rocky250 commited on
Commit
e6c079e
·
verified ·
1 Parent(s): f9cc141

Upload folder using huggingface_hub

Browse files
checkpoints/ablation_preprocessing_no_adaptive_pooling/meta.json CHANGED
@@ -1 +1 @@
1
- {"epochs_done": 1, "history": [{"epoch": 0, "total": 2.344102304180463, "articulatory": 0.3250249839363954, "phoneme": 2.3278510528497205, "semantic": 3.8904200627253607, "encoder_grad_norm": 3.0684217474399467, "phoneme_head_grad_norm": Infinity, "nan_inf_batch_pct": 0.0, "blank_prediction_pct": 76.32753656172171, "mean_decoded_phoneme_len": 16.92624659291326, "mean_target_phoneme_len": 26.89738656405323, "train_per": 0.6279484260158918, "learning_rate": 0.0009935547237482068, "artic_target_positive_rate": 0.12211586766135998, "artic_pred_positive_rate": 0.15161320740977924, "weight_articulatory": 0.05, "weight_phoneme": 1.0, "weight_semantic": 0.0, "conf_warmup_alpha": 0.0, "gate_mean": 0.5111079263786522, "gate_min": 0.5039503574371338, "gate_max": 0.6817492842674255, "reliability_mean": 0.589669914415331, "reliability_min": 9.999999747378752e-05, "reliability_max": 0.9999998807907104, "drift_severity_mean": 0.04467889739187631, "reference_initialized": true, "ctc_infeasible_pct_this_epoch": 0.0, "ctc_infeasible_pct_cumulative": 0.0, "val_per": 0.3177183350012171, "val_heuristic_wer": 0.9369172326464833, "val_heuristic_cer": 0.5225953762724764, "val_lexicon_wer": 0.5457580292390929, "val_lexicon_cer": 0.39275520882884596, "val_wer": 0.5457580292390929, "val_cer": 0.39275520882884596}], "cfg": {"d_model": 256, "n_heads": 8, "graph_dim": 32, "graph_heads": 4, "n_graph_layers": 2, "trunk_layers": 2, "trunk_kind": "gru", "patch_bins": 8, "semantic_embed_dim": 256, "baseline_d_model": 256, "graph_patch_chunk_size": 16, "dropout": 0.1, "sinkhorn_iters": 50, "batch_size": 8, "accumulation_steps": 1, "epochs": 30, "ablation_epochs": 10, "lr": 0.001, "weight_decay": 0.01, "warmup_frac": 0.05, "min_lr_frac": 0.05, "aug_noise_std": 0.3, "aug_offset_std": 0.1, "aug_channel_drop": 0.08, "aug_time_mask_n": 2, "aug_time_mask_max": 10, "ema_decay": 0.999, "use_session_affine": true, "collapse_patience_epochs": 4, "log_every": 100, "cl_lr": 0.0001, "cl_base_epochs": 3, "cl_task_epochs": 3, "cl_adapt_epochs": 8, "cl_lora_rank": 8, "cl_ot_shrink": 0.05, "cl_replay_ratio": 0.5, "cl_ewc_lambda": 100.0}, "pipeline_version": "chgft_v6_strict_cl_ema_affine_ot"}
 
