:tada: new model

config.json

@@ -1,6 +1,6 @@
 {
   "_commit_hash": null,
-  "_name_or_path": "epochs_10/",
+  "_name_or_path": "augmented/",
   "architectures": [
     "VisionEncoderDecoderModel"
   ],
@@ -78,7 +78,7 @@
     "top_p": 1.0,
     "torch_dtype": null,
     "torchscript": false,
-    "transformers_version": "4.30.2",
+    "transformers_version": "4.31.0",
     "typical_p": 1.0,
     "use_bfloat16": false,
     "use_cache": false,
@@ -161,7 +161,7 @@
     "top_p": 1.0,
     "torch_dtype": null,
     "torchscript": false,
-    "transformers_version": "4.30.2",
+    "transformers_version": "4.31.0",
     "typical_p": 1.0,
     "use_bfloat16": false
   },

generation_config.json

@@ -8,6 +8,6 @@
   "no_repeat_ngram_size": 3,
   "num_beams": 4,
   "pad_token_id": 1,
-  "transformers_version": "4.30.2",
+  "transformers_version": "4.31.0",
   "use_cache": false
 }

optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c4093e367e8b29225f56e34454e7ba3592096419b9593738fca1cc2866372c58
+oid sha256:60c6e67668afb8a54d383ee7f38b2e5308a28f7b4816abe9695c2d69b2e9ac72
 size 6036958

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9dc981772b6511fe1414a2ba29687bd62aaa3b2ac8324a14e77a0412109c3f40
+oid sha256:7945ac4e83a3b8ac82b0d341e4b7c8d88768f033877a7a26b55350805234d1b0
 size 2233126973

rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:31f264dff15f601257953d94b9d974863b70da54bcdd7f88d8dd26fa6d88f00c
-size 14511
+oid sha256:aefac5f2fad16add953ca932a5337a8d82f2042e2e10e8d28fad93231ca61b3d
+size 14575

scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:70a0fed77283d69bee0d73b41955100ee1ff922fc17871087fb2cef1e4c37fa4
+oid sha256:dfc93a53013405da66ec2c0d110a75687cfb2b37e70a9b12123d59f66cc9b54c
 size 627

