Training in progress, step 1099, checkpoint

checkpoint-1099/1_Pooling/config.json

@@ -0,0 +1,10 @@
+{
+  "word_embedding_dimension": 1536,
+  "pooling_mode_cls_token": false,
+  "pooling_mode_mean_tokens": true,
+  "pooling_mode_max_tokens": false,
+  "pooling_mode_mean_sqrt_len_tokens": false,
+  "pooling_mode_weightedmean_tokens": false,
+  "pooling_mode_lasttoken": false,
+  "include_prompt": true
+}

checkpoint-1099/added_tokens.json

@@ -0,0 +1,3 @@
+{
+  "[MASK]": 128000
+}

checkpoint-1099/config.json

@@ -0,0 +1,38 @@
+{
+  "_name_or_path": "microsoft/deberta-v2-xlarge",
+  "architectures": [
+    "DebertaV2Model"
+  ],
+  "attention_head_size": 64,
+  "attention_probs_dropout_prob": 0.1,
+  "conv_act": "gelu",
+  "conv_kernel_size": 3,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 1536,
+  "initializer_range": 0.02,
+  "intermediate_size": 6144,
+  "layer_norm_eps": 1e-07,
+  "max_position_embeddings": 512,
+  "max_relative_positions": -1,
+  "model_type": "deberta-v2",
+  "norm_rel_ebd": "layer_norm",
+  "num_attention_heads": 24,
+  "num_hidden_layers": 24,
+  "pad_token_id": 0,
+  "pooler_dropout": 0,
+  "pooler_hidden_act": "gelu",
+  "pooler_hidden_size": 1536,
+  "pos_att_type": [
+    "p2c",
+    "c2p"
+  ],
+  "position_biased_input": false,
+  "position_buckets": 256,
+  "relative_attention": true,
+  "share_att_key": true,
+  "torch_dtype": "float32",
+  "transformers_version": "4.42.4",
+  "type_vocab_size": 0,
+  "vocab_size": 128100
+}

checkpoint-1099/config_sentence_transformers.json

@@ -0,0 +1,10 @@
+{
+  "__version__": {
+    "sentence_transformers": "3.0.1",
+    "transformers": "4.42.4",
+    "pytorch": "2.4.0+cu121"
+  },
+  "prompts": {},
+  "default_prompt_name": null,
+  "similarity_fn_name": null
+}

checkpoint-1099/modules.json

@@ -0,0 +1,14 @@
+[
+  {
+    "idx": 0,
+    "name": "0",
+    "path": "",
+    "type": "sentence_transformers.models.Transformer"
+  },
+  {
+    "idx": 1,
+    "name": "1",
+    "path": "1_Pooling",
+    "type": "sentence_transformers.models.Pooling"
+  }
+]

checkpoint-1099/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d794ef35e20014638a15cd85b9d10710fcfe19a72a77a84bbdc2208b46897e0a
+size 7077084396

checkpoint-1099/pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:de769edb3a50516e2f021ccd2e84e0b7313cc5dd7145f74df9a996ce7e3f4b47
+size 3538506546

checkpoint-1099/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:89fd4570abd6ab157d4f8a77d660aa5b0b48567334df6e9eee25d739c27047dc
+size 14244

checkpoint-1099/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5fcf6801a71e0b6bea6edd0851a575ddeb9661adc438e0121096f8d4fc7bfe5a
+size 1064

checkpoint-1099/sentence_bert_config.json

@@ -0,0 +1,4 @@
+{
+  "max_seq_length": 512,
+  "do_lower_case": false
+}

checkpoint-1099/special_tokens_map.json

@@ -0,0 +1,15 @@
+{
+  "bos_token": "[CLS]",
+  "cls_token": "[CLS]",
+  "eos_token": "[SEP]",
+  "mask_token": "[MASK]",
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "unk_token": {
+    "content": "[UNK]",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  }
+}

checkpoint-1099/spm.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5598d5e96f339a8d980c15f9afd405a2e5e1be7db41de3ed13b0f03fac1e8c17
+size 2447305

checkpoint-1099/tokenizer_config.json

@@ -0,0 +1,58 @@
+{
+  "added_tokens_decoder": {
+    "0": {
+      "content": "[PAD]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "[CLS]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "[SEP]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "3": {
+      "content": "[UNK]",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128000": {
+      "content": "[MASK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "[CLS]",
+  "clean_up_tokenization_spaces": true,
+  "cls_token": "[CLS]",
+  "do_lower_case": false,
+  "eos_token": "[SEP]",
+  "mask_token": "[MASK]",
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "sp_model_kwargs": {},
+  "split_by_punct": false,
+  "tokenizer_class": "DebertaV2Tokenizer",
+  "unk_token": "[UNK]",
+  "vocab_type": "spm"
+}