1
+ {"epochs_done": 2, "history": [{"epoch": 0, "total": 2.344102304180463, "articulatory": 0.3250249839363954, "phoneme": 2.3278510528497205, "semantic": 3.8904200627253607, "encoder_grad_norm": 3.0684217474399467, "phoneme_head_grad_norm": Infinity, "nan_inf_batch_pct": 0.0, "blank_prediction_pct": 76.32753656172171, "mean_decoded_phoneme_len": 16.92624659291326, "mean_target_phoneme_len": 26.89738656405323, "train_per": 0.6279484260158918, "learning_rate": 0.0009935547237482068, "artic_target_positive_rate": 0.12211586766135998, "artic_pred_positive_rate": 0.15161320740977924, "weight_articulatory": 0.05, "weight_phoneme": 1.0, "weight_semantic": 0.0, "conf_warmup_alpha": 0.0, "gate_mean": 0.5111079263786522, "gate_min": 0.5039503574371338, "gate_max": 0.6817492842674255, "reliability_mean": 0.589669914415331, "reliability_min": 9.999999747378752e-05, "reliability_max": 0.9999998807907104, "drift_severity_mean": 0.04467889739187631, "reference_initialized": true, "ctc_infeasible_pct_this_epoch": 0.0, "ctc_infeasible_pct_cumulative": 0.0, "val_per": 0.3177183350012171, "val_heuristic_wer": 0.9369172326464833, "val_heuristic_cer": 0.5225953762724764, "val_lexicon_wer": 0.5457580292390929, "val_lexicon_cer": 0.39275520882884596, "val_wer": 0.5457580292390929, "val_cer": 0.39275520882884596}, {"epoch": 1, "total": 1.625613810160221, "articulatory": 0.29659276337195667, "phoneme": 1.2444457140488503, "semantic": 3.5150881027564025, "encoder_grad_norm": 2.4900145321320264, "phoneme_head_grad_norm": 1.7476189685173524, "nan_inf_batch_pct": 0.0, "blank_prediction_pct": 66.73945641041892, "mean_decoded_phoneme_len": 25.116562449895785, "mean_target_phoneme_len": 26.89738656405323, "train_per": 0.3715448947597446, "learning_rate": 0.000942845842533964, "artic_target_positive_rate": 0.12211523922399069, "artic_pred_positive_rate": 0.1228452838670749, "weight_articulatory": 0.1, "weight_phoneme": 1.0, "weight_semantic": 0.09999999999999999, "conf_warmup_alpha": 0.3333333333333333, "gate_mean": 0.5116789211385121, "gate_min": 0.5043685436248779, "gate_max": 0.6823399066925049, "reliability_mean": 0.583455979451358, "reliability_min": 9.999999747378752e-05, "reliability_max": 0.9999997019767761, "drift_severity_mean": 0.04577792228864175, "reference_initialized": true, "ctc_infeasible_pct_this_epoch": 0.0, "ctc_infeasible_pct_cumulative": 0.0, "val_per": 0.23925567305871853, "val_heuristic_wer": 0.9232186025094969, "val_heuristic_cer": 0.4796641613547712, "val_lexicon_wer": 0.41648440197997005, "val_lexicon_cer": 0.301898011606888, "val_wer": 0.41648440197997005, "val_cer": 0.301898011606888}], "cfg": {"d_model": 256, "n_heads": 8, "graph_dim": 32, "graph_heads": 4, "n_graph_layers": 2, "trunk_layers": 2, "trunk_kind": "gru", "patch_bins": 8, "semantic_embed_dim": 256, "baseline_d_model": 256, "graph_patch_chunk_size": 16, "dropout": 0.1, "sinkhorn_iters": 50, "batch_size": 8, "accumulation_steps": 1, "epochs": 30, "ablation_epochs": 10, "lr": 0.001, "weight_decay": 0.01, "warmup_frac": 0.05, "min_lr_frac": 0.05, "aug_noise_std": 0.3, "aug_offset_std": 0.1, "aug_channel_drop": 0.08, "aug_time_mask_n": 2, "aug_time_mask_max": 10, "ema_decay": 0.999, "use_session_affine": true, "collapse_patience_epochs": 4, "log_every": 100, "cl_lr": 0.0001, "cl_base_epochs": 3, "cl_task_epochs": 3, "cl_adapt_epochs": 8, "cl_lora_rank": 8, "cl_ot_shrink": 0.05, "cl_replay_ratio": 0.5, "cl_ewc_lambda": 100.0}, "pipeline_version": "chgft_v6_strict_cl_ema_affine_ot"}
checkpoints/ablation_preprocessing_no_adaptive_pooling/model.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:72cce5ea13747d18c74ef98c44470beab86023f32128dda7762d7b72d2c37b4b
3
  size 38563175
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bf67556344cc79bd397bfbe7ac61afb7a14c388fb269e31c62c6539a4b3d6e53
3
  size 38563175
checkpoints/ablation_preprocessing_no_adaptive_pooling/optimizer.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:570ab25bc2eeadc214506e43449de72a2e9d1de767513ac0d60875af02c75492
3
  size 38062603
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:432e08baa8320e3f1d1ccb4ee38d429a9ba74290700e56c38d76aa8f2591257a
3
  size 38062603
checkpoints/ablation_preprocessing_no_adaptive_pooling/scaler.pt CHANGED
Binary files a/checkpoints/ablation_preprocessing_no_adaptive_pooling/scaler.pt and b/checkpoints/ablation_preprocessing_no_adaptive_pooling/scaler.pt differ