trt_vocoded-voxceleb2-ssl-<id>.json`) | |
| Only the SSL(s) run on TensorRT (one engine per SSL: normalised wav [B, 64600] β [B, 201, 1024]); the | |
| waveform layer_norm, P2's difference, the MC back end with the identical per-clip masks, and the score stay | |
| in PyTorch. FP16 kept iff 64-clip ITW Spearman β₯ 0.9999, else FP32. | |
| - FP16 failed that bar everywhere (Spearman 0.9966β0.9996); a first FP32 build with TRT's default TF32 was | |
| also off (max |Ξ| 0.2β0.75, Spearman 0.9993β0.9997): the logit(mean prob) score amplifies small feature | |
| errors. Final engines: **strict FP32, TF32 cleared** (`TRT_NO_TF32=1 TRT_FORCE_FP32=1 ... export`) β | |
| 64-clip build parity max |Ξ| β€ 4.3e-4, Spearman 1.0 on both cards. | |
| - P2: two engines; MAX_BATCH_CAP 16 (32 β context-creation OOM with other agents on the card). The export | |
| first created a second copy of the engines for the sweep (OOM) β fixed (reuses the parity engines). | |
| - J-SPAW_LA (2397 trials) TRT vs PyTorch (`results_pt` = PyTorch on RTX 4070 Ti SUPER; `cmp_scores.py`): | |
| | model | PT EER | 4070 TRT EER / Spearman / max Ξ | 5060 TRT EER / Spearman / max Ξ | TRT utt/s 4070 / 5060 (best batch) | | |
| |---|---|---|---|---| | |
| | P1 | 19.125 | 19.125 / 1.000000 / 8e-4 β | 19.286 / **0.998762** / 0.063 β β PyTorch on RTX_5060_Ti | 63.4 (16) / β | | |
| | P2 | 7.451 | 7.451 / 1.000000 / 4e-4 β | 7.625 (Ξ0.17) / 0.999640 / 0.075 β | 33.3 (16) / 20.5 (16) | | |
| | P3 | 9.625 | 9.625 / 1.000000 / 2e-4 β | 9.625 / 0.999948 / 0.051 β | 64.1 (32) / 38.0 (32) | | |
| P1's RTX_5060_Ti key and engine were deleted (results_trt3 kept as evidence). The 5060 deviations are | |
| cross-card (5060 kernels vs the 4070 PyTorch reference); the 64-clip same-card build parity was exact. | |
| PyTorch utt/s: P1/P3 ~58β61 (4070) / 20 (5060, bs 4); P2 32.7 (4070). | |