checkpoint-1099/trainer_state.json

@@ -0,0 +1,1860 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.7026854219948849,
+  "eval_steps": 157,
+  "global_step": 1099,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.010230179028132993,
+      "grad_norm": 200.60667419433594,
+      "learning_rate": 2.129925452609159e-07,
+      "loss": 7.1882,
+      "step": 16
+    },
+    {
+      "epoch": 0.020460358056265986,
+      "grad_norm": 174.4742889404297,
+      "learning_rate": 5.324813631522898e-07,
+      "loss": 9.1489,
+      "step": 32
+    },
+    {
+      "epoch": 0.030690537084398978,
+      "grad_norm": 193.9903106689453,
+      "learning_rate": 8.732694355697552e-07,
+      "loss": 8.805,
+      "step": 48
+    },
+    {
+      "epoch": 0.04092071611253197,
+      "grad_norm": 3077.94775390625,
+      "learning_rate": 1.192758253461129e-06,
+      "loss": 5.7489,
+      "step": 64
+    },
+    {
+      "epoch": 0.05115089514066496,
+      "grad_norm": 80.12879943847656,
+      "learning_rate": 1.5335463258785943e-06,
+      "loss": 4.7163,
+      "step": 80
+    },
+    {
+      "epoch": 0.061381074168797956,
+      "grad_norm": 99.18599700927734,
+      "learning_rate": 1.8743343982960599e-06,
+      "loss": 3.176,
+      "step": 96
+    },
+    {
+      "epoch": 0.07161125319693094,
+      "grad_norm": 23.476715087890625,
+      "learning_rate": 2.1938232161874336e-06,
+      "loss": 2.034,
+      "step": 112
+    },
+    {
+      "epoch": 0.08184143222506395,
+      "grad_norm": 39.820777893066406,
+      "learning_rate": 2.534611288604899e-06,
+      "loss": 1.1278,
+      "step": 128
+    },
+    {
+      "epoch": 0.09207161125319693,
+      "grad_norm": 12.513270378112793,
+      "learning_rate": 2.8753993610223648e-06,
+      "loss": 0.7996,
+      "step": 144
+    },
+    {
+      "epoch": 0.100383631713555,
+      "eval_Qnli-dev_cosine_accuracy": 0.662109375,
+      "eval_Qnli-dev_cosine_accuracy_threshold": 0.8656402826309204,
+      "eval_Qnli-dev_cosine_ap": 0.6654869266843996,
+      "eval_Qnli-dev_cosine_f1": 0.6504559270516718,
+      "eval_Qnli-dev_cosine_f1_threshold": 0.6888570785522461,
+      "eval_Qnli-dev_cosine_precision": 0.5071090047393365,
+      "eval_Qnli-dev_cosine_recall": 0.9067796610169492,
+      "eval_Qnli-dev_dot_accuracy": 0.654296875,
+      "eval_Qnli-dev_dot_accuracy_threshold": 905.306640625,
+      "eval_Qnli-dev_dot_ap": 0.6262830221036416,
+      "eval_Qnli-dev_dot_f1": 0.65625,
+      "eval_Qnli-dev_dot_f1_threshold": 810.4493408203125,
+      "eval_Qnli-dev_dot_precision": 0.5558823529411765,
+      "eval_Qnli-dev_dot_recall": 0.8008474576271186,
+      "eval_Qnli-dev_euclidean_accuracy": 0.662109375,
+      "eval_Qnli-dev_euclidean_accuracy_threshold": 17.05492401123047,
+      "eval_Qnli-dev_euclidean_ap": 0.6671756125841228,
+      "eval_Qnli-dev_euclidean_f1": 0.6495176848874598,
+      "eval_Qnli-dev_euclidean_f1_threshold": 23.73143768310547,
+      "eval_Qnli-dev_euclidean_precision": 0.5233160621761658,
+      "eval_Qnli-dev_euclidean_recall": 0.8559322033898306,
+      "eval_Qnli-dev_manhattan_accuracy": 0.65625,
+      "eval_Qnli-dev_manhattan_accuracy_threshold": 526.1962890625,
+      "eval_Qnli-dev_manhattan_ap": 0.6662986170174698,
+      "eval_Qnli-dev_manhattan_f1": 0.6487341772151899,
+      "eval_Qnli-dev_manhattan_f1_threshold": 748.320068359375,
+      "eval_Qnli-dev_manhattan_precision": 0.5176767676767676,
+      "eval_Qnli-dev_manhattan_recall": 0.8686440677966102,
+      "eval_Qnli-dev_max_accuracy": 0.662109375,
+      "eval_Qnli-dev_max_accuracy_threshold": 905.306640625,
+      "eval_Qnli-dev_max_ap": 0.6671756125841228,
+      "eval_Qnli-dev_max_f1": 0.65625,
+      "eval_Qnli-dev_max_f1_threshold": 810.4493408203125,
+      "eval_Qnli-dev_max_precision": 0.5558823529411765,
+      "eval_Qnli-dev_max_recall": 0.9067796610169492,
+      "eval_allNLI-dev_cosine_accuracy": 0.6875,
+      "eval_allNLI-dev_cosine_accuracy_threshold": 0.9388645887374878,
+      "eval_allNLI-dev_cosine_ap": 0.5213973850248246,
+      "eval_allNLI-dev_cosine_f1": 0.5590361445783132,
+      "eval_allNLI-dev_cosine_f1_threshold": 0.7874077558517456,
+      "eval_allNLI-dev_cosine_precision": 0.4793388429752066,
+      "eval_allNLI-dev_cosine_recall": 0.6705202312138728,
+      "eval_allNLI-dev_dot_accuracy": 0.6953125,
+      "eval_allNLI-dev_dot_accuracy_threshold": 942.6085205078125,
+      "eval_allNLI-dev_dot_ap": 0.520777376545363,
+      "eval_allNLI-dev_dot_f1": 0.5538461538461539,
+      "eval_allNLI-dev_dot_f1_threshold": 817.1332397460938,
+      "eval_allNLI-dev_dot_precision": 0.44680851063829785,
+      "eval_allNLI-dev_dot_recall": 0.7283236994219653,
+      "eval_allNLI-dev_euclidean_accuracy": 0.6875,
+      "eval_allNLI-dev_euclidean_accuracy_threshold": 12.643478393554688,
+      "eval_allNLI-dev_euclidean_ap": 0.518083458930424,
+      "eval_allNLI-dev_euclidean_f1": 0.5605381165919282,
+      "eval_allNLI-dev_euclidean_f1_threshold": 22.352760314941406,
+      "eval_allNLI-dev_euclidean_precision": 0.45787545787545786,
+      "eval_allNLI-dev_euclidean_recall": 0.7225433526011561,
+      "eval_allNLI-dev_manhattan_accuracy": 0.6875,
+      "eval_allNLI-dev_manhattan_accuracy_threshold": 392.8375549316406,
+      "eval_allNLI-dev_manhattan_ap": 0.5229585168795372,
+      "eval_allNLI-dev_manhattan_f1": 0.5590909090909091,
+      "eval_allNLI-dev_manhattan_f1_threshold": 690.5735473632812,
+      "eval_allNLI-dev_manhattan_precision": 0.4606741573033708,
+      "eval_allNLI-dev_manhattan_recall": 0.7109826589595376,
+      "eval_allNLI-dev_max_accuracy": 0.6953125,
+      "eval_allNLI-dev_max_accuracy_threshold": 942.6085205078125,
+      "eval_allNLI-dev_max_ap": 0.5229585168795372,
+      "eval_allNLI-dev_max_f1": 0.5605381165919282,
+      "eval_allNLI-dev_max_f1_threshold": 817.1332397460938,
+      "eval_allNLI-dev_max_precision": 0.4793388429752066,
+      "eval_allNLI-dev_max_recall": 0.7283236994219653,
+      "eval_sequential_score": 0.6671756125841228,
+      "eval_sts-test_pearson_cosine": 0.8718645025760672,
+      "eval_sts-test_pearson_dot": 0.8442027854618407,
+      "eval_sts-test_pearson_euclidean": 0.8902978035400368,
+      "eval_sts-test_pearson_manhattan": 0.8909390707647828,
+      "eval_sts-test_pearson_max": 0.8909390707647828,
+      "eval_sts-test_spearman_cosine": 0.8775905391512535,
+      "eval_sts-test_spearman_dot": 0.8282565695883292,
+      "eval_sts-test_spearman_euclidean": 0.8785767674497972,
+      "eval_sts-test_spearman_manhattan": 0.8796654313856657,
+      "eval_sts-test_spearman_max": 0.8796654313856657,
+      "eval_vitaminc-pairs_loss": 3.82141375541687,
+      "eval_vitaminc-pairs_runtime": 4.6321,
+      "eval_vitaminc-pairs_samples_per_second": 27.633,
+      "eval_vitaminc-pairs_steps_per_second": 0.216,
+      "step": 157
+    },
+    {
+      "epoch": 0.100383631713555,
+      "eval_negation-triplets_loss": 1.2526816129684448,
+      "eval_negation-triplets_runtime": 3.4008,
+      "eval_negation-triplets_samples_per_second": 37.638,
+      "eval_negation-triplets_steps_per_second": 0.294,
+      "step": 157
+    },
+    {
+      "epoch": 0.100383631713555,
+      "eval_scitail-pairs-pos_loss": 0.07802588492631912,
+      "eval_scitail-pairs-pos_runtime": 2.7191,
+      "eval_scitail-pairs-pos_samples_per_second": 47.075,
+      "eval_scitail-pairs-pos_steps_per_second": 0.368,
+      "step": 157
+    },
+    {
+      "epoch": 0.100383631713555,
+      "eval_scitail-pairs-qa_loss": 0.11051780730485916,
+      "eval_scitail-pairs-qa_runtime": 2.1249,
+      "eval_scitail-pairs-qa_samples_per_second": 60.237,
+      "eval_scitail-pairs-qa_steps_per_second": 0.471,
+      "step": 157
+    },
+    {
+      "epoch": 0.100383631713555,
+      "eval_xsum-pairs_loss": 0.1444225311279297,
+      "eval_xsum-pairs_runtime": 3.075,
+      "eval_xsum-pairs_samples_per_second": 41.626,
