Training in progress, step 500, checkpoint

checkpoint-500/config.json

@@ -84,7 +84,7 @@
   "speech_decoder_prenet_layers": 2,
   "speech_decoder_prenet_units": 256,
   "tie_word_embeddings": true,
-  "transformers_version": "5.19.0.dev0",
+  "transformers_version": "5.19.0",
   "use_cache": false,
   "use_guided_attention_loss": false,
   "vocab_size": 81

checkpoint-500/generation_config.json

@@ -6,6 +6,6 @@
   "output_attentions": false,
   "output_hidden_states": false,
   "pad_token_id": 1,
-  "transformers_version": "5.19.0.dev0",
+  "transformers_version": "5.19.0",
   "use_cache": true
 }

checkpoint-500/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d6a4bc723ca1560db78e14eb822970b4fb6b0b19d8d8924470d35bcb34b28cb6
+oid sha256:95441933ea86b91163c07c3fb1a6a720f6a695ac8df596b542ad1f220d91ba5f
 size 577789320

checkpoint-500/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3f454d8572e9f95337459abf88e57c6b6890ade0792e91775b12fb77f67a9673
+oid sha256:6fc8527ef8b7c491e74272e19413e7fd3fa23f5723bc4a1986494a568c0c6a41
 size 1155772634

checkpoint-500/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f4d5c8ff66cdfd5cd04d9e4d3e28cf0c7fa2ba0ca68ae85fe4a53f5a2cddd7b6
+oid sha256:89f5da95b0303a02fa2a705db6e7bc7e9855f2a5c6b1511bc20c1a5758bfb737
 size 14645

