Upload folder using huggingface_hub

checkpoint-14990/config.json

@@ -0,0 +1,66 @@
+{
+  "_name_or_path": "openai/whisper-large-v2",
+  "activation_dropout": 0.0,
+  "activation_function": "gelu",
+  "apply_spec_augment": false,
+  "architectures": [
+    "WhisperForConditionalGeneration"
+  ],
+  "attention_dropout": 0.0,
+  "begin_suppress_tokens": [
+    220,
+    50257
+  ],
+  "bos_token_id": 50257,
+  "classifier_proj_size": 256,
+  "d_model": 1280,
+  "decoder_attention_heads": 20,
+  "decoder_ffn_dim": 5120,
+  "decoder_layerdrop": 0.0,
+  "decoder_layers": 32,
+  "decoder_start_token_id": 50258,
+  "dropout": 0.0,
+  "encoder_attention_heads": 20,
+  "encoder_ffn_dim": 5120,
+  "encoder_layerdrop": 0.0,
+  "encoder_layers": 32,
+  "eos_token_id": 50257,
+  "forced_decoder_ids": [
+    [
+      1,
+      50259
+    ],
+    [
+      2,
+      50359
+    ],
+    [
+      3,
+      50363
+    ]
+  ],
+  "init_std": 0.02,
+  "is_encoder_decoder": true,
+  "language": "English",
+  "mask_feature_length": 10,
+  "mask_feature_min_masks": 0,
+  "mask_feature_prob": 0.0,
+  "mask_time_length": 10,
+  "mask_time_min_masks": 2,
+  "mask_time_prob": 0.05,
+  "max_length": 448,
+  "max_source_positions": 1500,
+  "max_target_positions": 448,
+  "median_filter_width": 7,
+  "model_type": "whisper",
+  "num_hidden_layers": 32,
+  "num_mel_bins": 80,
+  "pad_token_id": 50257,
+  "scale_embedding": false,
+  "suppress_tokens": [],
+  "torch_dtype": "float32",
+  "transformers_version": "4.37.2",
+  "use_cache": true,
+  "use_weighted_layer_sum": false,
+  "vocab_size": 51865
+}

checkpoint-14990/generation_config.json

@@ -0,0 +1,317 @@
+{
+  "alignment_heads": [
+    [
+      10,
+      12
+    ],
+    [
+      13,
+      17
+    ],
+    [
+      16,
+      11
+    ],
+    [
+      16,
+      12
+    ],
+    [
+      16,
+      13
+    ],
+    [
+      17,
+      15
+    ],
+    [
+      17,
+      16
+    ],
+    [
+      18,
+      4
+    ],
+    [
+      18,
+      11
+    ],
+    [
+      18,
+      19
+    ],
+    [
+      19,
+      11
+    ],
+    [
+      21,
+      2
+    ],
+    [
+      21,
+      3
+    ],
+    [
+      22,
+      3
+    ],
+    [
+      22,
+      9
+    ],
+    [
+      22,
+      12
+    ],
+    [
+      23,
+      5
+    ],
+    [
+      23,
+      7
+    ],
+    [
+      23,
+      13
+    ],
+    [
+      25,
+      5
+    ],
+    [
+      26,
+      1
+    ],
+    [
+      26,
+      12
+    ],
+    [
+      27,
+      15
+    ]
+  ],
+  "begin_suppress_tokens": [
+    220,
+    50257
+  ],
+  "bos_token_id": 50257,
+  "decoder_start_token_id": 50258,
+  "eos_token_id": 50257,
+  "forced_decoder_ids": [
+    [
+      1,
+      null
+    ],
+    [
+      2,
+      50359
+    ]
+  ],
+  "is_multilingual": true,
+  "lang_to_id": {
+    "<|af|>": 50327,
+    "<|am|>": 50334,
+    "<|ar|>": 50272,
+    "<|as|>": 50350,
+    "<|az|>": 50304,
+    "<|ba|>": 50355,
+    "<|be|>": 50330,
+    "<|bg|>": 50292,
+    "<|bn|>": 50302,
+    "<|bo|>": 50347,
+    "<|br|>": 50309,
+    "<|bs|>": 50315,
+    "<|ca|>": 50270,
+    "<|cs|>": 50283,
+    "<|cy|>": 50297,
+    "<|da|>": 50285,
+    "<|de|>": 50261,
+    "<|el|>": 50281,
+    "<|en|>": 50259,
+    "<|es|>": 50262,
+    "<|et|>": 50307,
+    "<|eu|>": 50310,
+    "<|fa|>": 50300,
+    "<|fi|>": 50277,
+    "<|fo|>": 50338,
+    "<|fr|>": 50265,
+    "<|gl|>": 50319,
+    "<|gu|>": 50333,
+    "<|haw|>": 50352,
+    "<|ha|>": 50354,
+    "<|he|>": 50279,
+    "<|hi|>": 50276,
+    "<|hr|>": 50291,
+    "<|ht|>": 50339,
+    "<|hu|>": 50286,
+    "<|hy|>": 50312,
+    "<|id|>": 50275,
+    "<|is|>": 50311,
+    "<|it|>": 50274,
+    "<|ja|>": 50266,
+    "<|jw|>": 50356,
+    "<|ka|>": 50329,
+    "<|kk|>": 50316,
+    "<|km|>": 50323,
+    "<|kn|>": 50306,
+    "<|ko|>": 50264,
+    "<|la|>": 50294,
+    "<|lb|>": 50345,
+    "<|ln|>": 50353,
+    "<|lo|>": 50336,
+    "<|lt|>": 50293,
+    "<|lv|>": 50301,
+    "<|mg|>": 50349,
+    "<|mi|>": 50295,
+    "<|mk|>": 50308,
+    "<|ml|>": 50296,
+    "<|mn|>": 50314,
+    "<|mr|>": 50320,
+    "<|ms|>": 50282,
+    "<|mt|>": 50343,
+    "<|my|>": 50346,
+    "<|ne|>": 50313,
+    "<|nl|>": 50271,
+    "<|nn|>": 50342,
+    "<|no|>": 50288,
+    "<|oc|>": 50328,
+    "<|pa|>": 50321,
+    "<|pl|>": 50269,
+    "<|ps|>": 50340,
+    "<|pt|>": 50267,
+    "<|ro|>": 50284,
+    "<|ru|>": 50263,
+    "<|sa|>": 50344,
+    "<|sd|>": 50332,
+    "<|si|>": 50322,
+    "<|sk|>": 50298,
+    "<|sl|>": 50305,
+    "<|sn|>": 50324,
+    "<|so|>": 50326,
+    "<|sq|>": 50317,
+    "<|sr|>": 50303,
+    "<|su|>": 50357,
+    "<|sv|>": 50273,
+    "<|sw|>": 50318,
+    "<|ta|>": 50287,
+    "<|te|>": 50299,
+    "<|tg|>": 50331,
+    "<|th|>": 50289,
+    "<|tk|>": 50341,
+    "<|tl|>": 50348,
+    "<|tr|>": 50268,
+    "<|tt|>": 50351,
+    "<|uk|>": 50280,
+    "<|ur|>": 50290,
+    "<|uz|>": 50337,
+    "<|vi|>": 50278,
+    "<|yi|>": 50335,
+    "<|yo|>": 50325,
+    "<|zh|>": 50260
+  },
+  "max_initial_timestamp_index": 50,
+  "max_length": 448,
+  "no_timestamps_token_id": 50363,
+  "pad_token_id": 50257,
+  "prev_sot_token_id": 50361,
+  "return_timestamps": false,
+  "suppress_tokens": [
+    1,
+    2,
+    7,
+    8,
+    9,
+    10,
+    14,
+    25,
+    26,
+    27,
+    28,
+    29,
+    31,
+    58,
+    59,
+    60,
+    61,
+    62,
+    63,
+    90,
+    91,
+    92,
+    93,
+    359,
+    503,
+    522,
+    542,
+    873,
+    893,
+    902,
+    918,
+    922,
+    931,
+    1350,
+    1853,
+    1982,
+    2460,
+    2627,
+    3246,
+    3253,
+    3268,
+    3536,
+    3846,
+    3961,
+    4183,
+    4667,
+    6585,
+    6647,
+    7273,
+    9061,
+    9383,
+    10428,
+    10929,
+    11938,
+    12033,
+    12331,
+    12562,
+    13793,
+    14157,
+    14635,
+    15265,
+    15618,
+    16553,
+    16604,
+    18362,
+    18956,
+    20075,
+    21675,
+    22520,
+    26130,
+    26161,
+    26435,
+    28279,
+    29464,
+    31650,
+    32302,
+    32470,
+    36865,
+    42863,
+    47425,
+    49870,
+    50254,
+    50258,
+    50358,
+    50359,
+    50360,
+    50361,
+    50362
+  ],
+  "task": "transcribe",
+  "task_to_id": {
+    "transcribe": 50359,
+    "translate": 50358
+  },
+  "transformers_version": "4.37.2"
+}

checkpoint-14990/model-00001-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:80253fe0c12ffea0eca4106a68a72d32bfddf94bbbaf6e9ec340009353cb89e8
+size 4992706480

checkpoint-14990/model-00002-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:42bbc8e3b48518e7efd194e7645ef7c0729a04f7b3939f437bb31f119f34f928
+size 1180663192

checkpoint-14990/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:94eabb29139c00d1a91dece97d349c8eb882955d60f79cd7f5f3b7dade5609b9
+size 3095074288

checkpoint-14990/preprocessor_config.json

@@ -0,0 +1,14 @@
+{
+  "chunk_length": 30,
+  "feature_extractor_type": "WhisperFeatureExtractor",
+  "feature_size": 80,
+  "hop_length": 160,
+  "n_fft": 400,
+  "n_samples": 480000,
+  "nb_max_frames": 3000,
+  "padding_side": "right",
+  "padding_value": 0.0,
+  "processor_class": "WhisperProcessor",
+  "return_attention_mask": false,
+  "sampling_rate": 16000
+}

checkpoint-14990/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:19c443289acfa60805307d69881b0ec1c387b572434a6fa73e290bf4e81910fa
+size 14244

checkpoint-14990/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6dace0fc99542e955f7bc68ef4f03cbb24f7ceb849f0b8459bd1d4cd52c2d8cf
+size 1064