trainer_state.json

@@ -1,276 +1,272 @@
 {
-  "best_metric": 0.5421165227890015,
-  "best_model_checkpoint": "./20230701_models/checkpoint-1099",
-  "epoch": 19.94540491355778,
-  "global_step": 5480,
+  "best_metric": 0.5082262754440308,
+  "best_model_checkpoint": "./checkpoint-1174",
+  "epoch": 18.998296422487222,
+  "global_step": 5576,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 1.0,
-      "eval_cer": 0.10819683151455398,
-      "eval_loss": 0.6883996725082397,
-      "eval_runtime": 259.3566,
-      "eval_samples_per_second": 8.143,
-      "eval_steps_per_second": 0.509,
-      "eval_wer": 0.26681723843242583,
-      "step": 274
+      "eval_cer": 0.10684727658186562,
+      "eval_loss": 0.6773447394371033,
+      "eval_runtime": 207.0582,
+      "eval_samples_per_second": 11.026,
+      "eval_steps_per_second": 0.691,
+      "eval_wer": 0.26575449135102064,
+      "step": 293
     },
     {
-      "epoch": 1.82,
-      "learning_rate": 3.635766423357665e-05,
-      "loss": 0.9397,
+      "epoch": 1.7,
+      "learning_rate": 3.660068259385666e-05,
+      "loss": 0.9117,
       "step": 500
     },
     {
       "epoch": 2.0,
-      "eval_cer": 0.08392901158491496,
-      "eval_loss": 0.5709623694419861,
-      "eval_runtime": 251.7707,
-      "eval_samples_per_second": 8.389,
-      "eval_steps_per_second": 0.524,
-      "eval_wer": 0.2248809669149066,
-      "step": 549
+      "eval_cer": 0.08164138943248532,
+      "eval_loss": 0.536730945110321,
+      "eval_runtime": 206.1216,
+      "eval_samples_per_second": 11.076,
+      "eval_steps_per_second": 0.694,
+      "eval_wer": 0.21280382668669004,
+      "step": 587
     },
     {
       "epoch": 3.0,
-      "eval_cer": 0.07773321083648912,
-      "eval_loss": 0.5516401529312134,
-      "eval_runtime": 253.1328,
-      "eval_samples_per_second": 8.343,
-      "eval_steps_per_second": 0.521,
-      "eval_wer": 0.20919301672567453,
-      "step": 824
+      "eval_cer": 0.07224396607958251,
+      "eval_loss": 0.5175191164016724,
+      "eval_runtime": 202.5864,
+      "eval_samples_per_second": 11.269,
+      "eval_steps_per_second": 0.706,
+      "eval_wer": 0.19884309472161968,
+      "step": 880
     },
     {
-      "epoch": 3.64,
-      "learning_rate": 3.270802919708029e-05,
-      "loss": 0.2569,
+      "epoch": 3.41,
+      "learning_rate": 3.3187713310580205e-05,
+      "loss": 0.2625,
       "step": 1000
     },
     {
       "epoch": 4.0,
-      "eval_cer": 0.07259058417549914,
-      "eval_loss": 0.5421165227890015,
-      "eval_runtime": 253.562,
-      "eval_samples_per_second": 8.329,
-      "eval_steps_per_second": 0.521,
-      "eval_wer": 0.20198998901233062,
-      "step": 1099
+      "eval_cer": 0.07057240704500979,
+      "eval_loss": 0.5082262754440308,
+      "eval_runtime": 199.851,
+      "eval_samples_per_second": 11.424,
+      "eval_steps_per_second": 0.716,
+      "eval_wer": 0.1921130207464264,
+      "step": 1174
     },
     {
       "epoch": 5.0,
-      "eval_cer": 0.07377820601877787,
-      "eval_loss": 0.5494938492774963,
-      "eval_runtime": 254.0497,
-      "eval_samples_per_second": 8.313,
-      "eval_steps_per_second": 0.52,
-      "eval_wer": 0.2014406055426688,
-      "step": 1373
+      "eval_cer": 0.0687275766470972,
+      "eval_loss": 0.5106597542762756,
+      "eval_runtime": 201.3886,
+      "eval_samples_per_second": 11.336,
+      "eval_steps_per_second": 0.71,
+      "eval_wer": 0.1898325824573113,