checkpoint-500/trainer_state.json

@@ -1,8 +1,8 @@
 {
   "best_global_step": 500,
-  "best_metric": 1.2403377294540405,
+  "best_metric": 1.2534884214401245,
   "best_model_checkpoint": "/kaggle/working/tts_processed/checkpoint-500",
-  "epoch": 2.4755417956656345,
+  "epoch": 2.4752475247524752,
   "eval_steps": 500,
   "global_step": 500,
   "is_hyper_param_search": false,
@@ -10,151 +10,151 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.1238390092879257,
-      "grad_norm": 829.4996337890625,
+      "epoch": 0.12376237623762376,
+      "grad_norm": 363.43377685546875,
       "learning_rate": 4.8e-08,
-      "loss": 21.73699951171875,
+      "loss": 10.89256103515625,
       "step": 25
     },
     {
-      "epoch": 0.2476780185758514,
-      "grad_norm": 701.2330932617188,
+      "epoch": 0.24752475247524752,
+      "grad_norm": 336.6938171386719,
       "learning_rate": 9.8e-08,
-      "loss": 21.00220947265625,
+      "loss": 10.47541259765625,
       "step": 50
     },
     {
-      "epoch": 0.3715170278637771,
-      "grad_norm": 631.0809326171875,
+      "epoch": 0.3712871287128713,
+      "grad_norm": 255.0070343017578,
       "learning_rate": 1.4799999999999998e-07,
-      "loss": 19.772623291015623,
+      "loss": 10.08507568359375,
       "step": 75
     },
     {
-      "epoch": 0.4953560371517028,
-      "grad_norm": 556.35791015625,
+      "epoch": 0.49504950495049505,
+      "grad_norm": 247.54541015625,
       "learning_rate": 1.98e-07,
-      "loss": 18.46118896484375,
+      "loss": 9.282888793945313,
       "step": 100
     },
     {
-      "epoch": 0.6191950464396285,
-      "grad_norm": 536.2819213867188,
+      "epoch": 0.6188118811881188,
+      "grad_norm": 124.208740234375,
       "learning_rate": 2.48e-07,
-      "loss": 17.51901611328125,
+      "loss": 8.717689208984375,
       "step": 125
     },
     {
-      "epoch": 0.7430340557275542,
-      "grad_norm": 459.290283203125,
+      "epoch": 0.7425742574257426,
+      "grad_norm": 194.5498809814453,
       "learning_rate": 2.98e-07,
-      "loss": 16.43676513671875,
+      "loss": 8.212156372070313,
       "step": 150
     },
     {
-      "epoch": 0.8668730650154799,
-      "grad_norm": 230.56405639648438,
+      "epoch": 0.8663366336633663,
+      "grad_norm": 143.2777557373047,
       "learning_rate": 3.4799999999999994e-07,
-      "loss": 15.75008056640625,
+      "loss": 7.8293896484375,
       "step": 175
     },
     {
-      "epoch": 0.9907120743034056,
-      "grad_norm": 233.90635681152344,
+      "epoch": 0.9900990099009901,
+      "grad_norm": 55.50076675415039,
       "learning_rate": 3.98e-07,
-      "loss": 14.56845947265625,
+      "loss": 7.3809716796875,
       "step": 200
     },
     {
-      "epoch": 1.1139318885448917,
-      "grad_norm": 107.61080169677734,
+      "epoch": 1.113861386138614,
+      "grad_norm": 69.43047332763672,
       "learning_rate": 4.48e-07,
-      "loss": 13.99685546875,
+      "loss": 6.937587890625,
       "step": 225
     },
     {
-      "epoch": 1.2377708978328172,
-      "grad_norm": 115.48291015625,
+      "epoch": 1.2376237623762376,
+      "grad_norm": 49.69321060180664,
       "learning_rate": 4.979999999999999e-07,
-      "loss": 13.390751953125,
+      "loss": 6.8110888671875,
       "step": 250
     },
     {
-      "epoch": 1.361609907120743,
-      "grad_norm": 150.88845825195312,
+      "epoch": 1.3613861386138613,
+      "grad_norm": 26.72427749633789,
       "learning_rate": 5.48e-07,
-      "loss": 12.885958251953125,
+      "loss": 6.522671508789062,
       "step": 275
     },
     {
-      "epoch": 1.4854489164086688,
-      "grad_norm": 102.25688934326172,
+      "epoch": 1.4851485148514851,
+      "grad_norm": 51.36345672607422,
       "learning_rate": 5.979999999999999e-07,
-      "loss": 12.281358642578125,
+      "loss": 6.398966064453125,
       "step": 300
     },
     {
-      "epoch": 1.6092879256965944,
-      "grad_norm": 91.96704864501953,
+      "epoch": 1.608910891089109,
+      "grad_norm": 38.85439682006836,
       "learning_rate": 6.48e-07,
-      "loss": 12.51485107421875,
+      "loss": 6.198662109375,
       "step": 325
     },
     {
-      "epoch": 1.73312693498452,
-      "grad_norm": 89.19316864013672,
+      "epoch": 1.7326732673267327,
+      "grad_norm": 27.508665084838867,
       "learning_rate": 6.979999999999999e-07,
-      "loss": 11.758416748046875,
+      "loss": 6.075740356445312,
       "step": 350
     },
     {
-      "epoch": 1.8569659442724458,
-      "grad_norm": 43.06995391845703,
+      "epoch": 1.8564356435643563,
+      "grad_norm": 36.8642463684082,
       "learning_rate": 7.48e-07,
-      "loss": 11.7668603515625,
+      "loss": 6.0834722900390625,
       "step": 375
     },
     {
-      "epoch": 1.9808049535603716,
-      "grad_norm": 37.787078857421875,
+      "epoch": 1.9801980198019802,
+      "grad_norm": 27.07192039489746,
       "learning_rate": 7.98e-07,
-      "loss": 11.696016845703125,
+      "loss": 5.9166015625,
       "step": 400
     },
     {
-      "epoch": 2.1040247678018575,
-      "grad_norm": 52.63007354736328,
+      "epoch": 2.103960396039604,
+      "grad_norm": 22.465463638305664,
       "learning_rate": 8.48e-07,
-      "loss": 11.60255859375,
+      "loss": 5.795056762695313,
       "step": 425
     },
     {
-      "epoch": 2.2278637770897833,
-      "grad_norm": 37.73246383666992,
+      "epoch": 2.227722772277228,
+      "grad_norm": 34.86723709106445,
       "learning_rate": 8.98e-07,
-      "loss": 11.121099853515625,
+      "loss": 5.741965942382812,
       "step": 450
     },
     {
-      "epoch": 2.351702786377709,
-      "grad_norm": 50.654396057128906,
+      "epoch": 2.3514851485148514,
+      "grad_norm": 22.37932777404785,
       "learning_rate": 9.479999999999999e-07,
-      "loss": 10.981490478515624,
+      "loss": 5.716455688476563,
       "step": 475
     },
     {
-      "epoch": 2.4755417956656345,
-      "grad_norm": 34.40877151489258,
+      "epoch": 2.4752475247524752,
+      "grad_norm": 29.54213523864746,
       "learning_rate": 9.98e-07,
-      "loss": 10.89916015625,
+      "loss": 5.599214477539062,
       "step": 500
     },
     {
-      "epoch": 2.4755417956656345,
-      "eval_loss": 1.2403377294540405,
-      "eval_runtime": 78.2214,
-      "eval_samples_per_second": 20.659,
-      "eval_steps_per_second": 5.165,
+      "epoch": 2.4752475247524752,
+      "eval_loss": 1.2534884214401245,
+      "eval_runtime": 82.4141,
+      "eval_samples_per_second": 19.608,
+      "eval_steps_per_second": 4.902,
       "step": 500
     }
   ],
@@ -184,8 +184,8 @@
       "attributes": {}
     }
   },
-  "total_flos": 1946050805305728.0,
-  "train_batch_size": 4,
+  "total_flos": 2146045761867744.0,
+  "train_batch_size": 8,
   "trial_name": null,
   "trial_params": null
 }

checkpoint-500/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bc3ae1f3188fc78f14b1e24c35f0f902ff13c8475d4503f269d542bb33a09e53
+oid sha256:5d34645042a62af83117180e7d8c3a51dc32127e71ed6c053e7766610460e0b1
 size 5393