checkpoint-14990/trainer_state.json

@@ -0,0 +1,1425 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.6666370185893444,
+  "eval_steps": 500,
+  "global_step": 14990,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "learning_rate": 9.986880725784932e-07,
+      "loss": 5.3038,
+      "step": 64
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 9.972649648670284e-07,
+      "loss": 2.7293,
+      "step": 128
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 9.959085653295383e-07,
+      "loss": 1.6865,
+      "step": 192
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 9.944854576180734e-07,
+      "loss": 0.3066,
+      "step": 256
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 9.930623499066086e-07,
+      "loss": 0.2533,
+      "step": 320
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 9.916392421951437e-07,
+      "loss": 0.3489,
+      "step": 384
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 9.902161344836786e-07,
+      "loss": 0.2908,
+      "step": 448
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 9.887930267722138e-07,
+      "loss": 0.3666,
+      "step": 512
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 9.873699190607489e-07,
+      "loss": 0.3543,
+      "step": 576
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 9.85946811349284e-07,
+      "loss": 0.2938,
+      "step": 640
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 9.84523703637819e-07,
+      "loss": 0.2674,
+      "step": 704
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 9.83100595926354e-07,
+      "loss": 0.2186,
+      "step": 768
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 9.816774882148892e-07,
+      "loss": 0.3472,
+      "step": 832
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 9.802543805034244e-07,
+      "loss": 0.3157,
+      "step": 896
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 9.788312727919595e-07,
+      "loss": 0.432,
+      "step": 960
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 9.774081650804944e-07,
+      "loss": 0.2131,
+      "step": 1024
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 9.759850573690296e-07,
+      "loss": 0.249,
+      "step": 1088
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 9.745619496575647e-07,
+      "loss": 0.239,
+      "step": 1152
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 9.731388419460998e-07,
+      "loss": 0.3577,
+      "step": 1216
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 9.717157342346347e-07,
+      "loss": 0.1945,
+      "step": 1280
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 9.702926265231699e-07,
+      "loss": 0.3107,
+      "step": 1344
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 9.68869518811705e-07,
+      "loss": 0.1983,
+      "step": 1408
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 9.674464111002402e-07,
+      "loss": 0.188,
+      "step": 1472
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 9.660233033887753e-07,
+      "loss": 0.2659,
+      "step": 1536
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 9.646001956773104e-07,
+      "loss": 0.442,
+      "step": 1600
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 9.631770879658454e-07,
+      "loss": 0.1774,
+      "step": 1664
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 9.617539802543805e-07,
+      "loss": 0.2944,
+      "step": 1728
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 9.603308725429156e-07,
+      "loss": 0.3228,
+      "step": 1792
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 9.589077648314505e-07,
+      "loss": 0.2818,
+      "step": 1856
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 9.574846571199857e-07,
+      "loss": 0.2816,
+      "step": 1920
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 9.560615494085208e-07,
+      "loss": 0.3663,
+      "step": 1984
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 9.54638441697056e-07,
+      "loss": 0.2614,
+      "step": 2048
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 9.53215333985591e-07,
+      "loss": 0.2934,
+      "step": 2112
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 9.51792226274126e-07,
+      "loss": 0.4422,
+      "step": 2176
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 9.503691185626613e-07,
+      "loss": 0.2562,
+      "step": 2240
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 9.489460108511963e-07,
+      "loss": 0.3432,
+      "step": 2304
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 9.475229031397314e-07,
+      "loss": 0.3384,
+      "step": 2368
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 9.460997954282665e-07,
+      "loss": 0.1679,
+      "step": 2432
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 9.446766877168015e-07,
+      "loss": 0.1767,
+      "step": 2496
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 9.432535800053366e-07,
+      "loss": 0.2161,
+      "step": 2560
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 9.418304722938717e-07,
+      "loss": 0.2729,
+      "step": 2624
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 9.404073645824068e-07,
+      "loss": 0.3314,
+      "step": 2688
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 9.389842568709418e-07,
+      "loss": 0.242,
+      "step": 2752
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 9.37561149159477e-07,
+      "loss": 0.3446,
+      "step": 2816
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 9.36138041448012e-07,
+      "loss": 0.1292,
+      "step": 2880
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 9.347149337365472e-07,
+      "loss": 0.1854,
+      "step": 2944
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 9.332918260250823e-07,
+      "loss": 0.2536,
+      "step": 3008
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 9.318687183136174e-07,
+      "loss": 0.2224,
+      "step": 3072
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 9.304456106021524e-07,
+      "loss": 0.237,
+      "step": 3136
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 9.290225028906876e-07,
+      "loss": 0.3345,
+      "step": 3200
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 9.275993951792226e-07,
+      "loss": 0.175,
+      "step": 3264
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 9.261762874677576e-07,
+      "loss": 0.2693,
+      "step": 3328
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 9.247531797562928e-07,
+      "loss": 0.254,
+      "step": 3392
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 9.233300720448278e-07,
+      "loss": 0.2855,
+      "step": 3456
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 9.219069643333629e-07,
+      "loss": 0.1507,
+      "step": 3520
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 9.20483856621898e-07,
+      "loss": 0.2464,
+      "step": 3584
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 9.190607489104332e-07,
+      "loss": 0.2522,
+      "step": 3648
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 9.176376411989682e-07,
+      "loss": 0.2378,
+      "step": 3712
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 9.162145334875034e-07,
+      "loss": 0.2433,
+      "step": 3776
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 9.147914257760384e-07,
+      "loss": 0.2932,
+      "step": 3840
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 9.133683180645735e-07,
+      "loss": 0.2602,
+      "step": 3904
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 9.119452103531086e-07,
+      "loss": 0.3365,
+      "step": 3968
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 9.105221026416437e-07,
+      "loss": 0.1738,
+      "step": 4032
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 9.090989949301787e-07,
+      "loss": 0.3132,
+      "step": 4096
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 9.076758872187137e-07,
+      "loss": 0.2353,
+      "step": 4160
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 9.062527795072489e-07,
+      "loss": 0.2868,
+      "step": 4224
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 9.048296717957839e-07,
+      "loss": 0.1807,
+      "step": 4288
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 9.034065640843192e-07,
+      "loss": 0.3052,
+      "step": 4352
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 9.019834563728542e-07,
+      "loss": 0.2848,
+      "step": 4416
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 9.005603486613893e-07,
+      "loss": 0.2031,
+      "step": 4480
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 8.991372409499244e-07,
+      "loss": 0.2588,
+      "step": 4544
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 8.977141332384595e-07,
+      "loss": 0.2356,
+      "step": 4608
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 8.962910255269945e-07,
+      "loss": 0.2445,
+      "step": 4672
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 8.948679178155297e-07,
+      "loss": 0.2537,
+      "step": 4736
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 8.934448101040647e-07,
+      "loss": 0.2345,
+      "step": 4800
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 8.920217023925998e-07,
+      "loss": 0.2098,
+      "step": 4864
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 8.905985946811349e-07,
+      "loss": 0.3487,
+      "step": 4928
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 8.891754869696699e-07,
+      "loss": 0.1884,
+      "step": 4992
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 8.877523792582051e-07,
+      "loss": 0.273,
+      "step": 5056
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 8.863292715467402e-07,
+      "loss": 0.27,
+      "step": 5120
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 8.849061638352753e-07,
+      "loss": 0.1494,
+      "step": 5184
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 8.834830561238103e-07,
+      "loss": 0.1452,
+      "step": 5248
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 8.820599484123455e-07,
+      "loss": 0.177,
+      "step": 5312
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 8.806368407008805e-07,
+      "loss": 0.3317,
+      "step": 5376
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 8.792137329894156e-07,
+      "loss": 0.1638,
+      "step": 5440
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 8.777906252779507e-07,
+      "loss": 0.2928,
+      "step": 5504
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 8.763675175664858e-07,
+      "loss": 0.1851,
+      "step": 5568
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 8.749444098550208e-07,
+      "loss": 0.3108,
+      "step": 5632
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 8.73521302143556e-07,
+      "loss": 0.3002,
+      "step": 5696
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 8.720981944320911e-07,
+      "loss": 0.2726,
+      "step": 5760
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 8.706750867206262e-07,
+      "loss": 0.1775,
+      "step": 5824
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 8.692519790091613e-07,
+      "loss": 0.2632,
+      "step": 5888
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 8.678288712976963e-07,
+      "loss": 0.1625,
+      "step": 5952
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 8.664057635862314e-07,
+      "loss": 0.235,
+      "step": 6016
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 8.649826558747665e-07,
+      "loss": 0.2707,
+      "step": 6080
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 8.635595481633016e-07,
+      "loss": 0.1704,
+      "step": 6144
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 8.621364404518366e-07,
+      "loss": 0.1925,
+      "step": 6208
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 8.607133327403718e-07,
+      "loss": 0.2159,
+      "step": 6272
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 8.592902250289068e-07,
+      "loss": 0.2228,
+      "step": 6336
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 8.578671173174419e-07,
+      "loss": 0.2184,
+      "step": 6400
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 8.56444009605977e-07,
+      "loss": 0.2803,
+      "step": 6464
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 8.550209018945122e-07,
+      "loss": 0.1729,
+      "step": 6528
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 8.535977941830472e-07,
+      "loss": 0.2702,
+      "step": 6592
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 8.521746864715824e-07,
+      "loss": 0.2938,
+      "step": 6656
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 8.507515787601174e-07,
+      "loss": 0.2692,
+      "step": 6720
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 8.493284710486524e-07,
+      "loss": 0.3317,
+      "step": 6784
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 8.479053633371876e-07,
+      "loss": 0.2815,
+      "step": 6848
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 8.464822556257226e-07,
+      "loss": 0.2568,
+      "step": 6912
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 8.450591479142577e-07,
+      "loss": 0.2247,
+      "step": 6976
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 8.436360402027927e-07,
+      "loss": 0.2673,
+      "step": 7040
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 8.422129324913279e-07,
+      "loss": 0.1592,
+      "step": 7104
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 8.407898247798629e-07,
+      "loss": 0.2298,
+      "step": 7168
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 8.393667170683982e-07,
+      "loss": 0.2149,
+      "step": 7232
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 8.379436093569332e-07,
+      "loss": 0.3177,
+      "step": 7296
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 8.365205016454683e-07,
+      "loss": 0.2083,
+      "step": 7360
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 8.350973939340034e-07,
+      "loss": 0.2936,
+      "step": 7424
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 8.336742862225385e-07,
+      "loss": 0.2263,
+      "step": 7488
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 8.322511785110735e-07,
+      "loss": 0.2245,
+      "step": 7552
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 8.308280707996085e-07,
+      "loss": 0.1807,
+      "step": 7616
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 8.294049630881437e-07,
+      "loss": 0.2844,
+      "step": 7680
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 8.279818553766787e-07,
+      "loss": 0.1844,
+      "step": 7744
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 8.265587476652138e-07,
+      "loss": 0.1911,
+      "step": 7808
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 8.251578760117405e-07,
+      "loss": 0.2085,
+      "step": 7872
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 8.237347683002757e-07,
+      "loss": 0.1718,
+      "step": 7936
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 8.223116605888107e-07,
+      "loss": 0.2387,
+      "step": 8000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 8.208885528773458e-07,
+      "loss": 0.1725,
+      "step": 8064
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 8.19465445165881e-07,
+      "loss": 0.1448,
+      "step": 8128
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 8.180423374544161e-07,
+      "loss": 0.154,
+      "step": 8192
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 8.166192297429512e-07,
+      "loss": 0.3209,
+      "step": 8256
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 8.151961220314863e-07,
+      "loss": 0.2067,
+      "step": 8320
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 8.137730143200213e-07,
+      "loss": 0.2147,
+      "step": 8384
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 8.123499066085565e-07,
+      "loss": 0.1771,
+      "step": 8448
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 8.109267988970915e-07,
+      "loss": 0.1212,
+      "step": 8512
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 8.095036911856265e-07,
+      "loss": 0.1944,
+      "step": 8576
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 8.080805834741616e-07,
+      "loss": 0.2416,
+      "step": 8640
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 8.066574757626967e-07,
+      "loss": 0.2004,
+      "step": 8704
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 8.052343680512318e-07,
+      "loss": 0.1667,
+      "step": 8768
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 8.03811260339767e-07,
+      "loss": 0.203,
+      "step": 8832
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 8.023881526283021e-07,
+      "loss": 0.2355,
+      "step": 8896
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 8.009650449168371e-07,
+      "loss": 0.2247,
+      "step": 8960
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 7.995419372053723e-07,
+      "loss": 0.1768,
+      "step": 9024
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 7.981188294939073e-07,
+      "loss": 0.2669,
+      "step": 9088
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 7.966957217824424e-07,
+      "loss": 0.2288,
+      "step": 9152
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 7.952726140709774e-07,
+      "loss": 0.2397,
+      "step": 9216
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 7.938495063595126e-07,
+      "loss": 0.2276,
+      "step": 9280
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 7.924263986480476e-07,
+      "loss": 0.1367,
+      "step": 9344
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 7.910032909365826e-07,
+      "loss": 0.222,
+      "step": 9408
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 7.895801832251178e-07,
+      "loss": 0.2611,
+      "step": 9472
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 7.881570755136529e-07,
+      "loss": 0.2373,
+      "step": 9536
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 7.867339678021881e-07,
+      "loss": 0.2052,
+      "step": 9600
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 7.853108600907231e-07,
+      "loss": 0.1998,
+      "step": 9664
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 7.838877523792582e-07,
+      "loss": 0.2254,
+      "step": 9728
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 7.824646446677932e-07,
+      "loss": 0.1955,
+      "step": 9792
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 7.810415369563284e-07,
+      "loss": 0.1238,
+      "step": 9856
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 7.796184292448634e-07,
+      "loss": 0.1806,
+      "step": 9920
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 7.781953215333986e-07,
+      "loss": 0.31,
+      "step": 9984
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 7.767722138219336e-07,
+      "loss": 0.2349,
+      "step": 10048
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 7.753491061104687e-07,
+      "loss": 0.2277,
+      "step": 10112
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 7.739259983990037e-07,
+      "loss": 0.2239,
+      "step": 10176
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 7.725028906875389e-07,
+      "loss": 0.2707,
+      "step": 10240
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 7.71079782976074e-07,
+      "loss": 0.1957,
+      "step": 10304
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 7.69656675264609e-07,
+      "loss": 0.1649,
+      "step": 10368
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 7.682335675531442e-07,
+      "loss": 0.2773,
+      "step": 10432
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 7.668104598416792e-07,
+      "loss": 0.2146,
+      "step": 10496
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 7.653873521302144e-07,
+      "loss": 0.2134,
+      "step": 10560
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 7.639642444187494e-07,
+      "loss": 0.2168,
+      "step": 10624
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 7.625411367072845e-07,
+      "loss": 0.1122,
+      "step": 10688
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 7.611180289958195e-07,
+      "loss": 0.1628,
+      "step": 10752
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 7.596949212843547e-07,
+      "loss": 0.1566,
+      "step": 10816
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 7.582718135728897e-07,
+      "loss": 0.1824,
+      "step": 10880
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 7.56848705861425e-07,
+      "loss": 0.292,
+      "step": 10944
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 7.5542559814996e-07,
+      "loss": 0.1534,
+      "step": 11008
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 7.54002490438495e-07,
+      "loss": 0.2107,
+      "step": 11072
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 7.525793827270302e-07,
+      "loss": 0.1689,
+      "step": 11136
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 7.511562750155652e-07,
+      "loss": 0.3489,
+      "step": 11200
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 7.497331673041003e-07,
+      "loss": 0.1535,
+      "step": 11264
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 7.483100595926353e-07,
+      "loss": 0.1439,
+      "step": 11328
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 7.468869518811705e-07,
+      "loss": 0.2032,
+      "step": 11392
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 7.454638441697055e-07,
+      "loss": 0.4177,
+      "step": 11456
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 7.440407364582406e-07,
+      "loss": 0.2552,
+      "step": 11520
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 7.426176287467757e-07,
+      "loss": 0.181,
+      "step": 11584
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 7.411945210353108e-07,
+      "loss": 0.1128,
+      "step": 11648
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 7.39771413323846e-07,
+      "loss": 0.2184,
+      "step": 11712
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 7.383483056123811e-07,
+      "loss": 0.3284,
+      "step": 11776
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 7.369251979009161e-07,
+      "loss": 0.1976,
+      "step": 11840
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 7.355020901894511e-07,
+      "loss": 0.1274,
+      "step": 11904
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 7.340789824779863e-07,
+      "loss": 0.1578,
+      "step": 11968
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 7.326558747665213e-07,
+      "loss": 0.2599,
+      "step": 12032
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 7.312327670550564e-07,
+      "loss": 0.1812,
+      "step": 12096
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 7.298096593435915e-07,
+      "loss": 0.2296,
+      "step": 12160
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 7.283865516321266e-07,
+      "loss": 0.2716,
+      "step": 12224
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 7.269634439206616e-07,
+      "loss": 0.3587,
+      "step": 12288
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 7.255403362091968e-07,
+      "loss": 0.1644,
+      "step": 12352
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 7.241394645557235e-07,
+      "loss": 0.2707,
+      "step": 12416
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 7.227163568442586e-07,
+      "loss": 0.1747,
+      "step": 12480
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 7.212932491327936e-07,
+      "loss": 0.2109,
+      "step": 12544
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 7.198701414213289e-07,
+      "loss": 0.2764,
+      "step": 12608
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 7.184470337098639e-07,
+      "loss": 0.1857,
+      "step": 12672
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 7.170239259983991e-07,
+      "loss": 0.2993,
+      "step": 12736
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 7.156008182869341e-07,
+      "loss": 0.3019,
+      "step": 12800
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 7.141777105754691e-07,
+      "loss": 0.2171,
+      "step": 12864
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 7.127546028640042e-07,
+      "loss": 0.1616,
+      "step": 12928
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.113314951525393e-07,
+      "loss": 0.2056,
+      "step": 12992
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.099083874410744e-07,
+      "loss": 0.1741,
+      "step": 13056
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.084852797296094e-07,
+      "loss": 0.1881,
+      "step": 13120
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.070621720181446e-07,
+      "loss": 0.3325,
+      "step": 13184
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.056390643066796e-07,
+      "loss": 0.1797,
+      "step": 13248
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.042159565952149e-07,
+      "loss": 0.1904,
+      "step": 13312
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.027928488837499e-07,
+      "loss": 0.182,
+      "step": 13376
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.01369741172285e-07,
+      "loss": 0.1893,
+      "step": 13440
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 6.9994663346082e-07,
+      "loss": 0.2757,
+      "step": 13504
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 6.985235257493552e-07,
+      "loss": 0.1684,
+      "step": 13568
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.971004180378902e-07,
+      "loss": 0.1678,
+      "step": 13632
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.956773103264252e-07,
+      "loss": 0.2164,
+      "step": 13696
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.942542026149604e-07,
+      "loss": 0.3885,
+      "step": 13760
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.928310949034954e-07,
+      "loss": 0.153,
+      "step": 13824
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.914079871920305e-07,
+      "loss": 0.211,
+      "step": 13888
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.899848794805656e-07,
+      "loss": 0.2952,
+      "step": 13952
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.885617717691008e-07,
+      "loss": 0.227,
+      "step": 14016
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.871386640576358e-07,
+      "loss": 0.1933,
+      "step": 14080
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.85715556346171e-07,
+      "loss": 0.2457,
+      "step": 14144
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.84292448634706e-07,
+      "loss": 0.1388,
+      "step": 14208
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.828693409232412e-07,
+      "loss": 0.175,
+      "step": 14272
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.814462332117762e-07,
+      "loss": 0.1735,
+      "step": 14336
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.800231255003113e-07,
+      "loss": 0.2027,
+      "step": 14400
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.786000177888463e-07,
+      "loss": 0.3326,
+      "step": 14464
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 6.771769100773814e-07,
+      "loss": 0.2391,
+      "step": 14528
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 6.757538023659165e-07,
+      "loss": 0.1766,
+      "step": 14592
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 6.743306946544515e-07,
+      "loss": 0.195,
+      "step": 14656
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 6.729075869429868e-07,
+      "loss": 0.2176,
+      "step": 14720
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 6.714844792315218e-07,
+      "loss": 0.2863,
+      "step": 14784
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 6.700836075780485e-07,
+      "loss": 0.2664,
+      "step": 14848
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 6.686604998665835e-07,
+      "loss": 0.1482,
+      "step": 14912
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 6.672373921551188e-07,
+      "loss": 0.3435,
+      "step": 14976
+    }
+  ],
+  "logging_steps": 64,
+  "max_steps": 44972,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 2,
+  "save_steps": 14990,
+  "total_flos": 3.1826333601792e+19,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-14990/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2463dffe438013cfefdc9915eb0241a5214c5766ec68d593ec6bac88b5ea2196
+size 5048

checkpoint-29980/config.json

@@ -0,0 +1,66 @@
+{
+  "_name_or_path": "openai/whisper-large-v2",
+  "activation_dropout": 0.0,
+  "activation_function": "gelu",
+  "apply_spec_augment": false,
+  "architectures": [
+    "WhisperForConditionalGeneration"
+  ],
+  "attention_dropout": 0.0,
+  "begin_suppress_tokens": [
+    220,
+    50257
+  ],
+  "bos_token_id": 50257,
+  "classifier_proj_size": 256,
+  "d_model": 1280,
+  "decoder_attention_heads": 20,
+  "decoder_ffn_dim": 5120,
+  "decoder_layerdrop": 0.0,
+  "decoder_layers": 32,
+  "decoder_start_token_id": 50258,
+  "dropout": 0.0,
+  "encoder_attention_heads": 20,
+  "encoder_ffn_dim": 5120,
+  "encoder_layerdrop": 0.0,
+  "encoder_layers": 32,
+  "eos_token_id": 50257,
+  "forced_decoder_ids": [
+    [
+      1,
+      50259
+    ],
+    [
+      2,
+      50359
+    ],
+    [
+      3,
+      50363
+    ]
+  ],
+  "init_std": 0.02,
+  "is_encoder_decoder": true,
+  "language": "English",
+  "mask_feature_length": 10,
+  "mask_feature_min_masks": 0,
+  "mask_feature_prob": 0.0,
+  "mask_time_length": 10,
+  "mask_time_min_masks": 2,
+  "mask_time_prob": 0.05,
+  "max_length": 448,
+  "max_source_positions": 1500,
+  "max_target_positions": 448,
+  "median_filter_width": 7,
+  "model_type": "whisper",
+  "num_hidden_layers": 32,
+  "num_mel_bins": 80,
+  "pad_token_id": 50257,
+  "scale_embedding": false,
+  "suppress_tokens": [],
+  "torch_dtype": "float32",
+  "transformers_version": "4.37.2",
+  "use_cache": true,
+  "use_weighted_layer_sum": false,
+  "vocab_size": 51865
+}

checkpoint-29980/generation_config.json

@@ -0,0 +1,317 @@
+{
+  "alignment_heads": [
+    [
+      10,
+      12
+    ],
+    [
+      13,
+      17
+    ],
+    [
+      16,
+      11
+    ],
+    [
+      16,
+      12
+    ],
+    [
+      16,
+      13
+    ],
+    [
+      17,
+      15
+    ],
+    [
+      17,
+      16
+    ],
+    [
+      18,
+      4
+    ],
+    [
+      18,
+      11
+    ],
+    [
+      18,
+      19
+    ],
+    [
+      19,
+      11
+    ],
+    [
+      21,
+      2
+    ],
+    [
+      21,
+      3
+    ],
+    [
+      22,
+      3
+    ],
+    [
+      22,
+      9
+    ],
+    [
+      22,
+      12
+    ],
+    [
+      23,
+      5
+    ],
+    [
+      23,
+      7
+    ],
+    [
+      23,
+      13
+    ],
+    [
+      25,
+      5
+    ],
+    [
+      26,
+      1
+    ],
+    [
+      26,
+      12
+    ],
+    [
+      27,
+      15
+    ]
+  ],
+  "begin_suppress_tokens": [
+    220,
+    50257
+  ],
+  "bos_token_id": 50257,
+  "decoder_start_token_id": 50258,
+  "eos_token_id": 50257,
+  "forced_decoder_ids": [
+    [
+      1,
+      null
+    ],
+    [
+      2,
+      50359
+    ]
+  ],
+  "is_multilingual": true,
+  "lang_to_id": {
+    "<|af|>": 50327,
+    "<|am|>": 50334,
+    "<|ar|>": 50272,
+    "<|as|>": 50350,
+    "<|az|>": 50304,
+    "<|ba|>": 50355,
+    "<|be|>": 50330,
+    "<|bg|>": 50292,
+    "<|bn|>": 50302,
+    "<|bo|>": 50347,
+    "<|br|>": 50309,
+    "<|bs|>": 50315,
+    "<|ca|>": 50270,
+    "<|cs|>": 50283,
+    "<|cy|>": 50297,
+    "<|da|>": 50285,
+    "<|de|>": 50261,
+    "<|el|>": 50281,
+    "<|en|>": 50259,
+    "<|es|>": 50262,
+    "<|et|>": 50307,
+    "<|eu|>": 50310,
+    "<|fa|>": 50300,
+    "<|fi|>": 50277,
+    "<|fo|>": 50338,
+    "<|fr|>": 50265,
+    "<|gl|>": 50319,
+    "<|gu|>": 50333,
+    "<|haw|>": 50352,
+    "<|ha|>": 50354,
+    "<|he|>": 50279,
+    "<|hi|>": 50276,
+    "<|hr|>": 50291,
+    "<|ht|>": 50339,
+    "<|hu|>": 50286,
+    "<|hy|>": 50312,
+    "<|id|>": 50275,
+    "<|is|>": 50311,
+    "<|it|>": 50274,
+    "<|ja|>": 50266,
+    "<|jw|>": 50356,
+    "<|ka|>": 50329,
+    "<|kk|>": 50316,
+    "<|km|>": 50323,
+    "<|kn|>": 50306,
+    "<|ko|>": 50264,
+    "<|la|>": 50294,
+    "<|lb|>": 50345,
+    "<|ln|>": 50353,
+    "<|lo|>": 50336,
+    "<|lt|>": 50293,
+    "<|lv|>": 50301,
+    "<|mg|>": 50349,
+    "<|mi|>": 50295,
+    "<|mk|>": 50308,
+    "<|ml|>": 50296,
+    "<|mn|>": 50314,
+    "<|mr|>": 50320,
+    "<|ms|>": 50282,
+    "<|mt|>": 50343,
+    "<|my|>": 50346,
+    "<|ne|>": 50313,
+    "<|nl|>": 50271,
+    "<|nn|>": 50342,
+    "<|no|>": 50288,
+    "<|oc|>": 50328,
+    "<|pa|>": 50321,
+    "<|pl|>": 50269,
+    "<|ps|>": 50340,
+    "<|pt|>": 50267,
+    "<|ro|>": 50284,
+    "<|ru|>": 50263,
+    "<|sa|>": 50344,
+    "<|sd|>": 50332,
+    "<|si|>": 50322,
+    "<|sk|>": 50298,
+    "<|sl|>": 50305,
+    "<|sn|>": 50324,
+    "<|so|>": 50326,
+    "<|sq|>": 50317,
+    "<|sr|>": 50303,
+    "<|su|>": 50357,
+    "<|sv|>": 50273,
+    "<|sw|>": 50318,
+    "<|ta|>": 50287,
+    "<|te|>": 50299,
+    "<|tg|>": 50331,
+    "<|th|>": 50289,
+    "<|tk|>": 50341,
+    "<|tl|>": 50348,
+    "<|tr|>": 50268,
+    "<|tt|>": 50351,
+    "<|uk|>": 50280,
+    "<|ur|>": 50290,
+    "<|uz|>": 50337,
+    "<|vi|>": 50278,
+    "<|yi|>": 50335,
+    "<|yo|>": 50325,
+    "<|zh|>": 50260
+  },
+  "max_initial_timestamp_index": 50,
+  "max_length": 448,
+  "no_timestamps_token_id": 50363,
+  "pad_token_id": 50257,
+  "prev_sot_token_id": 50361,
+  "return_timestamps": false,
+  "suppress_tokens": [
+    1,
+    2,
+    7,
+    8,
+    9,
+    10,
+    14,
+    25,
+    26,
+    27,
+    28,
+    29,
+    31,
+    58,
+    59,
+    60,
+    61,
+    62,
+    63,
+    90,
+    91,
+    92,
+    93,
+    359,
+    503,
+    522,
+    542,
+    873,
+    893,
+    902,
+    918,
+    922,
+    931,
+    1350,
+    1853,
+    1982,
+    2460,
+    2627,
+    3246,
+    3253,
+    3268,
+    3536,
+    3846,
+    3961,
+    4183,
+    4667,
+    6585,
+    6647,
+    7273,
+    9061,
+    9383,
+    10428,
+    10929,
+    11938,
+    12033,
+    12331,
+    12562,
+    13793,
+    14157,
+    14635,
+    15265,
+    15618,
+    16553,
+    16604,
+    18362,
+    18956,
+    20075,
+    21675,
+    22520,
+    26130,
+    26161,
+    26435,
+    28279,
+    29464,
+    31650,
+    32302,
+    32470,
+    36865,
+    42863,
+    47425,
+    49870,
+    50254,
+    50258,
+    50358,
+    50359,
+    50360,
+    50361,
+    50362
+  ],
+  "task": "transcribe",
+  "task_to_id": {
+    "transcribe": 50359,
+    "translate": 50358
+  },
+  "transformers_version": "4.37.2"
+}

checkpoint-29980/model-00001-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7849c5b638cc6dae02c79f8ed73784c59a8e93a26b0e58c1814adeec177bc650
+size 4992706480

checkpoint-29980/model-00002-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:99d5d3949b132f6d23c2bd71d68343cafbf89ad0bd8228e2c5624b878f64d5ed
+size 1180663192

checkpoint-29980/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6f8cde1a5dcf41c196f8da9ed086500dc5565ae51779f8a3ed3f40cbad37ba28
+size 3095074288

checkpoint-29980/preprocessor_config.json

@@ -0,0 +1,14 @@
+{
+  "chunk_length": 30,
+  "feature_extractor_type": "WhisperFeatureExtractor",
+  "feature_size": 80,
+  "hop_length": 160,
+  "n_fft": 400,
+  "n_samples": 480000,
+  "nb_max_frames": 3000,
+  "padding_side": "right",
+  "padding_value": 0.0,
+  "processor_class": "WhisperProcessor",
+  "return_attention_mask": false,
+  "sampling_rate": 16000
+}

checkpoint-29980/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:be61adb96b72eade93dce1cedcd60b691d773048424e706d0cd5ac22d97dd85e
+size 14244

checkpoint-29980/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4308530455da1f5f90d491b8c3cf4b9e1f06e1a520107f9c49928909f63baf73
+size 1064