+      "eval_xsum-pairs_steps_per_second": 0.325,
+      "step": 157
+    },
+    {
+      "epoch": 0.100383631713555,
+      "eval_sciq_pairs_loss": 0.10900198668241501,
+      "eval_sciq_pairs_runtime": 4.6201,
+      "eval_sciq_pairs_samples_per_second": 27.705,
+      "eval_sciq_pairs_steps_per_second": 0.216,
+      "step": 157
+    },
+    {
+      "epoch": 0.100383631713555,
+      "eval_qasc_pairs_loss": 0.8441975116729736,
+      "eval_qasc_pairs_runtime": 2.0926,
+      "eval_qasc_pairs_samples_per_second": 61.168,
+      "eval_qasc_pairs_steps_per_second": 0.478,
+      "step": 157
+    },
+    {
+      "epoch": 0.100383631713555,
+      "eval_openbookqa_pairs_loss": 0.7852775454521179,
+      "eval_openbookqa_pairs_runtime": 2.2378,
+      "eval_openbookqa_pairs_samples_per_second": 57.198,
+      "eval_openbookqa_pairs_steps_per_second": 0.447,
+      "step": 157
+    },
+    {
+      "epoch": 0.100383631713555,
+      "eval_msmarco_pairs_loss": 0.7973570823669434,
+      "eval_msmarco_pairs_runtime": 2.2477,
+      "eval_msmarco_pairs_samples_per_second": 56.948,
+      "eval_msmarco_pairs_steps_per_second": 0.445,
+      "step": 157
+    },
+    {
+      "epoch": 0.100383631713555,
+      "eval_nq_pairs_loss": 1.149829387664795,
+      "eval_nq_pairs_runtime": 3.1982,
+      "eval_nq_pairs_samples_per_second": 40.023,
+      "eval_nq_pairs_steps_per_second": 0.313,
+      "step": 157
+    },
+    {
+      "epoch": 0.100383631713555,
+      "eval_trivia_pairs_loss": 0.46642398834228516,
+      "eval_trivia_pairs_runtime": 3.2778,
+      "eval_trivia_pairs_samples_per_second": 39.05,
+      "eval_trivia_pairs_steps_per_second": 0.305,
+      "step": 157
+    },
+    {
+      "epoch": 0.100383631713555,
+      "eval_gooaq_pairs_loss": 0.7757962346076965,
+      "eval_gooaq_pairs_runtime": 2.1385,
+      "eval_gooaq_pairs_samples_per_second": 59.856,
+      "eval_gooaq_pairs_steps_per_second": 0.468,
+      "step": 157
+    },
+    {
+      "epoch": 0.100383631713555,
+      "eval_paws-pos_loss": 0.03750575706362724,
+      "eval_paws-pos_runtime": 2.2521,
+      "eval_paws-pos_samples_per_second": 56.836,
+      "eval_paws-pos_steps_per_second": 0.444,
+      "step": 157
+    },
+    {
+      "epoch": 0.100383631713555,
+      "eval_global_dataset_loss": 0.48674729466438293,
+      "eval_global_dataset_runtime": 8.6002,
+      "eval_global_dataset_samples_per_second": 37.79,
+      "eval_global_dataset_steps_per_second": 0.349,
+      "step": 157
+    },
+    {
+      "epoch": 0.10230179028132992,
+      "grad_norm": 45.195247650146484,
+      "learning_rate": 3.2161874334398298e-06,
+      "loss": 0.54,
+      "step": 160
+    },
+    {
+      "epoch": 0.11253196930946291,
+      "grad_norm": 64.59953308105469,
+      "learning_rate": 3.556975505857295e-06,
+      "loss": 0.6267,
+      "step": 176
+    },
+    {
+      "epoch": 0.12276214833759591,
+      "grad_norm": 108.30097961425781,
+      "learning_rate": 3.8977635782747605e-06,
+      "loss": 0.401,
+      "step": 192
+    },
+    {
+      "epoch": 0.1329923273657289,
+      "grad_norm": 10.606786727905273,
+      "learning_rate": 4.238551650692226e-06,
+      "loss": 0.455,
+      "step": 208
+    },
+    {
+      "epoch": 0.1432225063938619,
+      "grad_norm": 10.502503395080566,
+      "learning_rate": 4.579339723109691e-06,
+      "loss": 0.308,
+      "step": 224
+    },
+    {
+      "epoch": 0.1534526854219949,
+      "grad_norm": 10.550800323486328,
+      "learning_rate": 4.920127795527157e-06,
+      "loss": 0.2808,
+      "step": 240
+    },
+    {
+      "epoch": 0.1636828644501279,
+      "grad_norm": 0.047671813517808914,
+      "learning_rate": 5.260915867944623e-06,
+      "loss": 0.319,
+      "step": 256
+    },
+    {
+      "epoch": 0.17391304347826086,
+      "grad_norm": 4.752868175506592,
+      "learning_rate": 5.6017039403620875e-06,
+      "loss": 0.3241,
+      "step": 272
+    },
+    {
+      "epoch": 0.18414322250639387,
+      "grad_norm": 3.744504690170288,
+      "learning_rate": 5.942492012779553e-06,
+      "loss": 0.2181,
+      "step": 288
+    },
+    {
+      "epoch": 0.19437340153452684,
+      "grad_norm": 0.10304329544305801,
+      "learning_rate": 6.283280085197019e-06,
+      "loss": 0.3247,
+      "step": 304
+    },
+    {
+      "epoch": 0.20076726342711,
+      "eval_Qnli-dev_cosine_accuracy": 0.68359375,
+      "eval_Qnli-dev_cosine_accuracy_threshold": 0.728395938873291,
+      "eval_Qnli-dev_cosine_ap": 0.7208629559865216,
+      "eval_Qnli-dev_cosine_f1": 0.6887661141804787,
+      "eval_Qnli-dev_cosine_f1_threshold": 0.671984076499939,
+      "eval_Qnli-dev_cosine_precision": 0.6091205211726385,
+      "eval_Qnli-dev_cosine_recall": 0.7923728813559322,
+      "eval_Qnli-dev_dot_accuracy": 0.681640625,
+      "eval_Qnli-dev_dot_accuracy_threshold": 713.8800048828125,
+      "eval_Qnli-dev_dot_ap": 0.6851873907589782,
+      "eval_Qnli-dev_dot_f1": 0.6788685524126457,
+      "eval_Qnli-dev_dot_f1_threshold": 571.084228515625,
+      "eval_Qnli-dev_dot_precision": 0.5589041095890411,
+      "eval_Qnli-dev_dot_recall": 0.864406779661017,
+      "eval_Qnli-dev_euclidean_accuracy": 0.689453125,
+      "eval_Qnli-dev_euclidean_accuracy_threshold": 23.01095962524414,
+      "eval_Qnli-dev_euclidean_ap": 0.7234217762808675,
+      "eval_Qnli-dev_euclidean_f1": 0.6831858407079647,
+      "eval_Qnli-dev_euclidean_f1_threshold": 25.087703704833984,
+      "eval_Qnli-dev_euclidean_precision": 0.5866261398176292,
+      "eval_Qnli-dev_euclidean_recall": 0.8177966101694916,
+      "eval_Qnli-dev_manhattan_accuracy": 0.6875,
+      "eval_Qnli-dev_manhattan_accuracy_threshold": 715.2478637695312,
+      "eval_Qnli-dev_manhattan_ap": 0.7243486016703601,
+      "eval_Qnli-dev_manhattan_f1": 0.6844919786096257,
+      "eval_Qnli-dev_manhattan_f1_threshold": 780.095947265625,
+      "eval_Qnli-dev_manhattan_precision": 0.5907692307692308,
+      "eval_Qnli-dev_manhattan_recall": 0.8135593220338984,
+      "eval_Qnli-dev_max_accuracy": 0.689453125,
+      "eval_Qnli-dev_max_accuracy_threshold": 715.2478637695312,
+      "eval_Qnli-dev_max_ap": 0.7243486016703601,
+      "eval_Qnli-dev_max_f1": 0.6887661141804787,
+      "eval_Qnli-dev_max_f1_threshold": 780.095947265625,
+      "eval_Qnli-dev_max_precision": 0.6091205211726385,
+      "eval_Qnli-dev_max_recall": 0.864406779661017,
+      "eval_allNLI-dev_cosine_accuracy": 0.720703125,
+      "eval_allNLI-dev_cosine_accuracy_threshold": 0.8353008031845093,
+      "eval_allNLI-dev_cosine_ap": 0.5917629293541244,
+      "eval_allNLI-dev_cosine_f1": 0.6000000000000001,
+      "eval_allNLI-dev_cosine_f1_threshold": 0.7087839245796204,
+      "eval_allNLI-dev_cosine_precision": 0.5019455252918288,
+      "eval_allNLI-dev_cosine_recall": 0.7456647398843931,
+      "eval_allNLI-dev_dot_accuracy": 0.71484375,
+      "eval_allNLI-dev_dot_accuracy_threshold": 817.4105224609375,
+      "eval_allNLI-dev_dot_ap": 0.5858245181656638,
+      "eval_allNLI-dev_dot_f1": 0.5980392156862746,
+      "eval_allNLI-dev_dot_f1_threshold": 674.387939453125,
+      "eval_allNLI-dev_dot_precision": 0.5191489361702127,
+      "eval_allNLI-dev_dot_recall": 0.7052023121387283,
+      "eval_allNLI-dev_euclidean_accuracy": 0.716796875,
+      "eval_allNLI-dev_euclidean_accuracy_threshold": 17.109304428100586,
+      "eval_allNLI-dev_euclidean_ap": 0.5912447191211185,
+      "eval_allNLI-dev_euclidean_f1": 0.6013986013986015,
+      "eval_allNLI-dev_euclidean_f1_threshold": 23.201446533203125,
+      "eval_allNLI-dev_euclidean_precision": 0.50390625,
+      "eval_allNLI-dev_euclidean_recall": 0.7456647398843931,
+      "eval_allNLI-dev_manhattan_accuracy": 0.71875,
+      "eval_allNLI-dev_manhattan_accuracy_threshold": 552.2703857421875,
+      "eval_allNLI-dev_manhattan_ap": 0.5910112781514775,
+      "eval_allNLI-dev_manhattan_f1": 0.5981735159817351,