+      "step": 1467
     },
     {
-      "epoch": 5.46,
-      "learning_rate": 2.9058394160583945e-05,
-      "loss": 0.109,
+      "epoch": 5.11,
+      "learning_rate": 2.9774744027303753e-05,
+      "loss": 0.112,
       "step": 1500
     },
     {
       "epoch": 6.0,
-      "eval_cer": 0.07154861406771686,
-      "eval_loss": 0.5665440559387207,
-      "eval_runtime": 253.9146,
-      "eval_samples_per_second": 8.318,
-      "eval_steps_per_second": 0.52,
-      "eval_wer": 0.1971676230008546,
-      "step": 1648
+      "eval_cer": 0.06810583822570124,
+      "eval_loss": 0.5201366543769836,
+      "eval_runtime": 202.2293,
+      "eval_samples_per_second": 11.289,
+      "eval_steps_per_second": 0.707,
+      "eval_wer": 0.18510484454085321,
+      "step": 1761
     },
     {
-      "epoch": 7.0,
-      "eval_cer": 0.07226566876554552,
-      "eval_loss": 0.5868140459060669,
-      "eval_runtime": 253.2966,
-      "eval_samples_per_second": 8.338,
-      "eval_steps_per_second": 0.521,
-      "eval_wer": 0.1968624099621536,
-      "step": 1923
-    },
-    {
-      "epoch": 7.28,
-      "learning_rate": 2.5408759124087593e-05,
-      "loss": 0.0481,
+      "epoch": 6.81,
+      "learning_rate": 2.6361774744027308e-05,
+      "loss": 0.0483,
       "step": 2000
     },
     {
-      "epoch": 8.0,
-      "eval_cer": 0.07029376834651668,
-      "eval_loss": 0.5920885801315308,
-      "eval_runtime": 252.718,
-      "eval_samples_per_second": 8.357,
-      "eval_steps_per_second": 0.522,
-      "eval_wer": 0.19570260041508974,
-      "step": 2198
+      "epoch": 7.0,
+      "eval_cer": 0.064925799086758,
+      "eval_loss": 0.5361577272415161,
+      "eval_runtime": 201.8712,
+      "eval_samples_per_second": 11.309,
+      "eval_steps_per_second": 0.708,
+      "eval_wer": 0.18154513599199065,
+      "step": 2054
     },
     {
-      "epoch": 9.0,
-      "eval_cer": 0.07013691263136666,
-      "eval_loss": 0.5932603478431702,
-      "eval_runtime": 254.5349,
-      "eval_samples_per_second": 8.297,
-      "eval_steps_per_second": 0.519,
-      "eval_wer": 0.19374923696740326,
-      "step": 2472
+      "epoch": 8.0,
+      "eval_cer": 0.06522137964774952,
+      "eval_loss": 0.5386066436767578,
+      "eval_runtime": 201.6773,
+      "eval_samples_per_second": 11.32,
+      "eval_steps_per_second": 0.709,
+      "eval_wer": 0.18009900439401524,
+      "step": 2348
     },
     {
-      "epoch": 9.1,
-      "learning_rate": 2.1759124087591242e-05,
+      "epoch": 8.52,
+      "learning_rate": 2.2948805460750853e-05,
       "loss": 0.022,
       "step": 2500
     },
+    {
+      "epoch": 9.0,
+      "eval_cer": 0.0632644324853229,
+      "eval_loss": 0.5573871731758118,
+      "eval_runtime": 201.4546,
+      "eval_samples_per_second": 11.333,
+      "eval_steps_per_second": 0.71,
+      "eval_wer": 0.17509316424717727,
+      "step": 2641
+    },
     {
       "epoch": 10.0,
-      "eval_cer": 0.06856835547986645,
-      "eval_loss": 0.6067739725112915,
-      "eval_runtime": 253.5231,
-      "eval_samples_per_second": 8.331,
-      "eval_steps_per_second": 0.521,
-      "eval_wer": 0.1925894274203394,
-      "step": 2747
+      "eval_cer": 0.06333577951728636,
+      "eval_loss": 0.5539766550064087,
+      "eval_runtime": 201.1768,
+      "eval_samples_per_second": 11.348,
+      "eval_steps_per_second": 0.711,
+      "eval_wer": 0.1772623616441404,
+      "step": 2935
     },
     {
-      "epoch": 10.92,