checkpoint-29980/trainer_state.json

@@ -0,0 +1,2829 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.3332740371786889,
+  "eval_steps": 500,
+  "global_step": 29980,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "learning_rate": 9.986880725784932e-07,
+      "loss": 5.3038,
+      "step": 64
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 9.972649648670284e-07,
+      "loss": 2.7293,
+      "step": 128
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 9.959085653295383e-07,
+      "loss": 1.6865,
+      "step": 192
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 9.944854576180734e-07,
+      "loss": 0.3066,
+      "step": 256
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 9.930623499066086e-07,
+      "loss": 0.2533,
+      "step": 320
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 9.916392421951437e-07,
+      "loss": 0.3489,
+      "step": 384
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 9.902161344836786e-07,
+      "loss": 0.2908,
+      "step": 448
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 9.887930267722138e-07,
+      "loss": 0.3666,
+      "step": 512
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 9.873699190607489e-07,
+      "loss": 0.3543,
+      "step": 576
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 9.85946811349284e-07,
+      "loss": 0.2938,
+      "step": 640
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 9.84523703637819e-07,
+      "loss": 0.2674,
+      "step": 704
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 9.83100595926354e-07,
+      "loss": 0.2186,
+      "step": 768
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 9.816774882148892e-07,
+      "loss": 0.3472,
+      "step": 832
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 9.802543805034244e-07,
+      "loss": 0.3157,
+      "step": 896
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 9.788312727919595e-07,
+      "loss": 0.432,
+      "step": 960
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 9.774081650804944e-07,
+      "loss": 0.2131,
+      "step": 1024
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 9.759850573690296e-07,
+      "loss": 0.249,
+      "step": 1088
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 9.745619496575647e-07,
+      "loss": 0.239,
+      "step": 1152
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 9.731388419460998e-07,
+      "loss": 0.3577,
+      "step": 1216
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 9.717157342346347e-07,
+      "loss": 0.1945,
+      "step": 1280
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 9.702926265231699e-07,
+      "loss": 0.3107,
+      "step": 1344
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 9.68869518811705e-07,
+      "loss": 0.1983,
+      "step": 1408
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 9.674464111002402e-07,
+      "loss": 0.188,
+      "step": 1472
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 9.660233033887753e-07,
+      "loss": 0.2659,
+      "step": 1536
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 9.646001956773104e-07,
+      "loss": 0.442,
+      "step": 1600
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 9.631770879658454e-07,
+      "loss": 0.1774,
+      "step": 1664
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 9.617539802543805e-07,
+      "loss": 0.2944,
+      "step": 1728
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 9.603308725429156e-07,
+      "loss": 0.3228,
+      "step": 1792
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 9.589077648314505e-07,
+      "loss": 0.2818,
+      "step": 1856
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 9.574846571199857e-07,
+      "loss": 0.2816,
+      "step": 1920
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 9.560615494085208e-07,
+      "loss": 0.3663,
+      "step": 1984
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 9.54638441697056e-07,
+      "loss": 0.2614,
+      "step": 2048
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 9.53215333985591e-07,
+      "loss": 0.2934,
+      "step": 2112
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 9.51792226274126e-07,
+      "loss": 0.4422,
+      "step": 2176
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 9.503691185626613e-07,
+      "loss": 0.2562,
+      "step": 2240
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 9.489460108511963e-07,
+      "loss": 0.3432,
+      "step": 2304
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 9.475229031397314e-07,
+      "loss": 0.3384,
+      "step": 2368
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 9.460997954282665e-07,
+      "loss": 0.1679,
+      "step": 2432
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 9.446766877168015e-07,
+      "loss": 0.1767,
+      "step": 2496
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 9.432535800053366e-07,
+      "loss": 0.2161,
+      "step": 2560
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 9.418304722938717e-07,
+      "loss": 0.2729,
+      "step": 2624
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 9.404073645824068e-07,
+      "loss": 0.3314,
+      "step": 2688
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 9.389842568709418e-07,
+      "loss": 0.242,
+      "step": 2752
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 9.37561149159477e-07,
+      "loss": 0.3446,
+      "step": 2816
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 9.36138041448012e-07,
+      "loss": 0.1292,
+      "step": 2880
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 9.347149337365472e-07,
+      "loss": 0.1854,
+      "step": 2944
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 9.332918260250823e-07,
+      "loss": 0.2536,
+      "step": 3008
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 9.318687183136174e-07,
+      "loss": 0.2224,
+      "step": 3072
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 9.304456106021524e-07,
+      "loss": 0.237,
+      "step": 3136
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 9.290225028906876e-07,
+      "loss": 0.3345,
+      "step": 3200
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 9.275993951792226e-07,
+      "loss": 0.175,
+      "step": 3264
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 9.261762874677576e-07,
+      "loss": 0.2693,
+      "step": 3328
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 9.247531797562928e-07,
+      "loss": 0.254,
+      "step": 3392
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 9.233300720448278e-07,
+      "loss": 0.2855,
+      "step": 3456
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 9.219069643333629e-07,
+      "loss": 0.1507,
+      "step": 3520
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 9.20483856621898e-07,
+      "loss": 0.2464,
+      "step": 3584
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 9.190607489104332e-07,
+      "loss": 0.2522,
+      "step": 3648
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 9.176376411989682e-07,
+      "loss": 0.2378,
+      "step": 3712
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 9.162145334875034e-07,
+      "loss": 0.2433,
+      "step": 3776
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 9.147914257760384e-07,
+      "loss": 0.2932,
+      "step": 3840
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 9.133683180645735e-07,
+      "loss": 0.2602,
+      "step": 3904
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 9.119452103531086e-07,
+      "loss": 0.3365,
+      "step": 3968
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 9.105221026416437e-07,
+      "loss": 0.1738,
+      "step": 4032
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 9.090989949301787e-07,
+      "loss": 0.3132,
+      "step": 4096
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 9.076758872187137e-07,
+      "loss": 0.2353,
+      "step": 4160
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 9.062527795072489e-07,
+      "loss": 0.2868,
+      "step": 4224
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 9.048296717957839e-07,
+      "loss": 0.1807,
+      "step": 4288
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 9.034065640843192e-07,
+      "loss": 0.3052,
+      "step": 4352
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 9.019834563728542e-07,
+      "loss": 0.2848,
+      "step": 4416
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 9.005603486613893e-07,
+      "loss": 0.2031,
+      "step": 4480
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 8.991372409499244e-07,
+      "loss": 0.2588,
+      "step": 4544
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 8.977141332384595e-07,
+      "loss": 0.2356,
+      "step": 4608
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 8.962910255269945e-07,
+      "loss": 0.2445,
+      "step": 4672
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 8.948679178155297e-07,
+      "loss": 0.2537,
+      "step": 4736
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 8.934448101040647e-07,
+      "loss": 0.2345,
+      "step": 4800
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 8.920217023925998e-07,
+      "loss": 0.2098,
+      "step": 4864
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 8.905985946811349e-07,
+      "loss": 0.3487,
+      "step": 4928
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 8.891754869696699e-07,
+      "loss": 0.1884,
+      "step": 4992
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 8.877523792582051e-07,
+      "loss": 0.273,
+      "step": 5056
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 8.863292715467402e-07,
+      "loss": 0.27,
+      "step": 5120
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 8.849061638352753e-07,
+      "loss": 0.1494,
+      "step": 5184
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 8.834830561238103e-07,
+      "loss": 0.1452,
+      "step": 5248
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 8.820599484123455e-07,
+      "loss": 0.177,
+      "step": 5312
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 8.806368407008805e-07,
+      "loss": 0.3317,
+      "step": 5376
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 8.792137329894156e-07,
+      "loss": 0.1638,
+      "step": 5440
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 8.777906252779507e-07,
+      "loss": 0.2928,
+      "step": 5504
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 8.763675175664858e-07,
+      "loss": 0.1851,
+      "step": 5568
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 8.749444098550208e-07,
+      "loss": 0.3108,
+      "step": 5632
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 8.73521302143556e-07,
+      "loss": 0.3002,
+      "step": 5696
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 8.720981944320911e-07,
+      "loss": 0.2726,
+      "step": 5760
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 8.706750867206262e-07,
+      "loss": 0.1775,
+      "step": 5824
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 8.692519790091613e-07,
+      "loss": 0.2632,
+      "step": 5888
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 8.678288712976963e-07,
+      "loss": 0.1625,
+      "step": 5952
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 8.664057635862314e-07,
+      "loss": 0.235,
+      "step": 6016
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 8.649826558747665e-07,
+      "loss": 0.2707,
+      "step": 6080
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 8.635595481633016e-07,
+      "loss": 0.1704,
+      "step": 6144
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 8.621364404518366e-07,
+      "loss": 0.1925,
+      "step": 6208
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 8.607133327403718e-07,
+      "loss": 0.2159,
+      "step": 6272
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 8.592902250289068e-07,
+      "loss": 0.2228,
+      "step": 6336
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 8.578671173174419e-07,
+      "loss": 0.2184,
+      "step": 6400
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 8.56444009605977e-07,
+      "loss": 0.2803,
+      "step": 6464
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 8.550209018945122e-07,
+      "loss": 0.1729,
+      "step": 6528
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 8.535977941830472e-07,
+      "loss": 0.2702,
+      "step": 6592
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 8.521746864715824e-07,
+      "loss": 0.2938,
+      "step": 6656
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 8.507515787601174e-07,
+      "loss": 0.2692,
+      "step": 6720
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 8.493284710486524e-07,
+      "loss": 0.3317,
+      "step": 6784
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 8.479053633371876e-07,
+      "loss": 0.2815,
+      "step": 6848
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 8.464822556257226e-07,
+      "loss": 0.2568,
+      "step": 6912
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 8.450591479142577e-07,
+      "loss": 0.2247,
+      "step": 6976
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 8.436360402027927e-07,
+      "loss": 0.2673,
+      "step": 7040
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 8.422129324913279e-07,
+      "loss": 0.1592,
+      "step": 7104
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 8.407898247798629e-07,
+      "loss": 0.2298,
+      "step": 7168
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 8.393667170683982e-07,
+      "loss": 0.2149,
+      "step": 7232
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 8.379436093569332e-07,
+      "loss": 0.3177,
+      "step": 7296
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 8.365205016454683e-07,
+      "loss": 0.2083,
+      "step": 7360
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 8.350973939340034e-07,
+      "loss": 0.2936,
+      "step": 7424
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 8.336742862225385e-07,
+      "loss": 0.2263,
+      "step": 7488
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 8.322511785110735e-07,
+      "loss": 0.2245,
+      "step": 7552
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 8.308280707996085e-07,
+      "loss": 0.1807,
+      "step": 7616
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 8.294049630881437e-07,
+      "loss": 0.2844,
+      "step": 7680
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 8.279818553766787e-07,
+      "loss": 0.1844,
+      "step": 7744
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 8.265587476652138e-07,
+      "loss": 0.1911,
+      "step": 7808
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 8.251578760117405e-07,
+      "loss": 0.2085,
+      "step": 7872
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 8.237347683002757e-07,
+      "loss": 0.1718,
+      "step": 7936
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 8.223116605888107e-07,
+      "loss": 0.2387,
+      "step": 8000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 8.208885528773458e-07,
+      "loss": 0.1725,
+      "step": 8064
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 8.19465445165881e-07,
+      "loss": 0.1448,
+      "step": 8128
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 8.180423374544161e-07,
+      "loss": 0.154,
+      "step": 8192
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 8.166192297429512e-07,
+      "loss": 0.3209,
+      "step": 8256
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 8.151961220314863e-07,
+      "loss": 0.2067,
+      "step": 8320
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 8.137730143200213e-07,
+      "loss": 0.2147,
+      "step": 8384
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 8.123499066085565e-07,
+      "loss": 0.1771,
+      "step": 8448
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 8.109267988970915e-07,
+      "loss": 0.1212,
+      "step": 8512
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 8.095036911856265e-07,
+      "loss": 0.1944,
+      "step": 8576
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 8.080805834741616e-07,
+      "loss": 0.2416,
+      "step": 8640
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 8.066574757626967e-07,
+      "loss": 0.2004,
+      "step": 8704
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 8.052343680512318e-07,
+      "loss": 0.1667,
+      "step": 8768
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 8.03811260339767e-07,
+      "loss": 0.203,
+      "step": 8832
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 8.023881526283021e-07,
+      "loss": 0.2355,
+      "step": 8896
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 8.009650449168371e-07,
+      "loss": 0.2247,
+      "step": 8960
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 7.995419372053723e-07,
+      "loss": 0.1768,
+      "step": 9024
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 7.981188294939073e-07,
+      "loss": 0.2669,
+      "step": 9088
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 7.966957217824424e-07,
+      "loss": 0.2288,
+      "step": 9152
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 7.952726140709774e-07,
+      "loss": 0.2397,
+      "step": 9216
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 7.938495063595126e-07,
+      "loss": 0.2276,
+      "step": 9280
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 7.924263986480476e-07,
+      "loss": 0.1367,
+      "step": 9344
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 7.910032909365826e-07,
+      "loss": 0.222,
+      "step": 9408
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 7.895801832251178e-07,
+      "loss": 0.2611,
+      "step": 9472
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 7.881570755136529e-07,
+      "loss": 0.2373,
+      "step": 9536
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 7.867339678021881e-07,
+      "loss": 0.2052,
+      "step": 9600
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 7.853108600907231e-07,
+      "loss": 0.1998,
+      "step": 9664
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 7.838877523792582e-07,
+      "loss": 0.2254,
+      "step": 9728
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 7.824646446677932e-07,
+      "loss": 0.1955,
+      "step": 9792
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 7.810415369563284e-07,
+      "loss": 0.1238,
+      "step": 9856
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 7.796184292448634e-07,
+      "loss": 0.1806,
+      "step": 9920
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 7.781953215333986e-07,
+      "loss": 0.31,
+      "step": 9984
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 7.767722138219336e-07,
+      "loss": 0.2349,
+      "step": 10048
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 7.753491061104687e-07,
+      "loss": 0.2277,
+      "step": 10112
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 7.739259983990037e-07,
+      "loss": 0.2239,
+      "step": 10176
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 7.725028906875389e-07,
+      "loss": 0.2707,
+      "step": 10240
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 7.71079782976074e-07,
+      "loss": 0.1957,
+      "step": 10304
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 7.69656675264609e-07,
+      "loss": 0.1649,
+      "step": 10368
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 7.682335675531442e-07,
+      "loss": 0.2773,
+      "step": 10432
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 7.668104598416792e-07,
+      "loss": 0.2146,
+      "step": 10496
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 7.653873521302144e-07,
+      "loss": 0.2134,
+      "step": 10560
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 7.639642444187494e-07,
+      "loss": 0.2168,
+      "step": 10624
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 7.625411367072845e-07,
+      "loss": 0.1122,
+      "step": 10688
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 7.611180289958195e-07,
+      "loss": 0.1628,
+      "step": 10752
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 7.596949212843547e-07,
+      "loss": 0.1566,
+      "step": 10816
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 7.582718135728897e-07,
+      "loss": 0.1824,
+      "step": 10880
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 7.56848705861425e-07,
+      "loss": 0.292,
+      "step": 10944
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 7.5542559814996e-07,
+      "loss": 0.1534,
+      "step": 11008
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 7.54002490438495e-07,
+      "loss": 0.2107,
+      "step": 11072
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 7.525793827270302e-07,
+      "loss": 0.1689,
+      "step": 11136
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 7.511562750155652e-07,
+      "loss": 0.3489,
+      "step": 11200
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 7.497331673041003e-07,
+      "loss": 0.1535,
+      "step": 11264
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 7.483100595926353e-07,
+      "loss": 0.1439,
+      "step": 11328
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 7.468869518811705e-07,
+      "loss": 0.2032,
+      "step": 11392
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 7.454638441697055e-07,
+      "loss": 0.4177,
+      "step": 11456
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 7.440407364582406e-07,
+      "loss": 0.2552,
+      "step": 11520
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 7.426176287467757e-07,
+      "loss": 0.181,
+      "step": 11584
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 7.411945210353108e-07,
+      "loss": 0.1128,
+      "step": 11648
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 7.39771413323846e-07,
+      "loss": 0.2184,
+      "step": 11712
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 7.383483056123811e-07,
+      "loss": 0.3284,
+      "step": 11776
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 7.369251979009161e-07,
+      "loss": 0.1976,
+      "step": 11840
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 7.355020901894511e-07,
+      "loss": 0.1274,
+      "step": 11904
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 7.340789824779863e-07,
+      "loss": 0.1578,
+      "step": 11968
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 7.326558747665213e-07,
+      "loss": 0.2599,
+      "step": 12032
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 7.312327670550564e-07,
+      "loss": 0.1812,
+      "step": 12096
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 7.298096593435915e-07,
+      "loss": 0.2296,
+      "step": 12160
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 7.283865516321266e-07,
+      "loss": 0.2716,
+      "step": 12224
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 7.269634439206616e-07,
+      "loss": 0.3587,
+      "step": 12288
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 7.255403362091968e-07,
+      "loss": 0.1644,
+      "step": 12352
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 7.241394645557235e-07,
+      "loss": 0.2707,
+      "step": 12416
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 7.227163568442586e-07,
+      "loss": 0.1747,
+      "step": 12480
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 7.212932491327936e-07,
+      "loss": 0.2109,
+      "step": 12544
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 7.198701414213289e-07,
+      "loss": 0.2764,
+      "step": 12608
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 7.184470337098639e-07,
+      "loss": 0.1857,
+      "step": 12672
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 7.170239259983991e-07,
+      "loss": 0.2993,
+      "step": 12736
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 7.156008182869341e-07,
+      "loss": 0.3019,
+      "step": 12800
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 7.141777105754691e-07,
+      "loss": 0.2171,
+      "step": 12864
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 7.127546028640042e-07,
+      "loss": 0.1616,
+      "step": 12928
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.113314951525393e-07,
+      "loss": 0.2056,
+      "step": 12992
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.099083874410744e-07,
+      "loss": 0.1741,
+      "step": 13056
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.084852797296094e-07,
+      "loss": 0.1881,
+      "step": 13120
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.070621720181446e-07,
+      "loss": 0.3325,
+      "step": 13184
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.056390643066796e-07,
+      "loss": 0.1797,
+      "step": 13248
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.042159565952149e-07,
+      "loss": 0.1904,
+      "step": 13312
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.027928488837499e-07,
+      "loss": 0.182,
+      "step": 13376
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.01369741172285e-07,
+      "loss": 0.1893,
+      "step": 13440
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 6.9994663346082e-07,
+      "loss": 0.2757,
+      "step": 13504
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 6.985235257493552e-07,
+      "loss": 0.1684,
+      "step": 13568
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.971004180378902e-07,
+      "loss": 0.1678,
+      "step": 13632
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.956773103264252e-07,
+      "loss": 0.2164,
+      "step": 13696
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.942542026149604e-07,
+      "loss": 0.3885,
+      "step": 13760
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.928310949034954e-07,
+      "loss": 0.153,
+      "step": 13824
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.914079871920305e-07,
+      "loss": 0.211,
+      "step": 13888
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.899848794805656e-07,
+      "loss": 0.2952,
+      "step": 13952
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.885617717691008e-07,
+      "loss": 0.227,
+      "step": 14016
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.871386640576358e-07,
+      "loss": 0.1933,
+      "step": 14080
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.85715556346171e-07,
+      "loss": 0.2457,
+      "step": 14144
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.84292448634706e-07,
+      "loss": 0.1388,
+      "step": 14208
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.828693409232412e-07,
+      "loss": 0.175,
+      "step": 14272
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.814462332117762e-07,
+      "loss": 0.1735,
+      "step": 14336
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.800231255003113e-07,
+      "loss": 0.2027,
+      "step": 14400
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.786000177888463e-07,
+      "loss": 0.3326,
+      "step": 14464
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 6.771769100773814e-07,
+      "loss": 0.2391,
+      "step": 14528
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 6.757538023659165e-07,
+      "loss": 0.1766,
+      "step": 14592
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 6.743306946544515e-07,
+      "loss": 0.195,
+      "step": 14656
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 6.729075869429868e-07,
+      "loss": 0.2176,
+      "step": 14720
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 6.714844792315218e-07,
+      "loss": 0.2863,
+      "step": 14784
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 6.700836075780485e-07,
+      "loss": 0.2664,
+      "step": 14848
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 6.686604998665835e-07,
+      "loss": 0.1482,
+      "step": 14912
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 6.672373921551188e-07,
+      "loss": 0.3435,
+      "step": 14976
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 6.658142844436538e-07,
+      "loss": 0.2416,
+      "step": 15040
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 6.64391176732189e-07,
+      "loss": 0.2352,
+      "step": 15104
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 6.62968069020724e-07,
+      "loss": 0.2318,
+      "step": 15168
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 6.615449613092591e-07,
+      "loss": 0.2973,
+      "step": 15232
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 6.601218535977941e-07,
+      "loss": 0.177,
+      "step": 15296
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 6.586987458863293e-07,
+      "loss": 0.2529,
+      "step": 15360
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 6.572756381748643e-07,
+      "loss": 0.1659,
+      "step": 15424
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 6.558747665213911e-07,
+      "loss": 0.1879,
+      "step": 15488
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 6.544516588099261e-07,
+      "loss": 0.296,
+      "step": 15552
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 6.530285510984612e-07,
+      "loss": 0.2029,
+      "step": 15616
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 6.516054433869963e-07,
+      "loss": 0.1833,
+      "step": 15680
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 6.501823356755313e-07,
+      "loss": 0.2408,
+      "step": 15744
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 6.487592279640665e-07,
+      "loss": 0.2729,
+      "step": 15808
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 6.473361202526015e-07,
+      "loss": 0.1371,
+      "step": 15872
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 6.459130125411368e-07,
+      "loss": 0.2234,
+      "step": 15936
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 6.444899048296718e-07,
+      "loss": 0.2095,
+      "step": 16000
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 6.430667971182069e-07,
+      "loss": 0.2612,
+      "step": 16064
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 6.41643689406742e-07,
+      "loss": 0.3269,
+      "step": 16128
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 6.402205816952771e-07,
+      "loss": 0.1677,
+      "step": 16192
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 6.387974739838121e-07,
+      "loss": 0.1625,
+      "step": 16256
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 6.373743662723472e-07,
+      "loss": 0.1675,
+      "step": 16320
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 6.359512585608823e-07,
+      "loss": 0.1452,
+      "step": 16384
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 6.345281508494173e-07,
+      "loss": 0.1458,
+      "step": 16448
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 6.331050431379524e-07,
+      "loss": 0.149,
+      "step": 16512
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 6.316819354264875e-07,
+      "loss": 0.329,
+      "step": 16576
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 6.302588277150227e-07,
+      "loss": 0.1425,
+      "step": 16640
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 6.288357200035577e-07,
+      "loss": 0.2696,
+      "step": 16704
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 6.274126122920929e-07,
+      "loss": 0.1641,
+      "step": 16768
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 6.259895045806279e-07,
+      "loss": 0.1661,
+      "step": 16832
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 6.24566396869163e-07,
+      "loss": 0.1707,
+      "step": 16896
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 6.231432891576981e-07,
+      "loss": 0.2579,
+      "step": 16960
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 6.217201814462332e-07,
+      "loss": 0.2654,
+      "step": 17024
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 6.202970737347682e-07,
+      "loss": 0.2443,
+      "step": 17088
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 6.188739660233034e-07,
+      "loss": 0.1523,
+      "step": 17152
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 6.174508583118384e-07,
+      "loss": 0.1256,
+      "step": 17216
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 6.160277506003734e-07,
+      "loss": 0.2002,
+      "step": 17280
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 6.146046428889087e-07,
+      "loss": 0.225,
+      "step": 17344
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 6.131815351774437e-07,
+      "loss": 0.2168,
+      "step": 17408
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 6.117584274659788e-07,
+      "loss": 0.1444,
+      "step": 17472
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 6.103353197545139e-07,
+      "loss": 0.1429,
+      "step": 17536
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 6.08912212043049e-07,
+      "loss": 0.1688,
+      "step": 17600
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 6.07489104331584e-07,
+      "loss": 0.2285,
+      "step": 17664
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 6.060659966201192e-07,
+      "loss": 0.2193,
+      "step": 17728
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 6.046428889086542e-07,
+      "loss": 0.1757,
+      "step": 17792
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 6.032197811971893e-07,
+      "loss": 0.2121,
+      "step": 17856
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 6.017966734857244e-07,
+      "loss": 0.2231,
+      "step": 17920
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 6.003735657742595e-07,
+      "loss": 0.213,
+      "step": 17984
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 5.989504580627946e-07,
+      "loss": 0.1744,
+      "step": 18048
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 5.975273503513297e-07,
+      "loss": 0.2606,
+      "step": 18112
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 5.961042426398648e-07,
+      "loss": 0.2494,
+      "step": 18176
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 5.946811349283998e-07,
+      "loss": 0.2053,
+      "step": 18240
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 5.93258027216935e-07,
+      "loss": 0.2282,
+      "step": 18304
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 5.9183491950547e-07,
+      "loss": 0.2375,
+      "step": 18368
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 5.904118117940051e-07,
+      "loss": 0.1211,
+      "step": 18432
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 5.889887040825402e-07,
+      "loss": 0.2294,
+      "step": 18496
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 5.875655963710753e-07,
+      "loss": 0.2236,
+      "step": 18560
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 5.861424886596103e-07,
+      "loss": 0.1577,
+      "step": 18624
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 5.847193809481455e-07,
+      "loss": 0.2457,
+      "step": 18688
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 5.832962732366806e-07,
+      "loss": 0.1634,
+      "step": 18752
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 5.818731655252158e-07,
+      "loss": 0.2724,
+      "step": 18816
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 5.804500578137508e-07,
+      "loss": 0.2411,
+      "step": 18880
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 5.790269501022858e-07,
+      "loss": 0.2015,
+      "step": 18944
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 5.776038423908209e-07,
+      "loss": 0.1731,
+      "step": 19008
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 5.76180734679356e-07,
+      "loss": 0.2573,
+      "step": 19072
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 5.747576269678911e-07,
+      "loss": 0.1844,
+      "step": 19136
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 5.733345192564261e-07,
+      "loss": 0.2007,
+      "step": 19200
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 5.719114115449613e-07,
+      "loss": 0.362,
+      "step": 19264
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 5.704883038334963e-07,
+      "loss": 0.2429,
+      "step": 19328
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 5.690651961220314e-07,
+      "loss": 0.2387,
+      "step": 19392
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 5.676420884105666e-07,
+      "loss": 0.1515,
+      "step": 19456
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 5.662189806991017e-07,
+      "loss": 0.0871,
+      "step": 19520
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 5.647958729876367e-07,
+      "loss": 0.2069,
+      "step": 19584
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 5.633727652761719e-07,
+      "loss": 0.2809,
+      "step": 19648
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 5.619496575647069e-07,
+      "loss": 0.1971,
+      "step": 19712
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 5.605265498532419e-07,
+      "loss": 0.2416,
+      "step": 19776
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 5.591034421417771e-07,
+      "loss": 0.2536,
+      "step": 19840
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 5.576803344303121e-07,
+      "loss": 0.199,
+      "step": 19904
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 5.562572267188472e-07,
+      "loss": 0.1612,
+      "step": 19968
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 5.548341190073823e-07,
+      "loss": 0.2325,
+      "step": 20032
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 5.534110112959174e-07,
+      "loss": 0.2497,
+      "step": 20096
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.519879035844525e-07,
+      "loss": 0.3172,
+      "step": 20160
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.505647958729877e-07,
+      "loss": 0.2571,
+      "step": 20224
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.491416881615227e-07,
+      "loss": 0.1585,
+      "step": 20288
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 5.477185804500578e-07,
+      "loss": 0.2229,
+      "step": 20352
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 5.462954727385929e-07,
+      "loss": 0.2502,
+      "step": 20416
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 5.44872365027128e-07,
+      "loss": 0.2151,
+      "step": 20480
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 5.43449257315663e-07,
+      "loss": 0.1644,
+      "step": 20544
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 5.420261496041981e-07,
+      "loss": 0.3373,
+      "step": 20608
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 5.406030418927332e-07,
+      "loss": 0.1322,
+      "step": 20672
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 5.391799341812682e-07,
+      "loss": 0.213,
+      "step": 20736
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 5.377568264698034e-07,
+      "loss": 0.169,
+      "step": 20800
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 5.363337187583385e-07,
+      "loss": 0.2652,
+      "step": 20864
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 5.349106110468736e-07,
+      "loss": 0.2234,
+      "step": 20928
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 5.334875033354087e-07,
+      "loss": 0.1933,
+      "step": 20992
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 5.320866316819354e-07,
+      "loss": 0.2141,
+      "step": 21056
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 5.306635239704705e-07,
+      "loss": 0.217,
+      "step": 21120
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 5.292404162590056e-07,
+      "loss": 0.2091,
+      "step": 21184
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 5.278173085475407e-07,
+      "loss": 0.2001,
+      "step": 21248
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 5.263942008360758e-07,
+      "loss": 0.1489,
+      "step": 21312
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 5.249710931246108e-07,
+      "loss": 0.2728,
+      "step": 21376
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 5.23547985413146e-07,
+      "loss": 0.1827,
+      "step": 21440
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 5.22124877701681e-07,
+      "loss": 0.1933,
+      "step": 21504
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 5.20701769990216e-07,
+      "loss": 0.1068,
+      "step": 21568
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 5.192786622787512e-07,
+      "loss": 0.2462,
+      "step": 21632
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 5.178555545672862e-07,
+      "loss": 0.3433,
+      "step": 21696
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 5.164324468558213e-07,
+      "loss": 0.2489,
+      "step": 21760
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 5.150093391443565e-07,
+      "loss": 0.2038,
+      "step": 21824
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 5.135862314328916e-07,
+      "loss": 0.2269,
+      "step": 21888
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 5.121631237214266e-07,
+      "loss": 0.1911,
+      "step": 21952
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 5.107400160099618e-07,
+      "loss": 0.2203,
+      "step": 22016
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 5.093169082984968e-07,
+      "loss": 0.2476,
+      "step": 22080
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 5.078938005870319e-07,
+      "loss": 0.2525,
+      "step": 22144
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 5.06470692875567e-07,
+      "loss": 0.1442,
+      "step": 22208
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 5.050475851641021e-07,
+      "loss": 0.1759,
+      "step": 22272
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 5.036244774526371e-07,
+      "loss": 0.2944,
+      "step": 22336
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 5.022013697411722e-07,
+      "loss": 0.1416,
+      "step": 22400
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 5.007782620297073e-07,
+      "loss": 0.2061,
+      "step": 22464
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.993551543182424e-07,
+      "loss": 0.2197,
+      "step": 22528
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.979320466067775e-07,
+      "loss": 0.1192,
+      "step": 22592
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 4.965089388953126e-07,
+      "loss": 0.1117,
+      "step": 22656
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 4.950858311838477e-07,
+      "loss": 0.1859,
+      "step": 22720
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 4.936627234723828e-07,
+      "loss": 0.1986,
+      "step": 22784
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 4.922396157609179e-07,
+      "loss": 0.2719,
+      "step": 22848
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 4.908165080494529e-07,
+      "loss": 0.1638,
+      "step": 22912
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 4.893934003379881e-07,
+      "loss": 0.0792,
+      "step": 22976
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 4.879702926265232e-07,
+      "loss": 0.1522,
+      "step": 23040
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 4.865471849150582e-07,
+      "loss": 0.1592,
+      "step": 23104
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 4.851240772035934e-07,
+      "loss": 0.1597,
+      "step": 23168
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 4.837009694921284e-07,
+      "loss": 0.1055,
+      "step": 23232
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 4.822778617806634e-07,
+      "loss": 0.1207,
+      "step": 23296
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 4.808547540691986e-07,
+      "loss": 0.1717,
+      "step": 23360
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 4.794538824157253e-07,
+      "loss": 0.1305,
+      "step": 23424
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 4.780307747042604e-07,
+      "loss": 0.0869,
+      "step": 23488
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 4.766076669927955e-07,
+      "loss": 0.1097,
+      "step": 23552
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 4.7518455928133063e-07,
+      "loss": 0.2356,
+      "step": 23616
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 4.737614515698657e-07,
+      "loss": 0.1326,
+      "step": 23680
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 4.7233834385840074e-07,
+      "loss": 0.1271,
+      "step": 23744
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 4.709152361469358e-07,
+      "loss": 0.1454,
+      "step": 23808
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 4.694921284354709e-07,
+      "loss": 0.1813,
+      "step": 23872
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 4.68069020724006e-07,
+      "loss": 0.1332,
+      "step": 23936
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 4.6664591301254113e-07,
+      "loss": 0.1769,
+      "step": 24000
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 4.652228053010762e-07,
+      "loss": 0.1574,
+      "step": 24064
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 4.637996975896113e-07,
+      "loss": 0.1397,
+      "step": 24128
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 4.623765898781464e-07,
+      "loss": 0.2307,
+      "step": 24192
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 4.6095348216668146e-07,
+      "loss": 0.0829,
+      "step": 24256
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 4.595303744552166e-07,
+      "loss": 0.1801,
+      "step": 24320
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 4.581072667437517e-07,
+      "loss": 0.1314,
+      "step": 24384
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 4.5668415903228676e-07,
+      "loss": 0.1002,
+      "step": 24448
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 4.5526105132082184e-07,
+      "loss": 0.1372,
+      "step": 24512
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 4.538379436093569e-07,
+      "loss": 0.1843,
+      "step": 24576
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 4.5241483589789196e-07,
+      "loss": 0.1166,
+      "step": 24640
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 4.509917281864271e-07,
+      "loss": 0.1739,
+      "step": 24704
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 4.495686204749622e-07,
+      "loss": 0.1977,
+      "step": 24768
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 4.4814551276349726e-07,
+      "loss": 0.2044,
+      "step": 24832
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 4.4672240505203234e-07,
+      "loss": 0.1625,
+      "step": 24896
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 4.452992973405674e-07,
+      "loss": 0.1712,
+      "step": 24960
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 4.4387618962910256e-07,
+      "loss": 0.0945,
+      "step": 25024
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 4.4245308191763764e-07,
+      "loss": 0.0845,
+      "step": 25088
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 4.4102997420617273e-07,
+      "loss": 0.1123,
+      "step": 25152
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 4.396068664947078e-07,
+      "loss": 0.1598,
+      "step": 25216
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 4.381837587832429e-07,
+      "loss": 0.173,
+      "step": 25280
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 4.36760651071778e-07,
+      "loss": 0.159,
+      "step": 25344
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 4.353375433603131e-07,
+      "loss": 0.1855,
+      "step": 25408
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 4.3391443564884814e-07,
+      "loss": 0.1274,
+      "step": 25472
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 4.324913279373832e-07,
+      "loss": 0.1939,
+      "step": 25536
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 4.310682202259183e-07,
+      "loss": 0.1458,
+      "step": 25600
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 4.296451125144534e-07,
+      "loss": 0.0993,
+      "step": 25664
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 4.282220048029885e-07,
+      "loss": 0.1084,
+      "step": 25728
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 4.267988970915236e-07,
+      "loss": 0.092,
+      "step": 25792
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 4.253757893800587e-07,
+      "loss": 0.1503,
+      "step": 25856
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 4.239526816685938e-07,
+      "loss": 0.144,
+      "step": 25920
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 4.2252957395712886e-07,
+      "loss": 0.075,
+      "step": 25984
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 4.2110646624566394e-07,
+      "loss": 0.1614,
+      "step": 26048
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 4.196833585341991e-07,
+      "loss": 0.1753,
+      "step": 26112
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 4.1826025082273416e-07,
+      "loss": 0.2336,
+      "step": 26176
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 4.1683714311126924e-07,
+      "loss": 0.1453,
+      "step": 26240
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 4.154140353998043e-07,
+      "loss": 0.1707,
+      "step": 26304
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 4.1399092768833936e-07,
+      "loss": 0.1123,
+      "step": 26368
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 4.1256781997687444e-07,
+      "loss": 0.1998,
+      "step": 26432
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 4.111447122654096e-07,
+      "loss": 0.1178,
+      "step": 26496
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 4.0972160455394466e-07,
+      "loss": 0.1781,
+      "step": 26560
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 4.0829849684247974e-07,
+      "loss": 0.0921,
+      "step": 26624
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 4.068753891310148e-07,
+      "loss": 0.0661,
+      "step": 26688
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 4.054522814195499e-07,
+      "loss": 0.1927,
+      "step": 26752
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 4.0402917370808504e-07,
+      "loss": 0.1816,
+      "step": 26816
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 4.0262830205461174e-07,
+      "loss": 0.2152,
+      "step": 26880
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 4.012051943431468e-07,
+      "loss": 0.181,
+      "step": 26944
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 3.997820866316819e-07,
+      "loss": 0.1332,
+      "step": 27008
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 3.9835897892021704e-07,
+      "loss": 0.077,
+      "step": 27072
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 3.9693587120875213e-07,
+      "loss": 0.1626,
+      "step": 27136
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 3.9551276349728716e-07,
+      "loss": 0.1894,
+      "step": 27200
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 3.9408965578582224e-07,
+      "loss": 0.1509,
+      "step": 27264
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 3.926665480743573e-07,
+      "loss": 0.0634,
+      "step": 27328
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 3.9124344036289246e-07,
+      "loss": 0.1287,
+      "step": 27392
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 3.8982033265142754e-07,
+      "loss": 0.0669,
+      "step": 27456
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 3.883972249399626e-07,
+      "loss": 0.1046,
+      "step": 27520
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 3.869741172284977e-07,
+      "loss": 0.0747,
+      "step": 27584
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 3.855510095170328e-07,
+      "loss": 0.1554,
+      "step": 27648
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 3.841279018055679e-07,
+      "loss": 0.1803,
+      "step": 27712
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 3.82704794094103e-07,
+      "loss": 0.1479,
+      "step": 27776
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 3.812816863826381e-07,
+      "loss": 0.1667,
+      "step": 27840
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 3.798585786711732e-07,
+      "loss": 0.2049,
+      "step": 27904
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 3.7843547095970826e-07,
+      "loss": 0.1617,
+      "step": 27968
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 3.770123632482433e-07,
+      "loss": 0.0975,
+      "step": 28032
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 3.755892555367784e-07,
+      "loss": 0.101,
+      "step": 28096
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 3.741661478253135e-07,
+      "loss": 0.1224,
+      "step": 28160
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 3.727430401138486e-07,
+      "loss": 0.1519,
+      "step": 28224
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 3.713199324023837e-07,
+      "loss": 0.1175,
+      "step": 28288
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 3.6989682469091876e-07,
+      "loss": 0.2126,
+      "step": 28352
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 3.6847371697945384e-07,
+      "loss": 0.1332,
+      "step": 28416
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 3.67050609267989e-07,
+      "loss": 0.1906,
+      "step": 28480
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 3.6562750155652406e-07,
+      "loss": 0.1032,
+      "step": 28544
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 3.6420439384505914e-07,
+      "loss": 0.096,
+      "step": 28608
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 3.627812861335942e-07,
+      "loss": 0.1452,
+      "step": 28672
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 3.613581784221293e-07,
+      "loss": 0.1455,
+      "step": 28736
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 3.5993507071066444e-07,
+      "loss": 0.103,
+      "step": 28800
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 3.5851196299919953e-07,
+      "loss": 0.1293,
+      "step": 28864
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 3.5708885528773456e-07,
+      "loss": 0.1759,
+      "step": 28928
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 3.5566574757626964e-07,
+      "loss": 0.1644,
+      "step": 28992
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 3.542426398648047e-07,
+      "loss": 0.0729,
+      "step": 29056
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 3.528195321533398e-07,
+      "loss": 0.1703,
+      "step": 29120
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 3.5139642444187494e-07,
+      "loss": 0.1259,
+      "step": 29184
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 3.4997331673041e-07,
+      "loss": 0.121,
+      "step": 29248
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 3.485502090189451e-07,
+      "loss": 0.1117,
+      "step": 29312
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 3.471271013074802e-07,
+      "loss": 0.1015,
+      "step": 29376
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 3.457039935960153e-07,
+      "loss": 0.1409,
+      "step": 29440
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 3.442808858845504e-07,
+      "loss": 0.1358,
+      "step": 29504
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 3.428577781730855e-07,
+      "loss": 0.1704,
+      "step": 29568
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 3.414346704616206e-07,
+      "loss": 0.2159,
+      "step": 29632
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 3.4001156275015566e-07,
+      "loss": 0.116,
+      "step": 29696
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 3.385884550386907e-07,
+      "loss": 0.2161,
+      "step": 29760
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 3.3716534732722577e-07,
+      "loss": 0.1037,
+      "step": 29824
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 3.357422396157609e-07,
+      "loss": 0.1729,
+      "step": 29888
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 3.34319131904296e-07,
+      "loss": 0.151,
+      "step": 29952
+    }
+  ],
+  "logging_steps": 64,
+  "max_steps": 44972,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 2,
+  "save_steps": 14990,
+  "total_flos": 6.3652667203584e+19,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-29980/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2463dffe438013cfefdc9915eb0241a5214c5766ec68d593ec6bac88b5ea2196
+size 5048