+      "eval_allNLI-dev_manhattan_f1_threshold": 731.6765747070312,
+      "eval_allNLI-dev_manhattan_precision": 0.49433962264150944,
+      "eval_allNLI-dev_manhattan_recall": 0.7572254335260116,
+      "eval_allNLI-dev_max_accuracy": 0.720703125,
+      "eval_allNLI-dev_max_accuracy_threshold": 817.4105224609375,
+      "eval_allNLI-dev_max_ap": 0.5917629293541244,
+      "eval_allNLI-dev_max_f1": 0.6013986013986015,
+      "eval_allNLI-dev_max_f1_threshold": 731.6765747070312,
+      "eval_allNLI-dev_max_precision": 0.5191489361702127,
+      "eval_allNLI-dev_max_recall": 0.7572254335260116,
+      "eval_sequential_score": 0.7243486016703601,
+      "eval_sts-test_pearson_cosine": 0.91107907465776,
+      "eval_sts-test_pearson_dot": 0.8802333333831622,
+      "eval_sts-test_pearson_euclidean": 0.92175721802385,
+      "eval_sts-test_pearson_manhattan": 0.9207411620779591,
+      "eval_sts-test_pearson_max": 0.92175721802385,
+      "eval_sts-test_spearman_cosine": 0.9170956996722621,
+      "eval_sts-test_spearman_dot": 0.8733774316475602,
+      "eval_sts-test_spearman_euclidean": 0.9165704988685727,
+      "eval_sts-test_spearman_manhattan": 0.9153493264477239,
+      "eval_sts-test_spearman_max": 0.9170956996722621,
+      "eval_vitaminc-pairs_loss": 3.3627231121063232,
+      "eval_vitaminc-pairs_runtime": 4.4746,
+      "eval_vitaminc-pairs_samples_per_second": 28.606,
+      "eval_vitaminc-pairs_steps_per_second": 0.223,
+      "step": 314
+    },
+    {
+      "epoch": 0.20076726342711,
+      "eval_negation-triplets_loss": 0.7009554505348206,
+      "eval_negation-triplets_runtime": 3.3424,
+      "eval_negation-triplets_samples_per_second": 38.296,
+      "eval_negation-triplets_steps_per_second": 0.299,
+      "step": 314
+    },
+    {
+      "epoch": 0.20076726342711,
+      "eval_scitail-pairs-pos_loss": 0.027083460241556168,
+      "eval_scitail-pairs-pos_runtime": 2.6253,
+      "eval_scitail-pairs-pos_samples_per_second": 48.757,
+      "eval_scitail-pairs-pos_steps_per_second": 0.381,
+      "step": 314
+    },
+    {
+      "epoch": 0.20076726342711,
+      "eval_scitail-pairs-qa_loss": 0.0013978304341435432,
+      "eval_scitail-pairs-qa_runtime": 2.0929,
+      "eval_scitail-pairs-qa_samples_per_second": 61.159,
+      "eval_scitail-pairs-qa_steps_per_second": 0.478,
+      "step": 314
+    },
+    {
+      "epoch": 0.20076726342711,
+      "eval_xsum-pairs_loss": 0.03819597512483597,
+      "eval_xsum-pairs_runtime": 3.0764,
+      "eval_xsum-pairs_samples_per_second": 41.607,
+      "eval_xsum-pairs_steps_per_second": 0.325,
+      "step": 314
+    },
+    {
+      "epoch": 0.20076726342711,
+      "eval_sciq_pairs_loss": 0.04831065982580185,
+      "eval_sciq_pairs_runtime": 4.6239,
+      "eval_sciq_pairs_samples_per_second": 27.682,
+      "eval_sciq_pairs_steps_per_second": 0.216,
+      "step": 314
+    },
+    {
+      "epoch": 0.20076726342711,
+      "eval_qasc_pairs_loss": 0.11294831335544586,
+      "eval_qasc_pairs_runtime": 2.1013,
+      "eval_qasc_pairs_samples_per_second": 60.915,
+      "eval_qasc_pairs_steps_per_second": 0.476,
+      "step": 314
+    },
+    {
+      "epoch": 0.20076726342711,
+      "eval_openbookqa_pairs_loss": 0.4045051634311676,
+      "eval_openbookqa_pairs_runtime": 2.2313,
+      "eval_openbookqa_pairs_samples_per_second": 57.367,
+      "eval_openbookqa_pairs_steps_per_second": 0.448,
+      "step": 314
+    },
+    {
+      "epoch": 0.20076726342711,
+      "eval_msmarco_pairs_loss": 0.17428961396217346,
+      "eval_msmarco_pairs_runtime": 2.2379,
+      "eval_msmarco_pairs_samples_per_second": 57.197,
+      "eval_msmarco_pairs_steps_per_second": 0.447,
+      "step": 314
+    },
+    {
+      "epoch": 0.20076726342711,
+      "eval_nq_pairs_loss": 0.18139179050922394,
+      "eval_nq_pairs_runtime": 3.1825,
+      "eval_nq_pairs_samples_per_second": 40.22,
+      "eval_nq_pairs_steps_per_second": 0.314,
+      "step": 314
+    },
+    {
+      "epoch": 0.20076726342711,
+      "eval_trivia_pairs_loss": 0.10668718814849854,
+      "eval_trivia_pairs_runtime": 3.2891,
+      "eval_trivia_pairs_samples_per_second": 38.917,
+      "eval_trivia_pairs_steps_per_second": 0.304,
+      "step": 314
+    },
+    {
+      "epoch": 0.20076726342711,
+      "eval_gooaq_pairs_loss": 0.2038198560476303,
+      "eval_gooaq_pairs_runtime": 2.209,
+      "eval_gooaq_pairs_samples_per_second": 57.945,
+      "eval_gooaq_pairs_steps_per_second": 0.453,
+      "step": 314
+    },
+    {
+      "epoch": 0.20076726342711,
+      "eval_paws-pos_loss": 0.02337946929037571,
+      "eval_paws-pos_runtime": 2.289,
+      "eval_paws-pos_samples_per_second": 55.92,
+      "eval_paws-pos_steps_per_second": 0.437,
+      "step": 314
+    },
+    {
+      "epoch": 0.20076726342711,
+      "eval_global_dataset_loss": 0.23699811100959778,
+      "eval_global_dataset_runtime": 8.6273,
+      "eval_global_dataset_samples_per_second": 37.671,
+      "eval_global_dataset_steps_per_second": 0.348,
+      "step": 314
+    },
+    {
+      "epoch": 0.20460358056265984,
+      "grad_norm": 7.919135570526123,
+      "learning_rate": 6.624068157614484e-06,
+      "loss": 0.233,
+      "step": 320
+    },
+    {
+      "epoch": 0.21483375959079284,
+      "grad_norm": 8.169248580932617,
+      "learning_rate": 6.96485623003195e-06,
+      "loss": 0.1946,
+      "step": 336
+    },
+    {
+      "epoch": 0.22506393861892582,
+      "grad_norm": 4.551661968231201,
+      "learning_rate": 7.305644302449415e-06,
+      "loss": 0.2393,
+      "step": 352
+    },
+    {
+      "epoch": 0.23529411764705882,
+      "grad_norm": 49.21135330200195,
+      "learning_rate": 7.64643237486688e-06,
+      "loss": 0.1476,
+      "step": 368
+    },
+    {
+      "epoch": 0.24552429667519182,
+      "grad_norm": 12.119938850402832,
+      "learning_rate": 7.987220447284347e-06,
+      "loss": 0.1976,
+      "step": 384
+    },
+    {
+      "epoch": 0.2557544757033248,
+      "grad_norm": 6.922146797180176,
+      "learning_rate": 8.32800851970181e-06,
+      "loss": 0.1139,
+      "step": 400
+    },
+    {
+      "epoch": 0.2659846547314578,
+      "grad_norm": 1.659225583076477,
+      "learning_rate": 8.668796592119276e-06,
+      "loss": 0.1986,
+      "step": 416
+    },
+    {
+      "epoch": 0.27621483375959077,
+      "grad_norm": 11.489442825317383,
+      "learning_rate": 9.009584664536743e-06,
+      "loss": 0.2405,
+      "step": 432
+    },
+    {
+      "epoch": 0.2864450127877238,
+      "grad_norm": 6.8070149421691895,
+      "learning_rate": 9.350372736954207e-06,
+      "loss": 0.1519,
+      "step": 448
+    },
+    {
+      "epoch": 0.2966751918158568,
+      "grad_norm": 8.569122314453125,
+      "learning_rate": 9.691160809371672e-06,
+      "loss": 0.1508,
+      "step": 464
+    },
+    {
+      "epoch": 0.30115089514066495,
+      "eval_Qnli-dev_cosine_accuracy": 0.689453125,
+      "eval_Qnli-dev_cosine_accuracy_threshold": 0.7045497894287109,
+      "eval_Qnli-dev_cosine_ap": 0.7262372250492908,
+      "eval_Qnli-dev_cosine_f1": 0.693103448275862,
+      "eval_Qnli-dev_cosine_f1_threshold": 0.6277430057525635,
+      "eval_Qnli-dev_cosine_precision": 0.5843023255813954,
+      "eval_Qnli-dev_cosine_recall": 0.8516949152542372,
+      "eval_Qnli-dev_dot_accuracy": 0.671875,
+      "eval_Qnli-dev_dot_accuracy_threshold": 652.669677734375,
+      "eval_Qnli-dev_dot_ap": 0.6928244257428712,
+      "eval_Qnli-dev_dot_f1": 0.6923076923076922,
+      "eval_Qnli-dev_dot_f1_threshold": 571.573486328125,
+      "eval_Qnli-dev_dot_precision": 0.5892857142857143,
+      "eval_Qnli-dev_dot_recall": 0.8389830508474576,
+      "eval_Qnli-dev_euclidean_accuracy": 0.689453125,
+      "eval_Qnli-dev_euclidean_accuracy_threshold": 23.253864288330078,
+      "eval_Qnli-dev_euclidean_ap": 0.7293266479878985,
+      "eval_Qnli-dev_euclidean_f1": 0.6928571428571428,
+      "eval_Qnli-dev_euclidean_f1_threshold": 25.396923065185547,