-      "learning_rate": 1.8116788321167883e-05,
-      "loss": 0.0108,
+      "epoch": 10.22,
+      "learning_rate": 1.9535836177474404e-05,
+      "loss": 0.0116,
       "step": 3000
     },
     {
       "epoch": 11.0,
-      "eval_cer": 0.06744795751450916,
-      "eval_loss": 0.6042998433113098,
-      "eval_runtime": 253.2695,
-      "eval_samples_per_second": 8.339,
-      "eval_steps_per_second": 0.521,
-      "eval_wer": 0.1902087657184715,
-      "step": 3022
+      "eval_cer": 0.06017612524461839,
+      "eval_loss": 0.5528882741928101,
+      "eval_runtime": 201.6037,
+      "eval_samples_per_second": 11.324,
+      "eval_steps_per_second": 0.709,
+      "eval_wer": 0.17019856499249125,
+      "step": 3228
     },
     {
-      "epoch": 12.0,
-      "eval_cer": 0.06674210679633406,
-      "eval_loss": 0.6130247116088867,
-      "eval_runtime": 253.5585,
-      "eval_samples_per_second": 8.329,
-      "eval_steps_per_second": 0.521,
-      "eval_wer": 0.18856061530948603,
-      "step": 3297
+      "epoch": 11.93,
+      "learning_rate": 1.6122866894197952e-05,
+      "loss": 0.005,
+      "step": 3500
     },
     {
-      "epoch": 12.74,
-      "learning_rate": 1.4467153284671533e-05,
-      "loss": 0.0052,
-      "step": 3500
+      "epoch": 12.0,
+      "eval_cer": 0.0598601598173516,
+      "eval_loss": 0.5630314946174622,
+      "eval_runtime": 201.2994,
+      "eval_samples_per_second": 11.341,
+      "eval_steps_per_second": 0.71,
+      "eval_wer": 0.16864119250236387,
+      "step": 3522
     },
     {
       "epoch": 13.0,
-      "eval_cer": 0.0659802361798911,
-      "eval_loss": 0.6161753535270691,
-      "eval_runtime": 254.095,
-      "eval_samples_per_second": 8.312,
-      "eval_steps_per_second": 0.519,
-      "eval_wer": 0.1861799536076181,
-      "step": 3571
+      "eval_cer": 0.06034939660795825,
+      "eval_loss": 0.5675850510597229,
+      "eval_runtime": 201.2983,
+      "eval_samples_per_second": 11.341,
+      "eval_steps_per_second": 0.71,
+      "eval_wer": 0.1693086378552756,
+      "step": 3815
     },
     {
-      "epoch": 14.0,
-      "eval_cer": 0.06466937056042306,
-      "eval_loss": 0.6252880692481995,
-      "eval_runtime": 254.0952,
-      "eval_samples_per_second": 8.312,
-      "eval_steps_per_second": 0.519,
-      "eval_wer": 0.18471493102185324,
-      "step": 3846
+      "epoch": 13.63,
+      "learning_rate": 1.2709897610921502e-05,
+      "loss": 0.0035,
+      "step": 4000
     },
     {
-      "epoch": 14.56,
-      "learning_rate": 1.0817518248175182e-05,
-      "loss": 0.0028,
-      "step": 4000
+      "epoch": 14.0,
+      "eval_cer": 0.05873899217221135,
+      "eval_loss": 0.5630576610565186,
+      "eval_runtime": 201.6395,
+      "eval_samples_per_second": 11.322,
+      "eval_steps_per_second": 0.709,
+      "eval_wer": 0.1666388564436287,
+      "step": 4109
     },
     {
       "epoch": 15.0,
-      "eval_cer": 0.06428843525220158,
-      "eval_loss": 0.6240524649620056,
-      "eval_runtime": 253.6298,
-      "eval_samples_per_second": 8.327,
-      "eval_steps_per_second": 0.52,
-      "eval_wer": 0.18270052496642655,
-      "step": 4121
+      "eval_cer": 0.058494373776908026,
+      "eval_loss": 0.5673080682754517,
+      "eval_runtime": 203.0728,
+      "eval_samples_per_second": 11.242,
+      "eval_steps_per_second": 0.704,
+      "eval_wer": 0.16613827242894488,
+      "step": 4402
     },
     {
-      "epoch": 16.0,
-      "eval_cer": 0.06460214668250162,