checkpoint-44970/config.json

@@ -0,0 +1,66 @@
+{
+  "_name_or_path": "openai/whisper-large-v2",
+  "activation_dropout": 0.0,
+  "activation_function": "gelu",
+  "apply_spec_augment": false,
+  "architectures": [
+    "WhisperForConditionalGeneration"
+  ],
+  "attention_dropout": 0.0,
+  "begin_suppress_tokens": [
+    220,
+    50257
+  ],
+  "bos_token_id": 50257,
+  "classifier_proj_size": 256,
+  "d_model": 1280,
+  "decoder_attention_heads": 20,
+  "decoder_ffn_dim": 5120,
+  "decoder_layerdrop": 0.0,
+  "decoder_layers": 32,
+  "decoder_start_token_id": 50258,
+  "dropout": 0.0,
+  "encoder_attention_heads": 20,
+  "encoder_ffn_dim": 5120,
+  "encoder_layerdrop": 0.0,
+  "encoder_layers": 32,
+  "eos_token_id": 50257,
+  "forced_decoder_ids": [
+    [
+      1,
+      50259
+    ],
+    [
+      2,
+      50359
+    ],
+    [
+      3,
+      50363
+    ]
+  ],
+  "init_std": 0.02,
+  "is_encoder_decoder": true,
+  "language": "English",
+  "mask_feature_length": 10,
+  "mask_feature_min_masks": 0,
+  "mask_feature_prob": 0.0,
+  "mask_time_length": 10,
+  "mask_time_min_masks": 2,
+  "mask_time_prob": 0.05,
+  "max_length": 448,
+  "max_source_positions": 1500,
+  "max_target_positions": 448,
+  "median_filter_width": 7,
+  "model_type": "whisper",
+  "num_hidden_layers": 32,
+  "num_mel_bins": 80,
+  "pad_token_id": 50257,
+  "scale_embedding": false,
+  "suppress_tokens": [],
+  "torch_dtype": "float32",
+  "transformers_version": "4.37.2",
+  "use_cache": true,
+  "use_weighted_layer_sum": false,
+  "vocab_size": 51865
+}

checkpoint-44970/generation_config.json

@@ -0,0 +1,317 @@
+{
+  "alignment_heads": [
+    [
+      10,
+      12
+    ],
+    [
+      13,
+      17
+    ],
+    [
+      16,
+      11
+    ],
+    [
+      16,
+      12
+    ],
+    [
+      16,
+      13
+    ],
+    [
+      17,
+      15
+    ],
+    [
+      17,
+      16
+    ],
+    [
+      18,
+      4
+    ],
+    [
+      18,
+      11
+    ],
+    [
+      18,
+      19
+    ],
+    [
+      19,
+      11
+    ],
+    [
+      21,
+      2
+    ],
+    [
+      21,
+      3
+    ],
+    [
+      22,
+      3
+    ],
+    [
+      22,
+      9
+    ],
+    [
+      22,
+      12
+    ],
+    [
+      23,
+      5
+    ],
+    [
+      23,
+      7
+    ],
+    [
+      23,
+      13
+    ],
+    [
+      25,
+      5
+    ],
+    [
+      26,
+      1
+    ],
+    [
+      26,
+      12
+    ],
+    [
+      27,
+      15
+    ]
+  ],
+  "begin_suppress_tokens": [
+    220,
+    50257
+  ],
+  "bos_token_id": 50257,
+  "decoder_start_token_id": 50258,
+  "eos_token_id": 50257,
+  "forced_decoder_ids": [
+    [
+      1,
+      null
+    ],
+    [
+      2,
+      50359
+    ]
+  ],
+  "is_multilingual": true,
+  "lang_to_id": {
+    "<|af|>": 50327,
+    "<|am|>": 50334,
+    "<|ar|>": 50272,
+    "<|as|>": 50350,
+    "<|az|>": 50304,
+    "<|ba|>": 50355,
+    "<|be|>": 50330,
+    "<|bg|>": 50292,
+    "<|bn|>": 50302,
+    "<|bo|>": 50347,
+    "<|br|>": 50309,
+    "<|bs|>": 50315,
+    "<|ca|>": 50270,
+    "<|cs|>": 50283,
+    "<|cy|>": 50297,
+    "<|da|>": 50285,
+    "<|de|>": 50261,
+    "<|el|>": 50281,
+    "<|en|>": 50259,
+    "<|es|>": 50262,
+    "<|et|>": 50307,
+    "<|eu|>": 50310,
+    "<|fa|>": 50300,
+    "<|fi|>": 50277,
+    "<|fo|>": 50338,
+    "<|fr|>": 50265,
+    "<|gl|>": 50319,
+    "<|gu|>": 50333,
+    "<|haw|>": 50352,
+    "<|ha|>": 50354,
+    "<|he|>": 50279,
+    "<|hi|>": 50276,
+    "<|hr|>": 50291,
+    "<|ht|>": 50339,
+    "<|hu|>": 50286,
+    "<|hy|>": 50312,
+    "<|id|>": 50275,
+    "<|is|>": 50311,
+    "<|it|>": 50274,
+    "<|ja|>": 50266,
+    "<|jw|>": 50356,
+    "<|ka|>": 50329,
+    "<|kk|>": 50316,
+    "<|km|>": 50323,
+    "<|kn|>": 50306,
+    "<|ko|>": 50264,
+    "<|la|>": 50294,
+    "<|lb|>": 50345,
+    "<|ln|>": 50353,
+    "<|lo|>": 50336,
+    "<|lt|>": 50293,
+    "<|lv|>": 50301,
+    "<|mg|>": 50349,
+    "<|mi|>": 50295,
+    "<|mk|>": 50308,
+    "<|ml|>": 50296,
+    "<|mn|>": 50314,
+    "<|mr|>": 50320,
+    "<|ms|>": 50282,
+    "<|mt|>": 50343,
+    "<|my|>": 50346,
+    "<|ne|>": 50313,
+    "<|nl|>": 50271,
+    "<|nn|>": 50342,
+    "<|no|>": 50288,
+    "<|oc|>": 50328,
+    "<|pa|>": 50321,
+    "<|pl|>": 50269,
+    "<|ps|>": 50340,
+    "<|pt|>": 50267,
+    "<|ro|>": 50284,
+    "<|ru|>": 50263,
+    "<|sa|>": 50344,
+    "<|sd|>": 50332,
+    "<|si|>": 50322,
+    "<|sk|>": 50298,
+    "<|sl|>": 50305,
+    "<|sn|>": 50324,
+    "<|so|>": 50326,
+    "<|sq|>": 50317,
+    "<|sr|>": 50303,
+    "<|su|>": 50357,
+    "<|sv|>": 50273,
+    "<|sw|>": 50318,
+    "<|ta|>": 50287,
+    "<|te|>": 50299,
+    "<|tg|>": 50331,
+    "<|th|>": 50289,
+    "<|tk|>": 50341,
+    "<|tl|>": 50348,
+    "<|tr|>": 50268,
+    "<|tt|>": 50351,
+    "<|uk|>": 50280,
+    "<|ur|>": 50290,
+    "<|uz|>": 50337,
+    "<|vi|>": 50278,
+    "<|yi|>": 50335,
+    "<|yo|>": 50325,
+    "<|zh|>": 50260
+  },
+  "max_initial_timestamp_index": 50,
+  "max_length": 448,
+  "no_timestamps_token_id": 50363,
+  "pad_token_id": 50257,
+  "prev_sot_token_id": 50361,
+  "return_timestamps": false,
+  "suppress_tokens": [
+    1,
+    2,
+    7,
+    8,
+    9,
+    10,
+    14,
+    25,
+    26,
+    27,
+    28,
+    29,
+    31,
+    58,
+    59,
+    60,
+    61,
+    62,
+    63,
+    90,
+    91,
+    92,
+    93,
+    359,
+    503,
+    522,
+    542,
+    873,
+    893,
+    902,
+    918,
+    922,
+    931,
+    1350,
+    1853,
+    1982,
+    2460,
+    2627,
+    3246,
+    3253,
+    3268,
+    3536,
+    3846,
+    3961,
+    4183,
+    4667,
+    6585,
+    6647,
+    7273,
+    9061,
+    9383,
+    10428,
+    10929,
+    11938,
+    12033,
+    12331,
+    12562,
+    13793,
+    14157,
+    14635,
+    15265,
+    15618,
+    16553,
+    16604,
+    18362,
+    18956,
+    20075,
+    21675,
+    22520,
+    26130,
+    26161,
+    26435,
+    28279,
+    29464,
+    31650,
+    32302,
+    32470,
+    36865,
+    42863,
+    47425,
+    49870,
+    50254,
+    50258,
+    50358,
+    50359,
+    50360,
+    50361,
+    50362
+  ],
+  "task": "transcribe",
+  "task_to_id": {
+    "transcribe": 50359,
+    "translate": 50358
+  },
+  "transformers_version": "4.37.2"
+}

checkpoint-44970/model-00001-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ae1543353453fed95f5a030792526a079ceea2e2a98338ecf7c40a7b19163f99
+size 4992706480

checkpoint-44970/model-00002-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bd856bdebbb525feb68794398ad8b57afc186f28efd03c381c349331f34a254b
+size 1180663192

checkpoint-44970/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b8d8abb2891bf59787b010f3222565dfa5287f54212128dfbb1fce461ab6b3c7
+size 3095074288

checkpoint-44970/preprocessor_config.json

@@ -0,0 +1,14 @@
+{
+  "chunk_length": 30,
+  "feature_extractor_type": "WhisperFeatureExtractor",
+  "feature_size": 80,
+  "hop_length": 160,
+  "n_fft": 400,
+  "n_samples": 480000,
+  "nb_max_frames": 3000,
+  "padding_side": "right",
+  "padding_value": 0.0,
+  "processor_class": "WhisperProcessor",
+  "return_attention_mask": false,
+  "sampling_rate": 16000
+}

checkpoint-44970/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:017534de4b6dec166023e1ed52fc8b116c76c212b0e10a54d409ed49978fa5af
+size 14244

checkpoint-44970/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5c21d2a87ef3e200406826d1b3eebed770b60f8d67ed77cd133aeaa515450c21
+size 1064