+      "eval_Qnli-dev_euclidean_precision": 0.5987654320987654,
+      "eval_Qnli-dev_euclidean_recall": 0.8220338983050848,
+      "eval_Qnli-dev_manhattan_accuracy": 0.6875,
+      "eval_Qnli-dev_manhattan_accuracy_threshold": 704.4737548828125,
+      "eval_Qnli-dev_manhattan_ap": 0.7283792835739776,
+      "eval_Qnli-dev_manhattan_f1": 0.6925675675675677,
+      "eval_Qnli-dev_manhattan_f1_threshold": 820.7425537109375,
+      "eval_Qnli-dev_manhattan_precision": 0.5758426966292135,
+      "eval_Qnli-dev_manhattan_recall": 0.8686440677966102,
+      "eval_Qnli-dev_max_accuracy": 0.689453125,
+      "eval_Qnli-dev_max_accuracy_threshold": 704.4737548828125,
+      "eval_Qnli-dev_max_ap": 0.7293266479878985,
+      "eval_Qnli-dev_max_f1": 0.693103448275862,
+      "eval_Qnli-dev_max_f1_threshold": 820.7425537109375,
+      "eval_Qnli-dev_max_precision": 0.5987654320987654,
+      "eval_Qnli-dev_max_recall": 0.8686440677966102,
+      "eval_allNLI-dev_cosine_accuracy": 0.716796875,
+      "eval_allNLI-dev_cosine_accuracy_threshold": 0.805169939994812,
+      "eval_allNLI-dev_cosine_ap": 0.5991158471614924,
+      "eval_allNLI-dev_cosine_f1": 0.6095617529880478,
+      "eval_allNLI-dev_cosine_f1_threshold": 0.6292704343795776,
+      "eval_allNLI-dev_cosine_precision": 0.46504559270516715,
+      "eval_allNLI-dev_cosine_recall": 0.884393063583815,
+      "eval_allNLI-dev_dot_accuracy": 0.7265625,
+      "eval_allNLI-dev_dot_accuracy_threshold": 685.67724609375,
+      "eval_allNLI-dev_dot_ap": 0.6022852154957883,
+      "eval_allNLI-dev_dot_f1": 0.603112840466926,
+      "eval_allNLI-dev_dot_f1_threshold": 566.2506103515625,
+      "eval_allNLI-dev_dot_precision": 0.45454545454545453,
+      "eval_allNLI-dev_dot_recall": 0.8959537572254336,
+      "eval_allNLI-dev_euclidean_accuracy": 0.71484375,
+      "eval_allNLI-dev_euclidean_accuracy_threshold": 20.33218765258789,
+      "eval_allNLI-dev_euclidean_ap": 0.5957908065503393,
+      "eval_allNLI-dev_euclidean_f1": 0.6140724946695097,
+      "eval_allNLI-dev_euclidean_f1_threshold": 25.308191299438477,
+      "eval_allNLI-dev_euclidean_precision": 0.4864864864864865,
+      "eval_allNLI-dev_euclidean_recall": 0.8323699421965318,
+      "eval_allNLI-dev_manhattan_accuracy": 0.716796875,
+      "eval_allNLI-dev_manhattan_accuracy_threshold": 638.1854248046875,
+      "eval_allNLI-dev_manhattan_ap": 0.5968486499581244,
+      "eval_allNLI-dev_manhattan_f1": 0.616052060737527,
+      "eval_allNLI-dev_manhattan_f1_threshold": 783.5068359375,
+      "eval_allNLI-dev_manhattan_precision": 0.4930555555555556,
+      "eval_allNLI-dev_manhattan_recall": 0.8208092485549133,
+      "eval_allNLI-dev_max_accuracy": 0.7265625,
+      "eval_allNLI-dev_max_accuracy_threshold": 685.67724609375,
+      "eval_allNLI-dev_max_ap": 0.6022852154957883,
+      "eval_allNLI-dev_max_f1": 0.616052060737527,
+      "eval_allNLI-dev_max_f1_threshold": 783.5068359375,
+      "eval_allNLI-dev_max_precision": 0.4930555555555556,
+      "eval_allNLI-dev_max_recall": 0.8959537572254336,
+      "eval_sequential_score": 0.7293266479878985,
+      "eval_sts-test_pearson_cosine": 0.9163990296196276,
+      "eval_sts-test_pearson_dot": 0.9067925652682547,
+      "eval_sts-test_pearson_euclidean": 0.9258813754113262,
+      "eval_sts-test_pearson_manhattan": 0.9250575950562571,
+      "eval_sts-test_pearson_max": 0.9258813754113262,
+      "eval_sts-test_spearman_cosine": 0.9200439127727275,
+      "eval_sts-test_spearman_dot": 0.9030931642366772,
+      "eval_sts-test_spearman_euclidean": 0.920441080217576,
+      "eval_sts-test_spearman_manhattan": 0.9196469690841858,
+      "eval_sts-test_spearman_max": 0.920441080217576,
+      "eval_vitaminc-pairs_loss": 2.4166860580444336,
+      "eval_vitaminc-pairs_runtime": 4.558,
+      "eval_vitaminc-pairs_samples_per_second": 28.082,
+      "eval_vitaminc-pairs_steps_per_second": 0.219,
+      "step": 471
+    },
+    {
+      "epoch": 0.30115089514066495,
+      "eval_negation-triplets_loss": 0.6827869415283203,
+      "eval_negation-triplets_runtime": 3.2935,
+      "eval_negation-triplets_samples_per_second": 38.864,
+      "eval_negation-triplets_steps_per_second": 0.304,
+      "step": 471
+    },
+    {
+      "epoch": 0.30115089514066495,
+      "eval_scitail-pairs-pos_loss": 0.0062057748436927795,
+      "eval_scitail-pairs-pos_runtime": 2.638,
+      "eval_scitail-pairs-pos_samples_per_second": 48.521,
+      "eval_scitail-pairs-pos_steps_per_second": 0.379,
+      "step": 471
+    },
+    {
+      "epoch": 0.30115089514066495,
+      "eval_scitail-pairs-qa_loss": 0.0004358906007837504,
+      "eval_scitail-pairs-qa_runtime": 2.0989,
+      "eval_scitail-pairs-qa_samples_per_second": 60.985,
+      "eval_scitail-pairs-qa_steps_per_second": 0.476,
+      "step": 471
+    },
+    {
+      "epoch": 0.30115089514066495,
+      "eval_xsum-pairs_loss": 0.03286559134721756,
+      "eval_xsum-pairs_runtime": 3.0772,
+      "eval_xsum-pairs_samples_per_second": 41.596,
+      "eval_xsum-pairs_steps_per_second": 0.325,
+      "step": 471
+    },
+    {
+      "epoch": 0.30115089514066495,
+      "eval_sciq_pairs_loss": 0.04141838476061821,
+      "eval_sciq_pairs_runtime": 4.69,
+      "eval_sciq_pairs_samples_per_second": 27.292,
+      "eval_sciq_pairs_steps_per_second": 0.213,
+      "step": 471
+    },
+    {
+      "epoch": 0.30115089514066495,
+      "eval_qasc_pairs_loss": 0.03906266763806343,
+      "eval_qasc_pairs_runtime": 2.104,
+      "eval_qasc_pairs_samples_per_second": 60.836,
+      "eval_qasc_pairs_steps_per_second": 0.475,
+      "step": 471
+    },
+    {
+      "epoch": 0.30115089514066495,
+      "eval_openbookqa_pairs_loss": 0.38474294543266296,
+      "eval_openbookqa_pairs_runtime": 2.2386,
+      "eval_openbookqa_pairs_samples_per_second": 57.18,
+      "eval_openbookqa_pairs_steps_per_second": 0.447,
+      "step": 471
+    },
+    {
+      "epoch": 0.30115089514066495,
+      "eval_msmarco_pairs_loss": 0.08647920191287994,
+      "eval_msmarco_pairs_runtime": 2.2858,
+      "eval_msmarco_pairs_samples_per_second": 55.999,
+      "eval_msmarco_pairs_steps_per_second": 0.437,
+      "step": 471
+    },
+    {
+      "epoch": 0.30115089514066495,
+      "eval_nq_pairs_loss": 0.10326902568340302,
+      "eval_nq_pairs_runtime": 3.1905,
+      "eval_nq_pairs_samples_per_second": 40.12,
+      "eval_nq_pairs_steps_per_second": 0.313,
+      "step": 471
+    },
+    {
+      "epoch": 0.30115089514066495,
+      "eval_trivia_pairs_loss": 0.09608602523803711,
+      "eval_trivia_pairs_runtime": 3.256,
+      "eval_trivia_pairs_samples_per_second": 39.312,
+      "eval_trivia_pairs_steps_per_second": 0.307,
+      "step": 471
+    },
+    {
+      "epoch": 0.30115089514066495,
+      "eval_gooaq_pairs_loss": 0.12804344296455383,
+      "eval_gooaq_pairs_runtime": 2.1753,
+      "eval_gooaq_pairs_samples_per_second": 58.841,
+      "eval_gooaq_pairs_steps_per_second": 0.46,
+      "step": 471
+    },
+    {
+      "epoch": 0.30115089514066495,
+      "eval_paws-pos_loss": 0.025196708738803864,
+      "eval_paws-pos_runtime": 2.3068,
+      "eval_paws-pos_samples_per_second": 55.488,
+      "eval_paws-pos_steps_per_second": 0.434,
+      "step": 471
+    },
+    {
+      "epoch": 0.30115089514066495,
+      "eval_global_dataset_loss": 0.17351965606212616,
+      "eval_global_dataset_runtime": 8.6124,
+      "eval_global_dataset_samples_per_second": 37.736,
+      "eval_global_dataset_steps_per_second": 0.348,
+      "step": 471
+    },
+    {
+      "epoch": 0.3069053708439898,
+      "grad_norm": 0.09573666751384735,
+      "learning_rate": 1.0031948881789138e-05,
+      "loss": 0.1457,
+      "step": 480
+    },
+    {
+      "epoch": 0.3171355498721228,
+      "grad_norm": 2.005824565887451,
+      "learning_rate": 1.0372736954206603e-05,
+      "loss": 0.1086,
+      "step": 496
+    },
+    {