-      "eval_loss": 0.6327393054962158,
-      "eval_runtime": 254.0581,
-      "eval_samples_per_second": 8.313,
-      "eval_steps_per_second": 0.52,
-      "eval_wer": 0.18428763276767182,
-      "step": 4396
+      "epoch": 15.33,
+      "learning_rate": 9.296928327645052e-06,
+      "loss": 0.0018,
+      "step": 4500
     },
     {
-      "epoch": 16.38,
-      "learning_rate": 7.167883211678832e-06,
-      "loss": 0.0016,
-      "step": 4500
+      "epoch": 16.0,
+      "eval_cer": 0.0591466894977169,
+      "eval_loss": 0.5768471956253052,
+      "eval_runtime": 202.2731,
+      "eval_samples_per_second": 11.287,
+      "eval_steps_per_second": 0.707,
+      "eval_wer": 0.16852995161021192,
+      "step": 4696
     },
     {
       "epoch": 17.0,
-      "eval_cer": 0.0632464651444193,
-      "eval_loss": 0.6320570707321167,
-      "eval_runtime": 253.728,
-      "eval_samples_per_second": 8.324,
-      "eval_steps_per_second": 0.52,
-      "eval_wer": 0.18190697106580392,
-      "step": 4670
+      "eval_cer": 0.05765859425962166,
+      "eval_loss": 0.5726827383041382,
+      "eval_runtime": 202.414,
+      "eval_samples_per_second": 11.279,
+      "eval_steps_per_second": 0.706,
+      "eval_wer": 0.16636075421324878,
+      "step": 4989
     },
     {
-      "epoch": 18.0,
-      "eval_cer": 0.06377305218813722,
-      "eval_loss": 0.6347479224205017,
-      "eval_runtime": 254.3801,
-      "eval_samples_per_second": 8.303,
-      "eval_steps_per_second": 0.519,
-      "eval_wer": 0.18239531192772554,
-      "step": 4945
+      "epoch": 17.04,
+      "learning_rate": 5.883959044368602e-06,
+      "loss": 0.0012,
+      "step": 5000
     },
     {
-      "epoch": 18.2,
-      "learning_rate": 3.5182481751824822e-06,
-      "loss": 0.0011,
-      "step": 5000
+      "epoch": 18.0,
+      "eval_cer": 0.05809686888454012,
+      "eval_loss": 0.5776260495185852,
+      "eval_runtime": 202.9462,
+      "eval_samples_per_second": 11.249,
+      "eval_steps_per_second": 0.705,
+      "eval_wer": 0.16647199510540076,
+      "step": 5283
     },
     {
-      "epoch": 19.0,
-      "eval_cer": 0.06367221637125507,
-      "eval_loss": 0.6383982300758362,
-      "eval_runtime": 254.2837,
-      "eval_samples_per_second": 8.306,
-      "eval_steps_per_second": 0.519,
-      "eval_wer": 0.1815407154193627,
-      "step": 5220
+      "epoch": 18.74,
+      "learning_rate": 2.4709897610921506e-06,
+      "loss": 0.0008,
+      "step": 5500
     },
     {
-      "epoch": 19.95,
-      "eval_cer": 0.06354897259506577,
-      "eval_loss": 0.6386234164237976,
-      "eval_runtime": 253.9317,
-      "eval_samples_per_second": 8.317,
-      "eval_steps_per_second": 0.52,
-      "eval_wer": 0.18233426931998534,
-      "step": 5480
+      "epoch": 19.0,
+      "eval_cer": 0.05747513046314416,
+      "eval_loss": 0.5768410563468933,
+      "eval_runtime": 201.776,
+      "eval_samples_per_second": 11.315,
+      "eval_steps_per_second": 0.709,
+      "eval_wer": 0.1651371043995773,
+      "step": 5576
     }
   ],
-  "max_steps": 5480,
+  "max_steps": 5860,
   "num_train_epochs": 20,
-  "total_flos": 9.42327114459343e+20,
+  "total_flos": 9.587705362442952e+20,
   "trial_name": null,
   "trial_params": null
 }

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:71d19633df0be7b388fccb41f0bcfa3f9e1a5625377ced62c9270cfe649830c4
+oid sha256:c4382fa122d3ef21fbc678c1a875e8243bd6b530ee29fd4ca4755e18e5bcfc9c
 size 4027