checkpoint-44970/trainer_state.json

@@ -0,0 +1,4233 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.9999110557680333,
+  "eval_steps": 500,
+  "global_step": 44970,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "learning_rate": 9.986880725784932e-07,
+      "loss": 5.3038,
+      "step": 64
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 9.972649648670284e-07,
+      "loss": 2.7293,
+      "step": 128
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 9.959085653295383e-07,
+      "loss": 1.6865,
+      "step": 192
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 9.944854576180734e-07,
+      "loss": 0.3066,
+      "step": 256
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 9.930623499066086e-07,
+      "loss": 0.2533,
+      "step": 320
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 9.916392421951437e-07,
+      "loss": 0.3489,
+      "step": 384
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 9.902161344836786e-07,
+      "loss": 0.2908,
+      "step": 448
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 9.887930267722138e-07,
+      "loss": 0.3666,
+      "step": 512
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 9.873699190607489e-07,
+      "loss": 0.3543,
+      "step": 576
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 9.85946811349284e-07,
+      "loss": 0.2938,
+      "step": 640
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 9.84523703637819e-07,
+      "loss": 0.2674,
+      "step": 704
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 9.83100595926354e-07,
+      "loss": 0.2186,
+      "step": 768
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 9.816774882148892e-07,
+      "loss": 0.3472,
+      "step": 832
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 9.802543805034244e-07,
+      "loss": 0.3157,
+      "step": 896
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 9.788312727919595e-07,
+      "loss": 0.432,
+      "step": 960
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 9.774081650804944e-07,
+      "loss": 0.2131,
+      "step": 1024
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 9.759850573690296e-07,
+      "loss": 0.249,
+      "step": 1088
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 9.745619496575647e-07,
+      "loss": 0.239,
+      "step": 1152
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 9.731388419460998e-07,
+      "loss": 0.3577,
+      "step": 1216
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 9.717157342346347e-07,
+      "loss": 0.1945,
+      "step": 1280
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 9.702926265231699e-07,
+      "loss": 0.3107,
+      "step": 1344
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 9.68869518811705e-07,
+      "loss": 0.1983,
+      "step": 1408
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 9.674464111002402e-07,
+      "loss": 0.188,
+      "step": 1472
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 9.660233033887753e-07,
+      "loss": 0.2659,
+      "step": 1536
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 9.646001956773104e-07,
+      "loss": 0.442,
+      "step": 1600
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 9.631770879658454e-07,
+      "loss": 0.1774,
+      "step": 1664
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 9.617539802543805e-07,
+      "loss": 0.2944,
+      "step": 1728
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 9.603308725429156e-07,
+      "loss": 0.3228,
+      "step": 1792
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 9.589077648314505e-07,
+      "loss": 0.2818,
+      "step": 1856
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 9.574846571199857e-07,
+      "loss": 0.2816,
+      "step": 1920
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 9.560615494085208e-07,
+      "loss": 0.3663,
+      "step": 1984
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 9.54638441697056e-07,
+      "loss": 0.2614,
+      "step": 2048
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 9.53215333985591e-07,
+      "loss": 0.2934,
+      "step": 2112
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 9.51792226274126e-07,
+      "loss": 0.4422,
+      "step": 2176
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 9.503691185626613e-07,
+      "loss": 0.2562,
+      "step": 2240
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 9.489460108511963e-07,
+      "loss": 0.3432,
+      "step": 2304
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 9.475229031397314e-07,
+      "loss": 0.3384,
+      "step": 2368
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 9.460997954282665e-07,
+      "loss": 0.1679,
+      "step": 2432
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 9.446766877168015e-07,
+      "loss": 0.1767,
+      "step": 2496
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 9.432535800053366e-07,
+      "loss": 0.2161,
+      "step": 2560
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 9.418304722938717e-07,
+      "loss": 0.2729,
+      "step": 2624
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 9.404073645824068e-07,
+      "loss": 0.3314,
+      "step": 2688
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 9.389842568709418e-07,
+      "loss": 0.242,
+      "step": 2752
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 9.37561149159477e-07,
+      "loss": 0.3446,
+      "step": 2816
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 9.36138041448012e-07,
+      "loss": 0.1292,
+      "step": 2880
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 9.347149337365472e-07,
+      "loss": 0.1854,
+      "step": 2944
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 9.332918260250823e-07,
+      "loss": 0.2536,
+      "step": 3008
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 9.318687183136174e-07,
+      "loss": 0.2224,
+      "step": 3072
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 9.304456106021524e-07,
+      "loss": 0.237,
+      "step": 3136
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 9.290225028906876e-07,
+      "loss": 0.3345,
+      "step": 3200
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 9.275993951792226e-07,
+      "loss": 0.175,
+      "step": 3264
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 9.261762874677576e-07,
+      "loss": 0.2693,
+      "step": 3328
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 9.247531797562928e-07,
+      "loss": 0.254,
+      "step": 3392
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 9.233300720448278e-07,
+      "loss": 0.2855,
+      "step": 3456
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 9.219069643333629e-07,
+      "loss": 0.1507,
+      "step": 3520
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 9.20483856621898e-07,
+      "loss": 0.2464,
+      "step": 3584
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 9.190607489104332e-07,
+      "loss": 0.2522,
+      "step": 3648
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 9.176376411989682e-07,
+      "loss": 0.2378,
+      "step": 3712
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 9.162145334875034e-07,
+      "loss": 0.2433,
+      "step": 3776
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 9.147914257760384e-07,
+      "loss": 0.2932,
+      "step": 3840
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 9.133683180645735e-07,
+      "loss": 0.2602,
+      "step": 3904
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 9.119452103531086e-07,
+      "loss": 0.3365,
+      "step": 3968
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 9.105221026416437e-07,
+      "loss": 0.1738,
+      "step": 4032
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 9.090989949301787e-07,
+      "loss": 0.3132,
+      "step": 4096
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 9.076758872187137e-07,
+      "loss": 0.2353,
+      "step": 4160
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 9.062527795072489e-07,
+      "loss": 0.2868,
+      "step": 4224
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 9.048296717957839e-07,
+      "loss": 0.1807,
+      "step": 4288
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 9.034065640843192e-07,
+      "loss": 0.3052,
+      "step": 4352
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 9.019834563728542e-07,
+      "loss": 0.2848,
+      "step": 4416
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 9.005603486613893e-07,
+      "loss": 0.2031,
+      "step": 4480
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 8.991372409499244e-07,
+      "loss": 0.2588,
+      "step": 4544
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 8.977141332384595e-07,
+      "loss": 0.2356,
+      "step": 4608
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 8.962910255269945e-07,
+      "loss": 0.2445,
+      "step": 4672
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 8.948679178155297e-07,
+      "loss": 0.2537,
+      "step": 4736
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 8.934448101040647e-07,
+      "loss": 0.2345,
+      "step": 4800
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 8.920217023925998e-07,
+      "loss": 0.2098,
+      "step": 4864
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 8.905985946811349e-07,
+      "loss": 0.3487,
+      "step": 4928
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 8.891754869696699e-07,
+      "loss": 0.1884,
+      "step": 4992
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 8.877523792582051e-07,
+      "loss": 0.273,
+      "step": 5056
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 8.863292715467402e-07,
+      "loss": 0.27,
+      "step": 5120
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 8.849061638352753e-07,
+      "loss": 0.1494,
+      "step": 5184
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 8.834830561238103e-07,
+      "loss": 0.1452,
+      "step": 5248
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 8.820599484123455e-07,
+      "loss": 0.177,
+      "step": 5312
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 8.806368407008805e-07,
+      "loss": 0.3317,
+      "step": 5376
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 8.792137329894156e-07,
+      "loss": 0.1638,
+      "step": 5440
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 8.777906252779507e-07,
+      "loss": 0.2928,
+      "step": 5504
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 8.763675175664858e-07,
+      "loss": 0.1851,
+      "step": 5568
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 8.749444098550208e-07,
+      "loss": 0.3108,
+      "step": 5632
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 8.73521302143556e-07,
+      "loss": 0.3002,
+      "step": 5696
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 8.720981944320911e-07,
+      "loss": 0.2726,
+      "step": 5760
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 8.706750867206262e-07,
+      "loss": 0.1775,
+      "step": 5824
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 8.692519790091613e-07,
+      "loss": 0.2632,
+      "step": 5888
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 8.678288712976963e-07,
+      "loss": 0.1625,
+      "step": 5952
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 8.664057635862314e-07,
+      "loss": 0.235,
+      "step": 6016
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 8.649826558747665e-07,
+      "loss": 0.2707,
+      "step": 6080
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 8.635595481633016e-07,
+      "loss": 0.1704,
+      "step": 6144
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 8.621364404518366e-07,
+      "loss": 0.1925,
+      "step": 6208
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 8.607133327403718e-07,
+      "loss": 0.2159,
+      "step": 6272
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 8.592902250289068e-07,
+      "loss": 0.2228,
+      "step": 6336
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 8.578671173174419e-07,
+      "loss": 0.2184,
+      "step": 6400
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 8.56444009605977e-07,
+      "loss": 0.2803,
+      "step": 6464
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 8.550209018945122e-07,
+      "loss": 0.1729,
+      "step": 6528
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 8.535977941830472e-07,
+      "loss": 0.2702,
+      "step": 6592
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 8.521746864715824e-07,
+      "loss": 0.2938,
+      "step": 6656
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 8.507515787601174e-07,
+      "loss": 0.2692,
+      "step": 6720
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 8.493284710486524e-07,
+      "loss": 0.3317,
+      "step": 6784
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 8.479053633371876e-07,
+      "loss": 0.2815,
+      "step": 6848
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 8.464822556257226e-07,
+      "loss": 0.2568,
+      "step": 6912
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 8.450591479142577e-07,
+      "loss": 0.2247,
+      "step": 6976
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 8.436360402027927e-07,
+      "loss": 0.2673,
+      "step": 7040
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 8.422129324913279e-07,
+      "loss": 0.1592,
+      "step": 7104
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 8.407898247798629e-07,
+      "loss": 0.2298,
+      "step": 7168
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 8.393667170683982e-07,
+      "loss": 0.2149,
+      "step": 7232
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 8.379436093569332e-07,
+      "loss": 0.3177,
+      "step": 7296
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 8.365205016454683e-07,
+      "loss": 0.2083,
+      "step": 7360
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 8.350973939340034e-07,
+      "loss": 0.2936,
+      "step": 7424
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 8.336742862225385e-07,
+      "loss": 0.2263,
+      "step": 7488
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 8.322511785110735e-07,
+      "loss": 0.2245,
+      "step": 7552
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 8.308280707996085e-07,
+      "loss": 0.1807,
+      "step": 7616
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 8.294049630881437e-07,
+      "loss": 0.2844,
+      "step": 7680
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 8.279818553766787e-07,
+      "loss": 0.1844,
+      "step": 7744
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 8.265587476652138e-07,
+      "loss": 0.1911,
+      "step": 7808
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 8.251578760117405e-07,
+      "loss": 0.2085,
+      "step": 7872
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 8.237347683002757e-07,
+      "loss": 0.1718,
+      "step": 7936
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 8.223116605888107e-07,
+      "loss": 0.2387,
+      "step": 8000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 8.208885528773458e-07,
+      "loss": 0.1725,
+      "step": 8064
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 8.19465445165881e-07,
+      "loss": 0.1448,
+      "step": 8128
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 8.180423374544161e-07,
+      "loss": 0.154,
+      "step": 8192
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 8.166192297429512e-07,
+      "loss": 0.3209,
+      "step": 8256
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 8.151961220314863e-07,
+      "loss": 0.2067,
+      "step": 8320
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 8.137730143200213e-07,
+      "loss": 0.2147,
+      "step": 8384
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 8.123499066085565e-07,
+      "loss": 0.1771,
+      "step": 8448
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 8.109267988970915e-07,
+      "loss": 0.1212,
+      "step": 8512
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 8.095036911856265e-07,
+      "loss": 0.1944,
+      "step": 8576
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 8.080805834741616e-07,
+      "loss": 0.2416,
+      "step": 8640
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 8.066574757626967e-07,
+      "loss": 0.2004,
+      "step": 8704
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 8.052343680512318e-07,
+      "loss": 0.1667,
+      "step": 8768
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 8.03811260339767e-07,
+      "loss": 0.203,
+      "step": 8832
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 8.023881526283021e-07,
+      "loss": 0.2355,
+      "step": 8896
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 8.009650449168371e-07,
+      "loss": 0.2247,
+      "step": 8960
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 7.995419372053723e-07,
+      "loss": 0.1768,
+      "step": 9024
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 7.981188294939073e-07,
+      "loss": 0.2669,
+      "step": 9088
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 7.966957217824424e-07,
+      "loss": 0.2288,
+      "step": 9152
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 7.952726140709774e-07,
+      "loss": 0.2397,
+      "step": 9216
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 7.938495063595126e-07,
+      "loss": 0.2276,
+      "step": 9280
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 7.924263986480476e-07,
+      "loss": 0.1367,
+      "step": 9344
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 7.910032909365826e-07,
+      "loss": 0.222,
+      "step": 9408
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 7.895801832251178e-07,
+      "loss": 0.2611,
+      "step": 9472
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 7.881570755136529e-07,
+      "loss": 0.2373,
+      "step": 9536
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 7.867339678021881e-07,
+      "loss": 0.2052,
+      "step": 9600
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 7.853108600907231e-07,
+      "loss": 0.1998,
+      "step": 9664
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 7.838877523792582e-07,
+      "loss": 0.2254,
+      "step": 9728
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 7.824646446677932e-07,
+      "loss": 0.1955,
+      "step": 9792
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 7.810415369563284e-07,
+      "loss": 0.1238,
+      "step": 9856
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 7.796184292448634e-07,
+      "loss": 0.1806,
+      "step": 9920
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 7.781953215333986e-07,
+      "loss": 0.31,
+      "step": 9984
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 7.767722138219336e-07,
+      "loss": 0.2349,
+      "step": 10048
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 7.753491061104687e-07,
+      "loss": 0.2277,
+      "step": 10112
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 7.739259983990037e-07,
+      "loss": 0.2239,
+      "step": 10176
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 7.725028906875389e-07,
+      "loss": 0.2707,
+      "step": 10240
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 7.71079782976074e-07,
+      "loss": 0.1957,
+      "step": 10304
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 7.69656675264609e-07,
+      "loss": 0.1649,
+      "step": 10368
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 7.682335675531442e-07,
+      "loss": 0.2773,
+      "step": 10432
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 7.668104598416792e-07,
+      "loss": 0.2146,
+      "step": 10496
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 7.653873521302144e-07,
+      "loss": 0.2134,
+      "step": 10560
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 7.639642444187494e-07,
+      "loss": 0.2168,
+      "step": 10624
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 7.625411367072845e-07,
+      "loss": 0.1122,
+      "step": 10688
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 7.611180289958195e-07,
+      "loss": 0.1628,
+      "step": 10752
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 7.596949212843547e-07,
+      "loss": 0.1566,
+      "step": 10816
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 7.582718135728897e-07,
+      "loss": 0.1824,
+      "step": 10880
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 7.56848705861425e-07,
+      "loss": 0.292,
+      "step": 10944
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 7.5542559814996e-07,
+      "loss": 0.1534,
+      "step": 11008
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 7.54002490438495e-07,
+      "loss": 0.2107,
+      "step": 11072
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 7.525793827270302e-07,
+      "loss": 0.1689,
+      "step": 11136
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 7.511562750155652e-07,
+      "loss": 0.3489,
+      "step": 11200
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 7.497331673041003e-07,
+      "loss": 0.1535,
+      "step": 11264
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 7.483100595926353e-07,
+      "loss": 0.1439,
+      "step": 11328
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 7.468869518811705e-07,
+      "loss": 0.2032,
+      "step": 11392
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 7.454638441697055e-07,
+      "loss": 0.4177,
+      "step": 11456
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 7.440407364582406e-07,
+      "loss": 0.2552,
+      "step": 11520
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 7.426176287467757e-07,
+      "loss": 0.181,
+      "step": 11584
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 7.411945210353108e-07,
+      "loss": 0.1128,
+      "step": 11648
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 7.39771413323846e-07,
+      "loss": 0.2184,
+      "step": 11712
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 7.383483056123811e-07,
+      "loss": 0.3284,
+      "step": 11776
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 7.369251979009161e-07,
+      "loss": 0.1976,
+      "step": 11840
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 7.355020901894511e-07,
+      "loss": 0.1274,
+      "step": 11904
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 7.340789824779863e-07,
+      "loss": 0.1578,
+      "step": 11968
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 7.326558747665213e-07,
+      "loss": 0.2599,
+      "step": 12032
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 7.312327670550564e-07,
+      "loss": 0.1812,
+      "step": 12096
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 7.298096593435915e-07,
+      "loss": 0.2296,
+      "step": 12160
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 7.283865516321266e-07,
+      "loss": 0.2716,
+      "step": 12224
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 7.269634439206616e-07,
+      "loss": 0.3587,
+      "step": 12288
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 7.255403362091968e-07,
+      "loss": 0.1644,
+      "step": 12352
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 7.241394645557235e-07,
+      "loss": 0.2707,
+      "step": 12416
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 7.227163568442586e-07,
+      "loss": 0.1747,
+      "step": 12480
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 7.212932491327936e-07,
+      "loss": 0.2109,
+      "step": 12544
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 7.198701414213289e-07,
+      "loss": 0.2764,
+      "step": 12608
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 7.184470337098639e-07,
+      "loss": 0.1857,
+      "step": 12672
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 7.170239259983991e-07,
+      "loss": 0.2993,
+      "step": 12736
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 7.156008182869341e-07,
+      "loss": 0.3019,
+      "step": 12800
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 7.141777105754691e-07,
+      "loss": 0.2171,
+      "step": 12864
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 7.127546028640042e-07,
+      "loss": 0.1616,
+      "step": 12928
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.113314951525393e-07,
+      "loss": 0.2056,
+      "step": 12992
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.099083874410744e-07,
+      "loss": 0.1741,
+      "step": 13056
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.084852797296094e-07,
+      "loss": 0.1881,
+      "step": 13120
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.070621720181446e-07,
+      "loss": 0.3325,
+      "step": 13184
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.056390643066796e-07,
+      "loss": 0.1797,
+      "step": 13248
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.042159565952149e-07,
+      "loss": 0.1904,
+      "step": 13312
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.027928488837499e-07,
+      "loss": 0.182,
+      "step": 13376
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.01369741172285e-07,
+      "loss": 0.1893,
+      "step": 13440
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 6.9994663346082e-07,
+      "loss": 0.2757,
+      "step": 13504
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 6.985235257493552e-07,
+      "loss": 0.1684,
+      "step": 13568
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.971004180378902e-07,
+      "loss": 0.1678,
+      "step": 13632
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.956773103264252e-07,
+      "loss": 0.2164,
+      "step": 13696
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.942542026149604e-07,
+      "loss": 0.3885,
+      "step": 13760
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.928310949034954e-07,
+      "loss": 0.153,
+      "step": 13824
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.914079871920305e-07,
+      "loss": 0.211,
+      "step": 13888
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.899848794805656e-07,
+      "loss": 0.2952,
+      "step": 13952
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.885617717691008e-07,
+      "loss": 0.227,
+      "step": 14016
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.871386640576358e-07,
+      "loss": 0.1933,
+      "step": 14080
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.85715556346171e-07,
+      "loss": 0.2457,
+      "step": 14144
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.84292448634706e-07,
+      "loss": 0.1388,
+      "step": 14208
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.828693409232412e-07,
+      "loss": 0.175,
+      "step": 14272
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.814462332117762e-07,
+      "loss": 0.1735,
+      "step": 14336
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.800231255003113e-07,
+      "loss": 0.2027,
+      "step": 14400
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.786000177888463e-07,
+      "loss": 0.3326,
+      "step": 14464
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 6.771769100773814e-07,
+      "loss": 0.2391,
+      "step": 14528
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 6.757538023659165e-07,
+      "loss": 0.1766,
+      "step": 14592
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 6.743306946544515e-07,
+      "loss": 0.195,
+      "step": 14656
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 6.729075869429868e-07,
+      "loss": 0.2176,
+      "step": 14720
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 6.714844792315218e-07,
+      "loss": 0.2863,
+      "step": 14784
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 6.700836075780485e-07,
+      "loss": 0.2664,
+      "step": 14848
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 6.686604998665835e-07,
+      "loss": 0.1482,
+      "step": 14912
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 6.672373921551188e-07,
+      "loss": 0.3435,
+      "step": 14976
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 6.658142844436538e-07,
+      "loss": 0.2416,
+      "step": 15040
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 6.64391176732189e-07,