+      "epoch": 0.3273657289002558,
+      "grad_norm": 10.186065673828125,
+      "learning_rate": 1.071352502662407e-05,
+      "loss": 0.1412,
+      "step": 512
+    },
+    {
+      "epoch": 0.3375959079283887,
+      "grad_norm": 29.2017822265625,
+      "learning_rate": 1.1054313099041534e-05,
+      "loss": 0.1538,
+      "step": 528
+    },
+    {
+      "epoch": 0.34782608695652173,
+      "grad_norm": 7.944432735443115,
+      "learning_rate": 1.1395101171459e-05,
+      "loss": 0.1013,
+      "step": 544
+    },
+    {
+      "epoch": 0.35805626598465473,
+      "grad_norm": 5.722199440002441,
+      "learning_rate": 1.1735889243876466e-05,
+      "loss": 0.1007,
+      "step": 560
+    },
+    {
+      "epoch": 0.36828644501278773,
+      "grad_norm": 10.10983943939209,
+      "learning_rate": 1.207667731629393e-05,
+      "loss": 0.0853,
+      "step": 576
+    },
+    {
+      "epoch": 0.37851662404092073,
+      "grad_norm": 0.017955342307686806,
+      "learning_rate": 1.2417465388711395e-05,
+      "loss": 0.0696,
+      "step": 592
+    },
+    {
+      "epoch": 0.3887468030690537,
+      "grad_norm": 87.81786346435547,
+      "learning_rate": 1.2758253461128862e-05,
+      "loss": 0.1468,
+      "step": 608
+    },
+    {
+      "epoch": 0.3989769820971867,
+      "grad_norm": 4.879648208618164,
+      "learning_rate": 1.3099041533546326e-05,
+      "loss": 0.1314,
+      "step": 624
+    },
+    {
+      "epoch": 0.40153452685422,
+      "eval_Qnli-dev_cosine_accuracy": 0.69140625,
+      "eval_Qnli-dev_cosine_accuracy_threshold": 0.723533034324646,
+      "eval_Qnli-dev_cosine_ap": 0.7379282328356602,
+      "eval_Qnli-dev_cosine_f1": 0.68824531516184,
+      "eval_Qnli-dev_cosine_f1_threshold": 0.6101292967796326,
+      "eval_Qnli-dev_cosine_precision": 0.5754985754985755,
+      "eval_Qnli-dev_cosine_recall": 0.8559322033898306,
+      "eval_Qnli-dev_dot_accuracy": 0.67578125,
+      "eval_Qnli-dev_dot_accuracy_threshold": 669.32177734375,
+      "eval_Qnli-dev_dot_ap": 0.7085572115324772,
+      "eval_Qnli-dev_dot_f1": 0.6891651865008882,
+      "eval_Qnli-dev_dot_f1_threshold": 557.8812255859375,
+      "eval_Qnli-dev_dot_precision": 0.5932721712538226,
+      "eval_Qnli-dev_dot_recall": 0.8220338983050848,
+      "eval_Qnli-dev_euclidean_accuracy": 0.69921875,
+      "eval_Qnli-dev_euclidean_accuracy_threshold": 24.064918518066406,
+      "eval_Qnli-dev_euclidean_ap": 0.7427699086648181,
+      "eval_Qnli-dev_euclidean_f1": 0.6867469879518073,
+      "eval_Qnli-dev_euclidean_f1_threshold": 24.312030792236328,
+      "eval_Qnli-dev_euclidean_precision": 0.6526717557251909,
+      "eval_Qnli-dev_euclidean_recall": 0.7245762711864406,
+      "eval_Qnli-dev_manhattan_accuracy": 0.697265625,
+      "eval_Qnli-dev_manhattan_accuracy_threshold": 756.5322875976562,
+      "eval_Qnli-dev_manhattan_ap": 0.7423808621909044,
+      "eval_Qnli-dev_manhattan_f1": 0.6892655367231638,
+      "eval_Qnli-dev_manhattan_f1_threshold": 785.015380859375,
+      "eval_Qnli-dev_manhattan_precision": 0.6203389830508474,
+      "eval_Qnli-dev_manhattan_recall": 0.7754237288135594,
+      "eval_Qnli-dev_max_accuracy": 0.69921875,
+      "eval_Qnli-dev_max_accuracy_threshold": 756.5322875976562,
+      "eval_Qnli-dev_max_ap": 0.7427699086648181,
+      "eval_Qnli-dev_max_f1": 0.6892655367231638,
+      "eval_Qnli-dev_max_f1_threshold": 785.015380859375,
+      "eval_Qnli-dev_max_precision": 0.6526717557251909,
+      "eval_Qnli-dev_max_recall": 0.8559322033898306,
+      "eval_allNLI-dev_cosine_accuracy": 0.716796875,
+      "eval_allNLI-dev_cosine_accuracy_threshold": 0.822735607624054,
+      "eval_allNLI-dev_cosine_ap": 0.6013654435047096,
+      "eval_allNLI-dev_cosine_f1": 0.6241758241758242,
+      "eval_allNLI-dev_cosine_f1_threshold": 0.6583813428878784,
+      "eval_allNLI-dev_cosine_precision": 0.5035460992907801,
+      "eval_allNLI-dev_cosine_recall": 0.8208092485549133,
+      "eval_allNLI-dev_dot_accuracy": 0.71484375,
+      "eval_allNLI-dev_dot_accuracy_threshold": 780.966796875,
+      "eval_allNLI-dev_dot_ap": 0.6035715304630727,
+      "eval_allNLI-dev_dot_f1": 0.6181015452538632,
+      "eval_allNLI-dev_dot_f1_threshold": 605.0504150390625,
+      "eval_allNLI-dev_dot_precision": 0.5,
+      "eval_allNLI-dev_dot_recall": 0.8092485549132948,
+      "eval_allNLI-dev_euclidean_accuracy": 0.71875,
+      "eval_allNLI-dev_euclidean_accuracy_threshold": 19.111133575439453,
+      "eval_allNLI-dev_euclidean_ap": 0.5987461911075059,
+      "eval_allNLI-dev_euclidean_f1": 0.6140724946695097,
+      "eval_allNLI-dev_euclidean_f1_threshold": 25.491098403930664,
+      "eval_allNLI-dev_euclidean_precision": 0.4864864864864865,
+      "eval_allNLI-dev_euclidean_recall": 0.8323699421965318,
+      "eval_allNLI-dev_manhattan_accuracy": 0.71875,
+      "eval_allNLI-dev_manhattan_accuracy_threshold": 601.7767333984375,
+      "eval_allNLI-dev_manhattan_ap": 0.5998202267289205,
+      "eval_allNLI-dev_manhattan_f1": 0.6160919540229884,
+      "eval_allNLI-dev_manhattan_f1_threshold": 765.3077392578125,
+      "eval_allNLI-dev_manhattan_precision": 0.5114503816793893,
+      "eval_allNLI-dev_manhattan_recall": 0.7745664739884393,
+      "eval_allNLI-dev_max_accuracy": 0.71875,
+      "eval_allNLI-dev_max_accuracy_threshold": 780.966796875,
+      "eval_allNLI-dev_max_ap": 0.6035715304630727,
+      "eval_allNLI-dev_max_f1": 0.6241758241758242,
+      "eval_allNLI-dev_max_f1_threshold": 765.3077392578125,
+      "eval_allNLI-dev_max_precision": 0.5114503816793893,
+      "eval_allNLI-dev_max_recall": 0.8323699421965318,
+      "eval_sequential_score": 0.7427699086648181,
+      "eval_sts-test_pearson_cosine": 0.9073197584934252,
+      "eval_sts-test_pearson_dot": 0.8968254992482173,
+      "eval_sts-test_pearson_euclidean": 0.9181231094029128,
+      "eval_sts-test_pearson_manhattan": 0.9171411172664841,
+      "eval_sts-test_pearson_max": 0.9181231094029128,
+      "eval_sts-test_spearman_cosine": 0.9153029193896612,
+      "eval_sts-test_spearman_dot": 0.897675084268759,
+      "eval_sts-test_spearman_euclidean": 0.9150786708188288,
+      "eval_sts-test_spearman_manhattan": 0.9142644663690737,
+      "eval_sts-test_spearman_max": 0.9153029193896612,
+      "eval_vitaminc-pairs_loss": 3.235485076904297,
+      "eval_vitaminc-pairs_runtime": 4.4491,
+      "eval_vitaminc-pairs_samples_per_second": 28.77,
+      "eval_vitaminc-pairs_steps_per_second": 0.225,
+      "step": 628
+    },
+    {
+      "epoch": 0.40153452685422,
+      "eval_negation-triplets_loss": 0.6763876080513,
+      "eval_negation-triplets_runtime": 3.2743,
+      "eval_negation-triplets_samples_per_second": 39.092,
+      "eval_negation-triplets_steps_per_second": 0.305,
+      "step": 628
+    },
+    {
+      "epoch": 0.40153452685422,
+      "eval_scitail-pairs-pos_loss": 0.009911377914249897,
+      "eval_scitail-pairs-pos_runtime": 2.585,
+      "eval_scitail-pairs-pos_samples_per_second": 49.516,
+      "eval_scitail-pairs-pos_steps_per_second": 0.387,
+      "step": 628
+    },
+    {
+      "epoch": 0.40153452685422,
+      "eval_scitail-pairs-qa_loss": 0.0001384670613333583,
+      "eval_scitail-pairs-qa_runtime": 2.0603,
+      "eval_scitail-pairs-qa_samples_per_second": 62.128,
+      "eval_scitail-pairs-qa_steps_per_second": 0.485,
+      "step": 628
+    },
+    {
+      "epoch": 0.40153452685422,
+      "eval_xsum-pairs_loss": 0.017213236540555954,
+      "eval_xsum-pairs_runtime": 3.042,
+      "eval_xsum-pairs_samples_per_second": 42.078,
+      "eval_xsum-pairs_steps_per_second": 0.329,
+      "step": 628
+    },
+    {
+      "epoch": 0.40153452685422,
+      "eval_sciq_pairs_loss": 0.029452303424477577,
+      "eval_sciq_pairs_runtime": 4.5793,
+      "eval_sciq_pairs_samples_per_second": 27.952,
+      "eval_sciq_pairs_steps_per_second": 0.218,
+      "step": 628
+    },
+    {
+      "epoch": 0.40153452685422,