+      "loss": 0.2352,
+      "step": 15104
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 6.62968069020724e-07,
+      "loss": 0.2318,
+      "step": 15168
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 6.615449613092591e-07,
+      "loss": 0.2973,
+      "step": 15232
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 6.601218535977941e-07,
+      "loss": 0.177,
+      "step": 15296
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 6.586987458863293e-07,
+      "loss": 0.2529,
+      "step": 15360
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 6.572756381748643e-07,
+      "loss": 0.1659,
+      "step": 15424
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 6.558747665213911e-07,
+      "loss": 0.1879,
+      "step": 15488
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 6.544516588099261e-07,
+      "loss": 0.296,
+      "step": 15552
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 6.530285510984612e-07,
+      "loss": 0.2029,
+      "step": 15616
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 6.516054433869963e-07,
+      "loss": 0.1833,
+      "step": 15680
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 6.501823356755313e-07,
+      "loss": 0.2408,
+      "step": 15744
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 6.487592279640665e-07,
+      "loss": 0.2729,
+      "step": 15808
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 6.473361202526015e-07,
+      "loss": 0.1371,
+      "step": 15872
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 6.459130125411368e-07,
+      "loss": 0.2234,
+      "step": 15936
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 6.444899048296718e-07,
+      "loss": 0.2095,
+      "step": 16000
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 6.430667971182069e-07,
+      "loss": 0.2612,
+      "step": 16064
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 6.41643689406742e-07,
+      "loss": 0.3269,
+      "step": 16128
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 6.402205816952771e-07,
+      "loss": 0.1677,
+      "step": 16192
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 6.387974739838121e-07,
+      "loss": 0.1625,
+      "step": 16256
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 6.373743662723472e-07,
+      "loss": 0.1675,
+      "step": 16320
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 6.359512585608823e-07,
+      "loss": 0.1452,
+      "step": 16384
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 6.345281508494173e-07,
+      "loss": 0.1458,
+      "step": 16448
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 6.331050431379524e-07,
+      "loss": 0.149,
+      "step": 16512
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 6.316819354264875e-07,
+      "loss": 0.329,
+      "step": 16576
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 6.302588277150227e-07,
+      "loss": 0.1425,
+      "step": 16640
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 6.288357200035577e-07,
+      "loss": 0.2696,
+      "step": 16704
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 6.274126122920929e-07,
+      "loss": 0.1641,
+      "step": 16768
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 6.259895045806279e-07,
+      "loss": 0.1661,
+      "step": 16832
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 6.24566396869163e-07,
+      "loss": 0.1707,
+      "step": 16896
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 6.231432891576981e-07,
+      "loss": 0.2579,
+      "step": 16960
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 6.217201814462332e-07,
+      "loss": 0.2654,
+      "step": 17024
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 6.202970737347682e-07,
+      "loss": 0.2443,
+      "step": 17088
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 6.188739660233034e-07,
+      "loss": 0.1523,
+      "step": 17152
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 6.174508583118384e-07,
+      "loss": 0.1256,
+      "step": 17216
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 6.160277506003734e-07,
+      "loss": 0.2002,
+      "step": 17280
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 6.146046428889087e-07,
+      "loss": 0.225,
+      "step": 17344
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 6.131815351774437e-07,
+      "loss": 0.2168,
+      "step": 17408
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 6.117584274659788e-07,
+      "loss": 0.1444,
+      "step": 17472
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 6.103353197545139e-07,
+      "loss": 0.1429,
+      "step": 17536
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 6.08912212043049e-07,
+      "loss": 0.1688,
+      "step": 17600
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 6.07489104331584e-07,
+      "loss": 0.2285,
+      "step": 17664
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 6.060659966201192e-07,
+      "loss": 0.2193,
+      "step": 17728
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 6.046428889086542e-07,
+      "loss": 0.1757,
+      "step": 17792
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 6.032197811971893e-07,
+      "loss": 0.2121,
+      "step": 17856
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 6.017966734857244e-07,
+      "loss": 0.2231,
+      "step": 17920
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 6.003735657742595e-07,
+      "loss": 0.213,
+      "step": 17984
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 5.989504580627946e-07,
+      "loss": 0.1744,
+      "step": 18048
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 5.975273503513297e-07,
+      "loss": 0.2606,
+      "step": 18112
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 5.961042426398648e-07,
+      "loss": 0.2494,
+      "step": 18176
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 5.946811349283998e-07,
+      "loss": 0.2053,
+      "step": 18240
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 5.93258027216935e-07,
+      "loss": 0.2282,
+      "step": 18304
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 5.9183491950547e-07,
+      "loss": 0.2375,
+      "step": 18368
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 5.904118117940051e-07,
+      "loss": 0.1211,
+      "step": 18432
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 5.889887040825402e-07,
+      "loss": 0.2294,
+      "step": 18496
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 5.875655963710753e-07,
+      "loss": 0.2236,
+      "step": 18560
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 5.861424886596103e-07,
+      "loss": 0.1577,
+      "step": 18624
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 5.847193809481455e-07,
+      "loss": 0.2457,
+      "step": 18688
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 5.832962732366806e-07,
+      "loss": 0.1634,
+      "step": 18752
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 5.818731655252158e-07,
+      "loss": 0.2724,
+      "step": 18816
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 5.804500578137508e-07,
+      "loss": 0.2411,
+      "step": 18880
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 5.790269501022858e-07,
+      "loss": 0.2015,
+      "step": 18944
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 5.776038423908209e-07,
+      "loss": 0.1731,
+      "step": 19008
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 5.76180734679356e-07,
+      "loss": 0.2573,
+      "step": 19072
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 5.747576269678911e-07,
+      "loss": 0.1844,
+      "step": 19136
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 5.733345192564261e-07,
+      "loss": 0.2007,
+      "step": 19200
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 5.719114115449613e-07,
+      "loss": 0.362,
+      "step": 19264
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 5.704883038334963e-07,
+      "loss": 0.2429,
+      "step": 19328
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 5.690651961220314e-07,
+      "loss": 0.2387,
+      "step": 19392
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 5.676420884105666e-07,
+      "loss": 0.1515,
+      "step": 19456
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 5.662189806991017e-07,
+      "loss": 0.0871,
+      "step": 19520
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 5.647958729876367e-07,
+      "loss": 0.2069,
+      "step": 19584
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 5.633727652761719e-07,
+      "loss": 0.2809,
+      "step": 19648
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 5.619496575647069e-07,
+      "loss": 0.1971,
+      "step": 19712
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 5.605265498532419e-07,
+      "loss": 0.2416,
+      "step": 19776
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 5.591034421417771e-07,
+      "loss": 0.2536,
+      "step": 19840
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 5.576803344303121e-07,
+      "loss": 0.199,
+      "step": 19904
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 5.562572267188472e-07,
+      "loss": 0.1612,
+      "step": 19968
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 5.548341190073823e-07,
+      "loss": 0.2325,
+      "step": 20032
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 5.534110112959174e-07,
+      "loss": 0.2497,
+      "step": 20096
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.519879035844525e-07,
+      "loss": 0.3172,
+      "step": 20160
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.505647958729877e-07,
+      "loss": 0.2571,
+      "step": 20224
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.491416881615227e-07,
+      "loss": 0.1585,
+      "step": 20288
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 5.477185804500578e-07,
+      "loss": 0.2229,
+      "step": 20352
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 5.462954727385929e-07,
+      "loss": 0.2502,
+      "step": 20416
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 5.44872365027128e-07,
+      "loss": 0.2151,
+      "step": 20480
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 5.43449257315663e-07,
+      "loss": 0.1644,
+      "step": 20544
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 5.420261496041981e-07,
+      "loss": 0.3373,
+      "step": 20608
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 5.406030418927332e-07,
+      "loss": 0.1322,
+      "step": 20672
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 5.391799341812682e-07,
+      "loss": 0.213,
+      "step": 20736
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 5.377568264698034e-07,
+      "loss": 0.169,
+      "step": 20800
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 5.363337187583385e-07,
+      "loss": 0.2652,
+      "step": 20864
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 5.349106110468736e-07,
+      "loss": 0.2234,
+      "step": 20928
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 5.334875033354087e-07,
+      "loss": 0.1933,
+      "step": 20992
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 5.320866316819354e-07,
+      "loss": 0.2141,
+      "step": 21056
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 5.306635239704705e-07,
+      "loss": 0.217,
+      "step": 21120
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 5.292404162590056e-07,
+      "loss": 0.2091,
+      "step": 21184
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 5.278173085475407e-07,
+      "loss": 0.2001,
+      "step": 21248
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 5.263942008360758e-07,
+      "loss": 0.1489,
+      "step": 21312
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 5.249710931246108e-07,
+      "loss": 0.2728,
+      "step": 21376
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 5.23547985413146e-07,
+      "loss": 0.1827,
+      "step": 21440
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 5.22124877701681e-07,
+      "loss": 0.1933,
+      "step": 21504
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 5.20701769990216e-07,
+      "loss": 0.1068,
+      "step": 21568
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 5.192786622787512e-07,
+      "loss": 0.2462,
+      "step": 21632
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 5.178555545672862e-07,
+      "loss": 0.3433,
+      "step": 21696
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 5.164324468558213e-07,
+      "loss": 0.2489,
+      "step": 21760
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 5.150093391443565e-07,
+      "loss": 0.2038,
+      "step": 21824
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 5.135862314328916e-07,
+      "loss": 0.2269,
+      "step": 21888
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 5.121631237214266e-07,
+      "loss": 0.1911,
+      "step": 21952
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 5.107400160099618e-07,
+      "loss": 0.2203,
+      "step": 22016
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 5.093169082984968e-07,
+      "loss": 0.2476,
+      "step": 22080
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 5.078938005870319e-07,
+      "loss": 0.2525,
+      "step": 22144
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 5.06470692875567e-07,
+      "loss": 0.1442,
+      "step": 22208
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 5.050475851641021e-07,
+      "loss": 0.1759,
+      "step": 22272
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 5.036244774526371e-07,
+      "loss": 0.2944,
+      "step": 22336
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 5.022013697411722e-07,
+      "loss": 0.1416,
+      "step": 22400
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 5.007782620297073e-07,
+      "loss": 0.2061,
+      "step": 22464
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.993551543182424e-07,
+      "loss": 0.2197,
+      "step": 22528
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.979320466067775e-07,
+      "loss": 0.1192,
+      "step": 22592
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 4.965089388953126e-07,
+      "loss": 0.1117,
+      "step": 22656
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 4.950858311838477e-07,
+      "loss": 0.1859,
+      "step": 22720
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 4.936627234723828e-07,
+      "loss": 0.1986,
+      "step": 22784
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 4.922396157609179e-07,
+      "loss": 0.2719,
+      "step": 22848
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 4.908165080494529e-07,
+      "loss": 0.1638,
+      "step": 22912
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 4.893934003379881e-07,
+      "loss": 0.0792,
+      "step": 22976
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 4.879702926265232e-07,
+      "loss": 0.1522,
+      "step": 23040
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 4.865471849150582e-07,
+      "loss": 0.1592,
+      "step": 23104
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 4.851240772035934e-07,
+      "loss": 0.1597,
+      "step": 23168
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 4.837009694921284e-07,
+      "loss": 0.1055,
+      "step": 23232
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 4.822778617806634e-07,
+      "loss": 0.1207,
+      "step": 23296
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 4.808547540691986e-07,
+      "loss": 0.1717,
+      "step": 23360
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 4.794538824157253e-07,
+      "loss": 0.1305,
+      "step": 23424
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 4.780307747042604e-07,
+      "loss": 0.0869,
+      "step": 23488
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 4.766076669927955e-07,
+      "loss": 0.1097,
+      "step": 23552
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 4.7518455928133063e-07,
+      "loss": 0.2356,
+      "step": 23616
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 4.737614515698657e-07,
+      "loss": 0.1326,
+      "step": 23680
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 4.7233834385840074e-07,
+      "loss": 0.1271,
+      "step": 23744
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 4.709152361469358e-07,
+      "loss": 0.1454,
+      "step": 23808
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 4.694921284354709e-07,
+      "loss": 0.1813,
+      "step": 23872
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 4.68069020724006e-07,
+      "loss": 0.1332,
+      "step": 23936
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 4.6664591301254113e-07,
+      "loss": 0.1769,
+      "step": 24000
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 4.652228053010762e-07,
+      "loss": 0.1574,
+      "step": 24064
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 4.637996975896113e-07,
+      "loss": 0.1397,
+      "step": 24128
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 4.623765898781464e-07,
+      "loss": 0.2307,
+      "step": 24192
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 4.6095348216668146e-07,
+      "loss": 0.0829,
+      "step": 24256
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 4.595303744552166e-07,
+      "loss": 0.1801,
+      "step": 24320
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 4.581072667437517e-07,
+      "loss": 0.1314,
+      "step": 24384
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 4.5668415903228676e-07,
+      "loss": 0.1002,
+      "step": 24448
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 4.5526105132082184e-07,
+      "loss": 0.1372,
+      "step": 24512
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 4.538379436093569e-07,
+      "loss": 0.1843,
+      "step": 24576
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 4.5241483589789196e-07,
+      "loss": 0.1166,
+      "step": 24640
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 4.509917281864271e-07,
+      "loss": 0.1739,
+      "step": 24704
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 4.495686204749622e-07,
+      "loss": 0.1977,
+      "step": 24768
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 4.4814551276349726e-07,
+      "loss": 0.2044,
+      "step": 24832
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 4.4672240505203234e-07,
+      "loss": 0.1625,
+      "step": 24896
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 4.452992973405674e-07,
+      "loss": 0.1712,
+      "step": 24960
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 4.4387618962910256e-07,
+      "loss": 0.0945,
+      "step": 25024
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 4.4245308191763764e-07,
+      "loss": 0.0845,
+      "step": 25088
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 4.4102997420617273e-07,
+      "loss": 0.1123,
+      "step": 25152
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 4.396068664947078e-07,
+      "loss": 0.1598,
+      "step": 25216
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 4.381837587832429e-07,
+      "loss": 0.173,
+      "step": 25280
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 4.36760651071778e-07,
+      "loss": 0.159,
+      "step": 25344
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 4.353375433603131e-07,
+      "loss": 0.1855,
+      "step": 25408
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 4.3391443564884814e-07,
+      "loss": 0.1274,
+      "step": 25472
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 4.324913279373832e-07,
+      "loss": 0.1939,
+      "step": 25536
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 4.310682202259183e-07,
+      "loss": 0.1458,
+      "step": 25600
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 4.296451125144534e-07,
+      "loss": 0.0993,
+      "step": 25664
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 4.282220048029885e-07,
+      "loss": 0.1084,
+      "step": 25728
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 4.267988970915236e-07,
+      "loss": 0.092,
+      "step": 25792
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 4.253757893800587e-07,
+      "loss": 0.1503,
+      "step": 25856
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 4.239526816685938e-07,
+      "loss": 0.144,
+      "step": 25920
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 4.2252957395712886e-07,
+      "loss": 0.075,
+      "step": 25984
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 4.2110646624566394e-07,
+      "loss": 0.1614,
+      "step": 26048
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 4.196833585341991e-07,
+      "loss": 0.1753,
+      "step": 26112
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 4.1826025082273416e-07,
+      "loss": 0.2336,
+      "step": 26176
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 4.1683714311126924e-07,
+      "loss": 0.1453,
+      "step": 26240
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 4.154140353998043e-07,
+      "loss": 0.1707,
+      "step": 26304
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 4.1399092768833936e-07,
+      "loss": 0.1123,
+      "step": 26368
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 4.1256781997687444e-07,
+      "loss": 0.1998,
+      "step": 26432
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 4.111447122654096e-07,
+      "loss": 0.1178,
+      "step": 26496
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 4.0972160455394466e-07,
+      "loss": 0.1781,
+      "step": 26560
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 4.0829849684247974e-07,
+      "loss": 0.0921,
+      "step": 26624
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 4.068753891310148e-07,
+      "loss": 0.0661,
+      "step": 26688
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 4.054522814195499e-07,
+      "loss": 0.1927,
+      "step": 26752
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 4.0402917370808504e-07,
+      "loss": 0.1816,
+      "step": 26816
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 4.0262830205461174e-07,
+      "loss": 0.2152,
+      "step": 26880
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 4.012051943431468e-07,
+      "loss": 0.181,
+      "step": 26944
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 3.997820866316819e-07,
+      "loss": 0.1332,
+      "step": 27008
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 3.9835897892021704e-07,
+      "loss": 0.077,
+      "step": 27072
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 3.9693587120875213e-07,
+      "loss": 0.1626,
+      "step": 27136
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 3.9551276349728716e-07,
+      "loss": 0.1894,
+      "step": 27200
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 3.9408965578582224e-07,
+      "loss": 0.1509,
+      "step": 27264
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 3.926665480743573e-07,
+      "loss": 0.0634,
+      "step": 27328
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 3.9124344036289246e-07,
+      "loss": 0.1287,
+      "step": 27392
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 3.8982033265142754e-07,
+      "loss": 0.0669,
+      "step": 27456
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 3.883972249399626e-07,
+      "loss": 0.1046,
+      "step": 27520
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 3.869741172284977e-07,
+      "loss": 0.0747,
+      "step": 27584
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 3.855510095170328e-07,
+      "loss": 0.1554,
+      "step": 27648
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 3.841279018055679e-07,
+      "loss": 0.1803,
+      "step": 27712
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 3.82704794094103e-07,
+      "loss": 0.1479,
+      "step": 27776
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 3.812816863826381e-07,
+      "loss": 0.1667,
+      "step": 27840
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 3.798585786711732e-07,
+      "loss": 0.2049,
+      "step": 27904
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 3.7843547095970826e-07,
+      "loss": 0.1617,
+      "step": 27968
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 3.770123632482433e-07,
+      "loss": 0.0975,
+      "step": 28032
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 3.755892555367784e-07,
+      "loss": 0.101,
+      "step": 28096
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 3.741661478253135e-07,
+      "loss": 0.1224,
+      "step": 28160
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 3.727430401138486e-07,
+      "loss": 0.1519,
+      "step": 28224
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 3.713199324023837e-07,
+      "loss": 0.1175,
+      "step": 28288
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 3.6989682469091876e-07,
+      "loss": 0.2126,
+      "step": 28352
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 3.6847371697945384e-07,
+      "loss": 0.1332,
+      "step": 28416
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 3.67050609267989e-07,
+      "loss": 0.1906,
+      "step": 28480
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 3.6562750155652406e-07,
+      "loss": 0.1032,
+      "step": 28544
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 3.6420439384505914e-07,
+      "loss": 0.096,
+      "step": 28608
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 3.627812861335942e-07,
+      "loss": 0.1452,
+      "step": 28672
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 3.613581784221293e-07,
+      "loss": 0.1455,
+      "step": 28736
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 3.5993507071066444e-07,
+      "loss": 0.103,
+      "step": 28800
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 3.5851196299919953e-07,
+      "loss": 0.1293,
+      "step": 28864
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 3.5708885528773456e-07,
+      "loss": 0.1759,
+      "step": 28928
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 3.5566574757626964e-07,
+      "loss": 0.1644,
+      "step": 28992
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 3.542426398648047e-07,
+      "loss": 0.0729,
+      "step": 29056
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 3.528195321533398e-07,
+      "loss": 0.1703,
+      "step": 29120
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 3.5139642444187494e-07,
+      "loss": 0.1259,
+      "step": 29184
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 3.4997331673041e-07,
+      "loss": 0.121,
+      "step": 29248
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 3.485502090189451e-07,
+      "loss": 0.1117,
+      "step": 29312
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 3.471271013074802e-07,
+      "loss": 0.1015,
+      "step": 29376
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 3.457039935960153e-07,
+      "loss": 0.1409,
+      "step": 29440
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 3.442808858845504e-07,
+      "loss": 0.1358,
+      "step": 29504
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 3.428577781730855e-07,
+      "loss": 0.1704,
+      "step": 29568
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 3.414346704616206e-07,
+      "loss": 0.2159,
+      "step": 29632
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 3.4001156275015566e-07,
+      "loss": 0.116,
+      "step": 29696
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 3.385884550386907e-07,
+      "loss": 0.2161,
+      "step": 29760
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 3.3716534732722577e-07,
+      "loss": 0.1037,
+      "step": 29824
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 3.357422396157609e-07,
+      "loss": 0.1729,
+      "step": 29888
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 3.34319131904296e-07,
+      "loss": 0.151,
+      "step": 29952
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 3.3289602419283107e-07,
+      "loss": 0.1541,
+      "step": 30016
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 3.3147291648136616e-07,
+      "loss": 0.1591,
+      "step": 30080
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 3.3004980876990124e-07,
+      "loss": 0.1226,
+      "step": 30144
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 3.286267010584364e-07,
+      "loss": 0.0741,
+      "step": 30208
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 3.2720359334697146e-07,
+      "loss": 0.0974,
+      "step": 30272
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 3.2578048563550654e-07,
+      "loss": 0.15,
+      "step": 30336
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 3.243573779240416e-07,
+      "loss": 0.1683,
+      "step": 30400
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 3.229342702125767e-07,
+      "loss": 0.1149,
+      "step": 30464
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 3.215111625011118e-07,
+      "loss": 0.1928,
+      "step": 30528
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 3.2008805478964687e-07,
+      "loss": 0.0922,
+      "step": 30592
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 3.1866494707818196e-07,
+      "loss": 0.0845,
+      "step": 30656
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 3.1724183936671704e-07,
+      "loss": 0.1364,
+      "step": 30720
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 3.158187316552521e-07,
+      "loss": 0.1298,
+      "step": 30784
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 3.143956239437872e-07,
+      "loss": 0.2509,
+      "step": 30848
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 3.129725162323223e-07,
+      "loss": 0.2089,
+      "step": 30912
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 3.115494085208574e-07,
+      "loss": 0.1076,
+      "step": 30976
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 3.101263008093925e-07,
+      "loss": 0.0883,
+      "step": 31040
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 3.087254291559192e-07,
+      "loss": 0.1692,
+      "step": 31104
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 3.0730232144445434e-07,
+      "loss": 0.1239,
+      "step": 31168
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 3.058792137329894e-07,
+      "loss": 0.0783,
+      "step": 31232
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 3.044561060215245e-07,
+      "loss": 0.1541,
+      "step": 31296
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 3.030329983100596e-07,
+      "loss": 0.1009,
+      "step": 31360
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 3.0160989059859467e-07,
+      "loss": 0.158,
+      "step": 31424
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 3.0018678288712976e-07,
+      "loss": 0.101,
+      "step": 31488
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 2.9876367517566484e-07,
+      "loss": 0.1887,
+      "step": 31552
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 2.973405674641999e-07,
+      "loss": 0.1327,
+      "step": 31616
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 2.95917459752735e-07,
+      "loss": 0.1198,
+      "step": 31680
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 2.944943520412701e-07,
+      "loss": 0.096,
+      "step": 31744
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 2.9307124432980517e-07,
+      "loss": 0.1474,
+      "step": 31808
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 2.916481366183403e-07,
+      "loss": 0.1221,
+      "step": 31872
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 2.902250289068754e-07,
+      "loss": 0.1119,
+      "step": 31936
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 2.8880192119541047e-07,
+      "loss": 0.1665,
+      "step": 32000
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 2.8737881348394556e-07,
+      "loss": 0.0913,
+      "step": 32064
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 2.8595570577248064e-07,
+      "loss": 0.1524,
+      "step": 32128
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 2.845325980610157e-07,
+      "loss": 0.1798,
+      "step": 32192
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 2.8310949034955086e-07,
+      "loss": 0.1023,
+      "step": 32256
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 2.8168638263808594e-07,
+      "loss": 0.1718,
+      "step": 32320
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 2.8026327492662097e-07,
+      "loss": 0.1651,
+      "step": 32384
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 2.788624032731478e-07,
+      "loss": 0.1502,
+      "step": 32448
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 2.774392955616828e-07,
+      "loss": 0.1074,
+      "step": 32512
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 2.760161878502179e-07,
+      "loss": 0.0905,
+      "step": 32576
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 2.7459308013875297e-07,
+      "loss": 0.1444,
+      "step": 32640
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 2.7316997242728805e-07,
+      "loss": 0.1053,
+      "step": 32704
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 2.7174686471582314e-07,
+      "loss": 0.0763,
+      "step": 32768
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 2.7032375700435827e-07,
+      "loss": 0.1183,
+      "step": 32832
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 2.6890064929289336e-07,
+      "loss": 0.1601,
+      "step": 32896
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 2.6747754158142844e-07,
+      "loss": 0.1325,
+      "step": 32960
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 2.660544338699635e-07,
+      "loss": 0.1957,
+      "step": 33024
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 2.646313261584986e-07,
+      "loss": 0.176,
+      "step": 33088
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 2.6320821844703374e-07,
+      "loss": 0.0789,
+      "step": 33152
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 2.617851107355688e-07,
+      "loss": 0.1672,
+      "step": 33216
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 2.603620030241039e-07,
+      "loss": 0.1331,
+      "step": 33280
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 2.5893889531263894e-07,
+      "loss": 0.1082,
+      "step": 33344
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 2.57515787601174e-07,
+      "loss": 0.0921,
+      "step": 33408
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 2.560926798897091e-07,
+      "loss": 0.086,
+      "step": 33472
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 2.5466957217824424e-07,
+      "loss": 0.1767,
+      "step": 33536
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 2.532464644667793e-07,
+      "loss": 0.2088,
+      "step": 33600
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 2.518233567553144e-07,
+      "loss": 0.1928,
+      "step": 33664
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 2.504002490438495e-07,
+      "loss": 0.176,
+      "step": 33728
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 2.4897714133238457e-07,
+      "loss": 0.1787,
+      "step": 33792
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 2.4755403362091965e-07,
+      "loss": 0.1343,
+      "step": 33856
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 2.4613092590945474e-07,
+      "loss": 0.1454,
+      "step": 33920
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 2.4470781819798987e-07,
+      "loss": 0.1752,
+      "step": 33984
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 2.4328471048652496e-07,
+      "loss": 0.1269,
+      "step": 34048
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.4186160277506004e-07,
+      "loss": 0.1861,
+      "step": 34112
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.404384950635951e-07,
+      "loss": 0.0963,
+      "step": 34176
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.390153873521302e-07,
+      "loss": 0.1724,
+      "step": 34240
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.3759227964066531e-07,
+      "loss": 0.2089,
+      "step": 34304
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.3616917192920037e-07,
+      "loss": 0.0815,
+      "step": 34368
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.3474606421773545e-07,
+      "loss": 0.1483,
+      "step": 34432
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.3332295650627056e-07,
+      "loss": 0.1122,
+      "step": 34496
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 2.3189984879480565e-07,
+      "loss": 0.0886,
+      "step": 34560
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 2.3047674108334073e-07,
+      "loss": 0.1228,
+      "step": 34624
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 2.2905363337187584e-07,
+      "loss": 0.113,
+      "step": 34688
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.2763052566041092e-07,
+      "loss": 0.115,
+      "step": 34752
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.2620741794894598e-07,
+      "loss": 0.23,
+      "step": 34816
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.247843102374811e-07,
+      "loss": 0.2181,
+      "step": 34880
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.2336120252601617e-07,
+      "loss": 0.145,
+      "step": 34944
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.2193809481455128e-07,
+      "loss": 0.099,
+      "step": 35008
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.2051498710308636e-07,
+      "loss": 0.1521,
+      "step": 35072
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.1909187939162145e-07,
+      "loss": 0.1015,
+      "step": 35136
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.1766877168015656e-07,
+      "loss": 0.1484,
+      "step": 35200
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.162456639686916e-07,
+      "loss": 0.1857,
+      "step": 35264
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.148225562572267e-07,
+      "loss": 0.1999,
+      "step": 35328
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.133994485457618e-07,
+      "loss": 0.2292,
+      "step": 35392
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 2.119763408342969e-07,
+      "loss": 0.1004,
+      "step": 35456
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 2.1055323312283197e-07,
+      "loss": 0.1026,
+      "step": 35520
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 2.0913012541136708e-07,
+      "loss": 0.177,
+      "step": 35584
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 2.0770701769990214e-07,
+      "loss": 0.1872,
+      "step": 35648
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 2.0628390998843722e-07,
+      "loss": 0.1365,
+      "step": 35712
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 2.0486080227697233e-07,
+      "loss": 0.2041,
+      "step": 35776
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 2.034376945655074e-07,
+      "loss": 0.1495,
+      "step": 35840
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 2.0201458685404252e-07,
+      "loss": 0.1752,
+      "step": 35904
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 2.005914791425776e-07,
+      "loss": 0.1263,
+      "step": 35968
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 1.991683714311127e-07,
+      "loss": 0.13,
+      "step": 36032
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 1.9774526371964777e-07,
+      "loss": 0.1357,
+      "step": 36096
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 1.9632215600818285e-07,
+      "loss": 0.1495,
+      "step": 36160
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 1.9489904829671794e-07,
+      "loss": 0.1125,
+      "step": 36224
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 1.9347594058525305e-07,
+      "loss": 0.1569,
+      "step": 36288
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 1.9205283287378813e-07,
+      "loss": 0.1422,
+      "step": 36352
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 1.9065196122031485e-07,
+      "loss": 0.0862,
+      "step": 36416
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 1.8922885350884996e-07,
+      "loss": 0.1289,
+      "step": 36480
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 1.8780574579738505e-07,
+      "loss": 0.2662,
+      "step": 36544
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 1.863826380859201e-07,
+      "loss": 0.1815,
+      "step": 36608
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 1.849595303744552e-07,
+      "loss": 0.1208,
+      "step": 36672
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 1.835364226629903e-07,
+      "loss": 0.189,
+      "step": 36736
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 1.8211331495152538e-07,
+      "loss": 0.1249,
+      "step": 36800
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 1.806902072400605e-07,
+      "loss": 0.1754,
+      "step": 36864
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 1.7926709952859557e-07,
+      "loss": 0.0879,
+      "step": 36928
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 1.7784399181713063e-07,
+      "loss": 0.1246,
+      "step": 36992
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 1.7642088410566574e-07,
+      "loss": 0.1831,
+      "step": 37056
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 1.7499777639420082e-07,
+      "loss": 0.0834,
+      "step": 37120
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 1.735746686827359e-07,
+      "loss": 0.1259,
+      "step": 37184
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 1.72151560971271e-07,
+      "loss": 0.0903,
+      "step": 37248
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 1.707284532598061e-07,
+      "loss": 0.1556,
+      "step": 37312
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 1.693053455483412e-07,
+      "loss": 0.1117,
+      "step": 37376
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 1.6788223783687626e-07,
+      "loss": 0.1567,
+      "step": 37440
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 1.6645913012541134e-07,
+      "loss": 0.1106,
+      "step": 37504
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 1.6503602241394645e-07,
+      "loss": 0.0943,
+      "step": 37568
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 1.6361291470248154e-07,
+      "loss": 0.1414,
+      "step": 37632
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 1.6218980699101662e-07,
+      "loss": 0.2007,
+      "step": 37696
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 1.6076669927955173e-07,
+      "loss": 0.0953,
+      "step": 37760
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 1.593435915680868e-07,
+      "loss": 0.1475,
+      "step": 37824
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 1.5792048385662187e-07,
+      "loss": 0.1893,
+      "step": 37888
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 1.5649737614515698e-07,
+      "loss": 0.0747,
+      "step": 37952
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 1.5507426843369206e-07,
+      "loss": 0.1341,
+      "step": 38016
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 1.5365116072222717e-07,
+      "loss": 0.168,
+      "step": 38080
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 1.5222805301076225e-07,
+      "loss": 0.1653,
+      "step": 38144
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 1.5080494529929734e-07,
+      "loss": 0.1036,
+      "step": 38208
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 1.4938183758783242e-07,
+      "loss": 0.1672,
+      "step": 38272
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 1.479587298763675e-07,
+      "loss": 0.0958,
+      "step": 38336
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 1.4653562216490259e-07,
+      "loss": 0.0857,
+      "step": 38400
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 1.451125144534377e-07,
+      "loss": 0.0764,
+      "step": 38464
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 1.4371164279996442e-07,
+      "loss": 0.1594,
+      "step": 38528
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 1.422885350884995e-07,
+      "loss": 0.1631,
+      "step": 38592
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 1.4086542737703459e-07,
+      "loss": 0.1066,
+      "step": 38656
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 1.394423196655697e-07,
+      "loss": 0.1881,
+      "step": 38720
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 1.3801921195410475e-07,
+      "loss": 0.1482,
+      "step": 38784
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 1.3659610424263986e-07,
+      "loss": 0.1915,
+      "step": 38848
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 1.3517299653117494e-07,
+      "loss": 0.0681,
+      "step": 38912
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 1.3374988881971003e-07,
+      "loss": 0.1975,
+      "step": 38976
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 1.3232678110824514e-07,
+      "loss": 0.1903,
+      "step": 39040
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 1.3090367339678022e-07,
+      "loss": 0.0811,
+      "step": 39104
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 1.294805656853153e-07,
+      "loss": 0.1073,
+      "step": 39168
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 1.2805745797385039e-07,
+      "loss": 0.1046,
+      "step": 39232
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 1.2663435026238547e-07,
+      "loss": 0.0984,
+      "step": 39296
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 1.2521124255092055e-07,
+      "loss": 0.1307,
+      "step": 39360
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 1.2378813483945566e-07,
+      "loss": 0.1867,
+      "step": 39424
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 1.2236502712799074e-07,
+      "loss": 0.0528,
+      "step": 39488
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 1.2094191941652583e-07,
+      "loss": 0.1276,
+      "step": 39552
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 1.195188117050609e-07,
+      "loss": 0.1381,
+      "step": 39616
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 1.1809570399359602e-07,
+      "loss": 0.1219,
+      "step": 39680
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 1.1667259628213109e-07,
+      "loss": 0.1109,
+      "step": 39744
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 1.1524948857066619e-07,
+      "loss": 0.0977,
+      "step": 39808
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 1.1382638085920128e-07,
+      "loss": 0.1619,
+      "step": 39872
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 1.1240327314773635e-07,
+      "loss": 0.1579,
+      "step": 39936
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 1.1098016543627145e-07,
+      "loss": 0.0653,
+      "step": 40000
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 1.0955705772480654e-07,
+      "loss": 0.1466,
+      "step": 40064
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 1.0813395001334164e-07,
+      "loss": 0.1393,
+      "step": 40128
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 1.0671084230187671e-07,
+      "loss": 0.1423,
+      "step": 40192
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 1.052877345904118e-07,
+      "loss": 0.134,
+      "step": 40256
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 1.038646268789469e-07,
+      "loss": 0.0889,
+      "step": 40320
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 1.0244151916748199e-07,
+      "loss": 0.2005,
+      "step": 40384
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 1.0101841145601707e-07,
+      "loss": 0.1085,
+      "step": 40448
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 9.959530374455216e-08,
+      "loss": 0.1331,
+      "step": 40512
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 9.819443209107889e-08,
+      "loss": 0.1805,
+      "step": 40576
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 9.677132437961399e-08,
+      "loss": 0.1644,
+      "step": 40640
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 9.534821666814907e-08,
+      "loss": 0.0803,
+      "step": 40704
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 9.392510895668415e-08,
+      "loss": 0.1879,
+      "step": 40768
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 9.250200124521925e-08,
+      "loss": 0.0765,
+      "step": 40832
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 9.110112959174597e-08,
+      "loss": 0.2374,
+      "step": 40896
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 8.967802188028107e-08,
+      "loss": 0.1095,
+      "step": 40960
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 8.825491416881614e-08,
+      "loss": 0.1828,
+      "step": 41024
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 8.683180645735124e-08,
+      "loss": 0.1268,
+      "step": 41088
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 8.540869874588633e-08,
+      "loss": 0.1552,
+      "step": 41152
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 8.398559103442141e-08,
+      "loss": 0.1431,
+      "step": 41216
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 8.25624833229565e-08,
+      "loss": 0.1276,
+      "step": 41280
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 8.11393756114916e-08,
+      "loss": 0.0916,
+      "step": 41344
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 7.971626790002668e-08,
+      "loss": 0.1484,
+      "step": 41408
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 7.829316018856177e-08,
+      "loss": 0.2089,
+      "step": 41472
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 7.687005247709686e-08,
+      "loss": 0.1541,
+      "step": 41536
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 7.544694476563194e-08,
+      "loss": 0.1452,
+      "step": 41600
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 7.402383705416703e-08,
+      "loss": 0.1424,
+      "step": 41664
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 7.260072934270212e-08,
+      "loss": 0.0817,
+      "step": 41728
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 7.117762163123721e-08,
+      "loss": 0.1659,
+      "step": 41792
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.97545139197723e-08,
+      "loss": 0.0918,
+      "step": 41856
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.83314062083074e-08,
+      "loss": 0.0709,
+      "step": 41920
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.690829849684248e-08,
+      "loss": 0.1059,
+      "step": 41984
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.548519078537756e-08,
+      "loss": 0.1332,
+      "step": 42048
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.406208307391266e-08,
+      "loss": 0.0878,
+      "step": 42112
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.263897536244775e-08,
+      "loss": 0.2336,
+      "step": 42176
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.121586765098282e-08,
+      "loss": 0.0781,
+      "step": 42240
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 5.979275993951792e-08,
+      "loss": 0.1154,
+      "step": 42304
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 5.836965222805301e-08,
+      "loss": 0.0801,
+      "step": 42368
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 5.69465445165881e-08,
+      "loss": 0.1331,
+      "step": 42432
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 5.552343680512318e-08,
+      "loss": 0.1225,
+      "step": 42496
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 5.4100329093658276e-08,
+      "loss": 0.1078,
+      "step": 42560
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 5.267722138219336e-08,
+      "loss": 0.1638,
+      "step": 42624
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 5.1254113670728456e-08,
+      "loss": 0.1839,
+      "step": 42688
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 4.983100595926354e-08,
+      "loss": 0.1643,
+      "step": 42752
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 4.840789824779863e-08,
+      "loss": 0.1923,
+      "step": 42816
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 4.698479053633372e-08,
+      "loss": 0.0508,
+      "step": 42880
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 4.55616828248688e-08,
+      "loss": 0.1679,
+      "step": 42944
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 4.41385751134039e-08,
+      "loss": 0.1319,
+      "step": 43008
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 4.271546740193898e-08,
+      "loss": 0.17,
+      "step": 43072
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 4.129235969047407e-08,
+      "loss": 0.1271,
+      "step": 43136
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 3.986925197900916e-08,
+      "loss": 0.106,
+      "step": 43200
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 3.844614426754425e-08,
+      "loss": 0.1248,
+      "step": 43264
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 3.702303655607934e-08,
+      "loss": 0.1422,
+      "step": 43328
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 3.559992884461443e-08,
+      "loss": 0.1535,
+      "step": 43392
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 3.417682113314951e-08,
+      "loss": 0.1323,
+      "step": 43456
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 3.27537134216846e-08,
+      "loss": 0.076,
+      "step": 43520
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 3.133060571021969e-08,
+      "loss": 0.1737,
+      "step": 43584
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 2.990749799875478e-08,
+      "loss": 0.089,
+      "step": 43648
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 2.848439028728987e-08,
+      "loss": 0.1577,
+      "step": 43712
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 2.7061282575824956e-08,
+      "loss": 0.1323,
+      "step": 43776
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 2.5638174864360045e-08,
+      "loss": 0.0963,
+      "step": 43840
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 2.4215067152895132e-08,
+      "loss": 0.1993,
+      "step": 43904
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 2.279195944143022e-08,
+      "loss": 0.1967,
+      "step": 43968
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 2.136885172996531e-08,
+      "loss": 0.1396,
+      "step": 44032
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 1.99457440185004e-08,
+      "loss": 0.1352,
+      "step": 44096
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 1.852263630703549e-08,
+      "loss": 0.0771,
+      "step": 44160
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.7099528595570576e-08,
+      "loss": 0.138,
+      "step": 44224
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.5676420884105666e-08,
+      "loss": 0.1512,
+      "step": 44288
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.4253313172640754e-08,
+      "loss": 0.0892,
+      "step": 44352
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 1.2830205461175842e-08,
+      "loss": 0.1576,
+      "step": 44416
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 1.1407097749710932e-08,
+      "loss": 0.131,
+      "step": 44480
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 9.983990038246018e-09,
+      "loss": 0.1815,
+      "step": 44544
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 8.560882326781107e-09,
+      "loss": 0.1482,
+      "step": 44608
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 7.137774615316196e-09,
+      "loss": 0.117,
+      "step": 44672
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.714666903851285e-09,
+      "loss": 0.1666,
+      "step": 44736
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 4.291559192386374e-09,
+      "loss": 0.2056,
+      "step": 44800
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 2.868451480921462e-09,
+      "loss": 0.2186,
+      "step": 44864
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 1.4675798274481899e-09,
+      "loss": 0.1433,
+      "step": 44928
+    }
+  ],
+  "logging_steps": 64,
+  "max_steps": 44972,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 2,
+  "save_steps": 14990,
+  "total_flos": 9.5479000805376e+19,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-44970/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2463dffe438013cfefdc9915eb0241a5214c5766ec68d593ec6bac88b5ea2196
+size 5048