+      "eval_qasc_pairs_loss": 0.0720791220664978,
+      "eval_qasc_pairs_runtime": 2.0524,
+      "eval_qasc_pairs_samples_per_second": 62.366,
+      "eval_qasc_pairs_steps_per_second": 0.487,
+      "step": 628
+    },
+    {
+      "epoch": 0.40153452685422,
+      "eval_openbookqa_pairs_loss": 0.3721720278263092,
+      "eval_openbookqa_pairs_runtime": 2.1848,
+      "eval_openbookqa_pairs_samples_per_second": 58.586,
+      "eval_openbookqa_pairs_steps_per_second": 0.458,
+      "step": 628
+    },
+    {
+      "epoch": 0.40153452685422,
+      "eval_msmarco_pairs_loss": 0.09755530953407288,
+      "eval_msmarco_pairs_runtime": 2.2104,
+      "eval_msmarco_pairs_samples_per_second": 57.907,
+      "eval_msmarco_pairs_steps_per_second": 0.452,
+      "step": 628
+    },
+    {
+      "epoch": 0.40153452685422,
+      "eval_nq_pairs_loss": 0.08931688219308853,
+      "eval_nq_pairs_runtime": 3.1702,
+      "eval_nq_pairs_samples_per_second": 40.376,
+      "eval_nq_pairs_steps_per_second": 0.315,
+      "step": 628
+    },
+    {
+      "epoch": 0.40153452685422,
+      "eval_trivia_pairs_loss": 0.0646594688296318,
+      "eval_trivia_pairs_runtime": 3.2357,
+      "eval_trivia_pairs_samples_per_second": 39.558,
+      "eval_trivia_pairs_steps_per_second": 0.309,
+      "step": 628
+    },
+    {
+      "epoch": 0.40153452685422,
+      "eval_gooaq_pairs_loss": 0.1432056725025177,
+      "eval_gooaq_pairs_runtime": 2.1108,
+      "eval_gooaq_pairs_samples_per_second": 60.641,
+      "eval_gooaq_pairs_steps_per_second": 0.474,
+      "step": 628
+    },
+    {
+      "epoch": 0.40153452685422,
+      "eval_paws-pos_loss": 0.02637900970876217,
+      "eval_paws-pos_runtime": 2.1958,
+      "eval_paws-pos_samples_per_second": 58.294,
+      "eval_paws-pos_steps_per_second": 0.455,
+      "step": 628
+    },
+    {
+      "epoch": 0.40153452685422,
+      "eval_global_dataset_loss": 0.2361193746328354,
+      "eval_global_dataset_runtime": 8.5048,
+      "eval_global_dataset_samples_per_second": 38.214,
+      "eval_global_dataset_steps_per_second": 0.353,
+      "step": 628
+    },
+    {
+      "epoch": 0.4092071611253197,
+      "grad_norm": 10.842616081237793,
+      "learning_rate": 1.3439829605963792e-05,
+      "loss": 0.149,
+      "step": 640
+    },
+    {
+      "epoch": 0.4194373401534527,
+      "grad_norm": 10.382250785827637,
+      "learning_rate": 1.3780617678381259e-05,
+      "loss": 0.1402,
+      "step": 656
+    },
+    {
+      "epoch": 0.4296675191815857,
+      "grad_norm": 11.57950496673584,
+      "learning_rate": 1.4121405750798722e-05,
+      "loss": 0.1056,
+      "step": 672
+    },
+    {
+      "epoch": 0.4398976982097187,
+      "grad_norm": 5.777312278747559,
+      "learning_rate": 1.4462193823216188e-05,
+      "loss": 0.0932,
+      "step": 688
+    },
+    {
+      "epoch": 0.45012787723785164,
+      "grad_norm": 0.2746630311012268,
+      "learning_rate": 1.4802981895633655e-05,
+      "loss": 0.0534,
+      "step": 704
+    },
+    {
+      "epoch": 0.46035805626598464,
+      "grad_norm": 0.021142609417438507,
+      "learning_rate": 1.5143769968051119e-05,
+      "loss": 0.2175,
+      "step": 720
+    },
+    {
+      "epoch": 0.47058823529411764,
+      "grad_norm": 0.00019583643006626517,
+      "learning_rate": 1.5484558040468584e-05,
+      "loss": 0.1107,
+      "step": 736
+    },
+    {
+      "epoch": 0.48081841432225064,
+      "grad_norm": 18.342988967895508,
+      "learning_rate": 1.582534611288605e-05,
+      "loss": 0.2301,
+      "step": 752
+    },
+    {
+      "epoch": 0.49104859335038364,
+      "grad_norm": 4.310429096221924,
+      "learning_rate": 1.6166134185303515e-05,
+      "loss": 0.2317,
+      "step": 768
+    },
+    {
+      "epoch": 0.5012787723785166,
+      "grad_norm": 4.9908766746521,
+      "learning_rate": 1.6506922257720982e-05,
+      "loss": 0.1084,
+      "step": 784
+    },
+    {
+      "epoch": 0.5019181585677749,
+      "eval_Qnli-dev_cosine_accuracy": 0.69921875,
+      "eval_Qnli-dev_cosine_accuracy_threshold": 0.692398190498352,
+      "eval_Qnli-dev_cosine_ap": 0.7373451033369327,
+      "eval_Qnli-dev_cosine_f1": 0.7100175746924429,
+      "eval_Qnli-dev_cosine_f1_threshold": 0.6075993776321411,
+      "eval_Qnli-dev_cosine_precision": 0.6066066066066066,
+      "eval_Qnli-dev_cosine_recall": 0.8559322033898306,
+      "eval_Qnli-dev_dot_accuracy": 0.68359375,
+      "eval_Qnli-dev_dot_accuracy_threshold": 593.7620849609375,
+      "eval_Qnli-dev_dot_ap": 0.7049377659461418,
+      "eval_Qnli-dev_dot_f1": 0.7008849557522123,
+      "eval_Qnli-dev_dot_f1_threshold": 556.5305786132812,
+      "eval_Qnli-dev_dot_precision": 0.601823708206687,
+      "eval_Qnli-dev_dot_recall": 0.8389830508474576,
+      "eval_Qnli-dev_euclidean_accuracy": 0.705078125,
+      "eval_Qnli-dev_euclidean_accuracy_threshold": 24.380191802978516,
+      "eval_Qnli-dev_euclidean_ap": 0.7421878778893789,
+      "eval_Qnli-dev_euclidean_f1": 0.7024952015355086,
+      "eval_Qnli-dev_euclidean_f1_threshold": 25.691165924072266,
+      "eval_Qnli-dev_euclidean_precision": 0.6421052631578947,
+      "eval_Qnli-dev_euclidean_recall": 0.7754237288135594,
+      "eval_Qnli-dev_manhattan_accuracy": 0.705078125,
+      "eval_Qnli-dev_manhattan_accuracy_threshold": 755.8370361328125,
+      "eval_Qnli-dev_manhattan_ap": 0.7420425528192431,
+      "eval_Qnli-dev_manhattan_f1": 0.7012987012987013,
+      "eval_Qnli-dev_manhattan_f1_threshold": 872.5838623046875,
+      "eval_Qnli-dev_manhattan_precision": 0.5684210526315789,
+      "eval_Qnli-dev_manhattan_recall": 0.9152542372881356,
+      "eval_Qnli-dev_max_accuracy": 0.705078125,
+      "eval_Qnli-dev_max_accuracy_threshold": 755.8370361328125,
+      "eval_Qnli-dev_max_ap": 0.7421878778893789,
+      "eval_Qnli-dev_max_f1": 0.7100175746924429,
+      "eval_Qnli-dev_max_f1_threshold": 872.5838623046875,
+      "eval_Qnli-dev_max_precision": 0.6421052631578947,
+      "eval_Qnli-dev_max_recall": 0.9152542372881356,
+      "eval_allNLI-dev_cosine_accuracy": 0.72265625,
+      "eval_allNLI-dev_cosine_accuracy_threshold": 0.8044645190238953,
+      "eval_allNLI-dev_cosine_ap": 0.6029077511537212,
+      "eval_allNLI-dev_cosine_f1": 0.6193548387096773,
+      "eval_allNLI-dev_cosine_f1_threshold": 0.6631404161453247,
+      "eval_allNLI-dev_cosine_precision": 0.4931506849315068,
+      "eval_allNLI-dev_cosine_recall": 0.8323699421965318,
+      "eval_allNLI-dev_dot_accuracy": 0.716796875,
+      "eval_allNLI-dev_dot_accuracy_threshold": 788.6072998046875,
+      "eval_allNLI-dev_dot_ap": 0.5937600770830995,
+      "eval_allNLI-dev_dot_f1": 0.6299559471365639,
+      "eval_allNLI-dev_dot_f1_threshold": 632.650634765625,
+      "eval_allNLI-dev_dot_precision": 0.5088967971530249,
+      "eval_allNLI-dev_dot_recall": 0.8265895953757225,
+      "eval_allNLI-dev_euclidean_accuracy": 0.71875,
+      "eval_allNLI-dev_euclidean_accuracy_threshold": 19.220054626464844,
+      "eval_allNLI-dev_euclidean_ap": 0.6000308340748244,
+      "eval_allNLI-dev_euclidean_f1": 0.6140350877192983,
+      "eval_allNLI-dev_euclidean_f1_threshold": 24.999332427978516,
+      "eval_allNLI-dev_euclidean_precision": 0.49469964664310956,
+      "eval_allNLI-dev_euclidean_recall": 0.8092485549132948,
+      "eval_allNLI-dev_manhattan_accuracy": 0.724609375,
+      "eval_allNLI-dev_manhattan_accuracy_threshold": 601.0503540039062,
+      "eval_allNLI-dev_manhattan_ap": 0.5994463454270988,
+      "eval_allNLI-dev_manhattan_f1": 0.6137339055793991,
+      "eval_allNLI-dev_manhattan_f1_threshold": 785.221435546875,
+      "eval_allNLI-dev_manhattan_precision": 0.4880546075085324,
+      "eval_allNLI-dev_manhattan_recall": 0.8265895953757225,
+      "eval_allNLI-dev_max_accuracy": 0.724609375,
+      "eval_allNLI-dev_max_accuracy_threshold": 788.6072998046875,
+      "eval_allNLI-dev_max_ap": 0.6029077511537212,