config.json

@@ -0,0 +1,66 @@
+{
+  "_name_or_path": "openai/whisper-large-v2",
+  "activation_dropout": 0.0,
+  "activation_function": "gelu",
+  "apply_spec_augment": false,
+  "architectures": [
+    "WhisperForConditionalGeneration"
+  ],
+  "attention_dropout": 0.0,
+  "begin_suppress_tokens": [
+    220,
+    50257
+  ],
+  "bos_token_id": 50257,
+  "classifier_proj_size": 256,
+  "d_model": 1280,
+  "decoder_attention_heads": 20,
+  "decoder_ffn_dim": 5120,
+  "decoder_layerdrop": 0.0,
+  "decoder_layers": 32,
+  "decoder_start_token_id": 50258,
+  "dropout": 0.0,
+  "encoder_attention_heads": 20,
+  "encoder_ffn_dim": 5120,
+  "encoder_layerdrop": 0.0,
+  "encoder_layers": 32,
+  "eos_token_id": 50257,
+  "forced_decoder_ids": [
+    [
+      1,
+      50259
+    ],
+    [
+      2,
+      50359
+    ],
+    [
+      3,
+      50363
+    ]
+  ],
+  "init_std": 0.02,
+  "is_encoder_decoder": true,
+  "language": "English",
+  "mask_feature_length": 10,
+  "mask_feature_min_masks": 0,
+  "mask_feature_prob": 0.0,
+  "mask_time_length": 10,
+  "mask_time_min_masks": 2,
+  "mask_time_prob": 0.05,
+  "max_length": 448,
+  "max_source_positions": 1500,
+  "max_target_positions": 448,
+  "median_filter_width": 7,
+  "model_type": "whisper",
+  "num_hidden_layers": 32,
+  "num_mel_bins": 80,
+  "pad_token_id": 50257,
+  "scale_embedding": false,
+  "suppress_tokens": [],
+  "torch_dtype": "float32",
+  "transformers_version": "4.37.2",
+  "use_cache": true,
+  "use_weighted_layer_sum": false,
+  "vocab_size": 51865
+}

generation_config.json

@@ -0,0 +1,317 @@
+{
+  "alignment_heads": [
+    [
+      10,
+      12
+    ],
+    [
+      13,
+      17
+    ],
+    [
+      16,
+      11
+    ],
+    [
+      16,
+      12
+    ],
+    [
+      16,
+      13
+    ],
+    [
+      17,
+      15
+    ],
+    [
+      17,
+      16
+    ],
+    [
+      18,
+      4
+    ],
+    [
+      18,
+      11
+    ],
+    [
+      18,
+      19
+    ],
+    [
+      19,
+      11
+    ],
+    [
+      21,
+      2
+    ],
+    [
+      21,
+      3
+    ],
+    [
+      22,
+      3
+    ],
+    [
+      22,
+      9
+    ],
+    [
+      22,
+      12
+    ],
+    [
+      23,
+      5
+    ],
+    [
+      23,
+      7
+    ],
+    [
+      23,
+      13
+    ],
+    [
+      25,
+      5
+    ],
+    [
+      26,
+      1
+    ],
+    [
+      26,
+      12
+    ],
+    [
+      27,
+      15
+    ]
+  ],
+  "begin_suppress_tokens": [
+    220,
+    50257
+  ],
+  "bos_token_id": 50257,
+  "decoder_start_token_id": 50258,
+  "eos_token_id": 50257,
+  "forced_decoder_ids": [
+    [
+      1,
+      null
+    ],
+    [
+      2,
+      50359
+    ]
+  ],
+  "is_multilingual": true,
+  "lang_to_id": {
+    "<|af|>": 50327,
+    "<|am|>": 50334,
+    "<|ar|>": 50272,
+    "<|as|>": 50350,
+    "<|az|>": 50304,
+    "<|ba|>": 50355,
+    "<|be|>": 50330,
+    "<|bg|>": 50292,
+    "<|bn|>": 50302,
+    "<|bo|>": 50347,
+    "<|br|>": 50309,
+    "<|bs|>": 50315,
+    "<|ca|>": 50270,
+    "<|cs|>": 50283,
+    "<|cy|>": 50297,
+    "<|da|>": 50285,
+    "<|de|>": 50261,
+    "<|el|>": 50281,
+    "<|en|>": 50259,
+    "<|es|>": 50262,
+    "<|et|>": 50307,
+    "<|eu|>": 50310,
+    "<|fa|>": 50300,
+    "<|fi|>": 50277,
+    "<|fo|>": 50338,
+    "<|fr|>": 50265,
+    "<|gl|>": 50319,
+    "<|gu|>": 50333,
+    "<|haw|>": 50352,
+    "<|ha|>": 50354,
+    "<|he|>": 50279,
+    "<|hi|>": 50276,
+    "<|hr|>": 50291,
+    "<|ht|>": 50339,
+    "<|hu|>": 50286,
+    "<|hy|>": 50312,
+    "<|id|>": 50275,
+    "<|is|>": 50311,
+    "<|it|>": 50274,
+    "<|ja|>": 50266,
+    "<|jw|>": 50356,
+    "<|ka|>": 50329,
+    "<|kk|>": 50316,
+    "<|km|>": 50323,
+    "<|kn|>": 50306,
+    "<|ko|>": 50264,
+    "<|la|>": 50294,
+    "<|lb|>": 50345,
+    "<|ln|>": 50353,
+    "<|lo|>": 50336,
+    "<|lt|>": 50293,
+    "<|lv|>": 50301,
+    "<|mg|>": 50349,
+    "<|mi|>": 50295,
+    "<|mk|>": 50308,
+    "<|ml|>": 50296,
+    "<|mn|>": 50314,
+    "<|mr|>": 50320,
+    "<|ms|>": 50282,
+    "<|mt|>": 50343,
+    "<|my|>": 50346,
+    "<|ne|>": 50313,
+    "<|nl|>": 50271,
+    "<|nn|>": 50342,
+    "<|no|>": 50288,
+    "<|oc|>": 50328,
+    "<|pa|>": 50321,
+    "<|pl|>": 50269,
+    "<|ps|>": 50340,
+    "<|pt|>": 50267,
+    "<|ro|>": 50284,
+    "<|ru|>": 50263,
+    "<|sa|>": 50344,
+    "<|sd|>": 50332,
+    "<|si|>": 50322,
+    "<|sk|>": 50298,
+    "<|sl|>": 50305,
+    "<|sn|>": 50324,
+    "<|so|>": 50326,
+    "<|sq|>": 50317,
+    "<|sr|>": 50303,
+    "<|su|>": 50357,
+    "<|sv|>": 50273,
+    "<|sw|>": 50318,
+    "<|ta|>": 50287,
+    "<|te|>": 50299,
+    "<|tg|>": 50331,
+    "<|th|>": 50289,
+    "<|tk|>": 50341,
+    "<|tl|>": 50348,
+    "<|tr|>": 50268,
+    "<|tt|>": 50351,
+    "<|uk|>": 50280,
+    "<|ur|>": 50290,
+    "<|uz|>": 50337,
+    "<|vi|>": 50278,
+    "<|yi|>": 50335,
+    "<|yo|>": 50325,
+    "<|zh|>": 50260
+  },
+  "max_initial_timestamp_index": 50,
+  "max_length": 448,
+  "no_timestamps_token_id": 50363,
+  "pad_token_id": 50257,
+  "prev_sot_token_id": 50361,
+  "return_timestamps": false,
+  "suppress_tokens": [
+    1,
+    2,
+    7,
+    8,
+    9,
+    10,
+    14,
+    25,
+    26,
+    27,
+    28,
+    29,
+    31,
+    58,
+    59,
+    60,
+    61,
+    62,
+    63,
+    90,
+    91,
+    92,
+    93,
+    359,
+    503,
+    522,
+    542,
+    873,
+    893,
+    902,
+    918,
+    922,
+    931,
+    1350,
+    1853,
+    1982,
+    2460,
+    2627,
+    3246,
+    3253,
+    3268,
+    3536,
+    3846,
+    3961,
+    4183,
+    4667,
+    6585,
+    6647,
+    7273,
+    9061,
+    9383,
+    10428,
+    10929,
+    11938,
+    12033,
+    12331,
+    12562,
+    13793,
+    14157,
+    14635,
+    15265,
+    15618,
+    16553,
+    16604,
+    18362,
+    18956,
+    20075,
+    21675,
+    22520,
+    26130,
+    26161,
+    26435,
+    28279,
+    29464,
+    31650,
+    32302,
+    32470,
+    36865,
+    42863,
+    47425,
+    49870,
+    50254,
+    50258,
+    50358,
+    50359,
+    50360,
+    50361,
+    50362
+  ],
+  "task": "transcribe",
+  "task_to_id": {
+    "transcribe": 50359,
+    "translate": 50358
+  },
+  "transformers_version": "4.37.2"
+}

model-00001-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:24ca37e9a48fcaf8ce6b4a93bf73ed540dac7015a0ab63baeff0547cb4dcc057
+size 4992706480

model-00002-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dc1133d719f22fae5b256e1137e1c5f4bbbc6b828978b8b97ce2fe4c91eefd8d
+size 1180663192

preprocessor_config.json

@@ -0,0 +1,14 @@
+{
+  "chunk_length": 30,
+  "feature_extractor_type": "WhisperFeatureExtractor",
+  "feature_size": 80,
+  "hop_length": 160,
+  "n_fft": 400,
+  "n_samples": 480000,
+  "nb_max_frames": 3000,
+  "padding_side": "right",
+  "padding_value": 0.0,
+  "processor_class": "WhisperProcessor",
+  "return_attention_mask": false,
+  "sampling_rate": 16000
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2463dffe438013cfefdc9915eb0241a5214c5766ec68d593ec6bac88b5ea2196
+size 5048