+      "eval_allNLI-dev_max_f1": 0.6299559471365639,
+      "eval_allNLI-dev_max_f1_threshold": 785.221435546875,
+      "eval_allNLI-dev_max_precision": 0.5088967971530249,
+      "eval_allNLI-dev_max_recall": 0.8323699421965318,
+      "eval_sequential_score": 0.7421878778893789,
+      "eval_sts-test_pearson_cosine": 0.9043291123731831,
+      "eval_sts-test_pearson_dot": 0.8879215841003365,
+      "eval_sts-test_pearson_euclidean": 0.9229060512014216,
+      "eval_sts-test_pearson_manhattan": 0.9223094827934384,
+      "eval_sts-test_pearson_max": 0.9229060512014216,
+      "eval_sts-test_spearman_cosine": 0.9171074245027466,
+      "eval_sts-test_spearman_dot": 0.8910804041363384,
+      "eval_sts-test_spearman_euclidean": 0.9186099481039627,
+      "eval_sts-test_spearman_manhattan": 0.9180492147987285,
+      "eval_sts-test_spearman_max": 0.9186099481039627,
+      "eval_vitaminc-pairs_loss": 3.900864362716675,
+      "eval_vitaminc-pairs_runtime": 4.4894,
+      "eval_vitaminc-pairs_samples_per_second": 28.512,
+      "eval_vitaminc-pairs_steps_per_second": 0.223,
+      "step": 785
+    },
+    {
+      "epoch": 0.5019181585677749,
+      "eval_negation-triplets_loss": 0.633728563785553,
+      "eval_negation-triplets_runtime": 3.4005,
+      "eval_negation-triplets_samples_per_second": 37.641,
+      "eval_negation-triplets_steps_per_second": 0.294,
+      "step": 785
+    },
+    {
+      "epoch": 0.5019181585677749,
+      "eval_scitail-pairs-pos_loss": 0.012064059264957905,
+      "eval_scitail-pairs-pos_runtime": 2.6615,
+      "eval_scitail-pairs-pos_samples_per_second": 48.092,
+      "eval_scitail-pairs-pos_steps_per_second": 0.376,
+      "step": 785
+    },
+    {
+      "epoch": 0.5019181585677749,
+      "eval_scitail-pairs-qa_loss": 7.389660459011793e-05,
+      "eval_scitail-pairs-qa_runtime": 2.0934,
+      "eval_scitail-pairs-qa_samples_per_second": 61.146,
+      "eval_scitail-pairs-qa_steps_per_second": 0.478,
+      "step": 785
+    },
+    {
+      "epoch": 0.5019181585677749,
+      "eval_xsum-pairs_loss": 0.009806518442928791,
+      "eval_xsum-pairs_runtime": 3.0655,
+      "eval_xsum-pairs_samples_per_second": 41.755,
+      "eval_xsum-pairs_steps_per_second": 0.326,
+      "step": 785
+    },
+    {
+      "epoch": 0.5019181585677749,
+      "eval_sciq_pairs_loss": 0.03242116793990135,
+      "eval_sciq_pairs_runtime": 4.6405,
+      "eval_sciq_pairs_samples_per_second": 27.583,
+      "eval_sciq_pairs_steps_per_second": 0.215,
+      "step": 785
+    },
+    {
+      "epoch": 0.5019181585677749,
+      "eval_qasc_pairs_loss": 0.07289526611566544,
+      "eval_qasc_pairs_runtime": 2.0997,
+      "eval_qasc_pairs_samples_per_second": 60.962,
+      "eval_qasc_pairs_steps_per_second": 0.476,
+      "step": 785
+    },
+    {
+      "epoch": 0.5019181585677749,
+      "eval_openbookqa_pairs_loss": 0.41281530261039734,
+      "eval_openbookqa_pairs_runtime": 2.2362,
+      "eval_openbookqa_pairs_samples_per_second": 57.24,
+      "eval_openbookqa_pairs_steps_per_second": 0.447,
+      "step": 785
+    },
+    {
+      "epoch": 0.5019181585677749,
+      "eval_msmarco_pairs_loss": 0.11559046804904938,
+      "eval_msmarco_pairs_runtime": 2.2303,
+      "eval_msmarco_pairs_samples_per_second": 57.393,
+      "eval_msmarco_pairs_steps_per_second": 0.448,
+      "step": 785
+    },
+    {
+      "epoch": 0.5019181585677749,
+      "eval_nq_pairs_loss": 0.10707583278417587,
+      "eval_nq_pairs_runtime": 3.2248,
+      "eval_nq_pairs_samples_per_second": 39.693,
+      "eval_nq_pairs_steps_per_second": 0.31,
+      "step": 785
+    },
+    {
+      "epoch": 0.5019181585677749,
+      "eval_trivia_pairs_loss": 0.09321507811546326,
+      "eval_trivia_pairs_runtime": 3.2802,
+      "eval_trivia_pairs_samples_per_second": 39.022,
+      "eval_trivia_pairs_steps_per_second": 0.305,
+      "step": 785
+    },
+    {
+      "epoch": 0.5019181585677749,
+      "eval_gooaq_pairs_loss": 0.1236836165189743,
+      "eval_gooaq_pairs_runtime": 2.166,
+      "eval_gooaq_pairs_samples_per_second": 59.094,
+      "eval_gooaq_pairs_steps_per_second": 0.462,
+      "step": 785
+    },
+    {
+      "epoch": 0.5019181585677749,
+      "eval_paws-pos_loss": 0.025379396975040436,
+      "eval_paws-pos_runtime": 2.2622,
+      "eval_paws-pos_samples_per_second": 56.581,
+      "eval_paws-pos_steps_per_second": 0.442,
+      "step": 785
+    },
+    {
+      "epoch": 0.5019181585677749,
+      "eval_global_dataset_loss": 0.2525935471057892,
+      "eval_global_dataset_runtime": 8.5173,
+      "eval_global_dataset_samples_per_second": 38.158,
+      "eval_global_dataset_steps_per_second": 0.352,
+      "step": 785
+    },
+    {
+      "epoch": 0.5115089514066496,
+      "grad_norm": 2.781236171722412,
+      "learning_rate": 1.6847710330138445e-05,
+      "loss": 0.0799,
+      "step": 800
+    },
+    {
+      "epoch": 0.5217391304347826,
+      "grad_norm": 4.296014785766602,
+      "learning_rate": 1.7188498402555913e-05,
+      "loss": 0.1466,
+      "step": 816
+    },
+    {
+      "epoch": 0.5319693094629157,
+      "grad_norm": 0.6779389977455139,
+      "learning_rate": 1.7529286474973376e-05,
+      "loss": 0.1208,
+      "step": 832
+    },
+    {
+      "epoch": 0.5421994884910486,
+      "grad_norm": 5.031887054443359,
+      "learning_rate": 1.7870074547390843e-05,
+      "loss": 0.1021,
+      "step": 848
+    },
+    {
+      "epoch": 0.5524296675191815,
+      "grad_norm": 3.492619276046753,
+      "learning_rate": 1.8210862619808307e-05,
+      "loss": 0.1391,
+      "step": 864
+    },
+    {
+      "epoch": 0.5626598465473146,
+      "grad_norm": 3.4289321899414062,
+      "learning_rate": 1.8551650692225774e-05,
+      "loss": 0.185,
+      "step": 880
+    },
+    {
+      "epoch": 0.5728900255754475,
+      "grad_norm": 0.48259153962135315,
+      "learning_rate": 1.8892438764643238e-05,
+      "loss": 0.1108,
+      "step": 896
+    },
+    {
+      "epoch": 0.5831202046035806,
+      "grad_norm": 1.390660047531128,
+      "learning_rate": 1.9233226837060705e-05,
+      "loss": 0.0926,
+      "step": 912
+    },
+    {
+      "epoch": 0.5933503836317136,
+      "grad_norm": 11.370463371276855,
+      "learning_rate": 1.957401490947817e-05,
+      "loss": 0.157,
+      "step": 928
+    },
+    {
+      "epoch": 0.6023017902813299,
+      "eval_Qnli-dev_cosine_accuracy": 0.6796875,
+      "eval_Qnli-dev_cosine_accuracy_threshold": 0.7110848426818848,
+      "eval_Qnli-dev_cosine_ap": 0.7098259706381642,
+      "eval_Qnli-dev_cosine_f1": 0.6848,
+      "eval_Qnli-dev_cosine_f1_threshold": 0.5396614074707031,
+      "eval_Qnli-dev_cosine_precision": 0.5501285347043702,
+      "eval_Qnli-dev_cosine_recall": 0.9067796610169492,
+      "eval_Qnli-dev_dot_accuracy": 0.66015625,
+      "eval_Qnli-dev_dot_accuracy_threshold": 612.0003662109375,
+      "eval_Qnli-dev_dot_ap": 0.6677871203032961,
+      "eval_Qnli-dev_dot_f1": 0.6847133757961782,
+      "eval_Qnli-dev_dot_f1_threshold": 489.4056091308594,
+      "eval_Qnli-dev_dot_precision": 0.548469387755102,
+      "eval_Qnli-dev_dot_recall": 0.9110169491525424,
+      "eval_Qnli-dev_euclidean_accuracy": 0.689453125,
+      "eval_Qnli-dev_euclidean_accuracy_threshold": 22.8848876953125,
+      "eval_Qnli-dev_euclidean_ap": 0.7192898440461163,
+      "eval_Qnli-dev_euclidean_f1": 0.6879194630872484,
+      "eval_Qnli-dev_euclidean_f1_threshold": 28.238075256347656,
+      "eval_Qnli-dev_euclidean_precision": 0.5694444444444444,
+      "eval_Qnli-dev_euclidean_recall": 0.8686440677966102,
+      "eval_Qnli-dev_manhattan_accuracy": 0.681640625,
+      "eval_Qnli-dev_manhattan_accuracy_threshold": 722.6270751953125,
+      "eval_Qnli-dev_manhattan_ap": 0.7187658029633445,
+      "eval_Qnli-dev_manhattan_f1": 0.6911519198664441,
+      "eval_Qnli-dev_manhattan_f1_threshold": 881.6773681640625,
+      "eval_Qnli-dev_manhattan_precision": 0.5702479338842975,
+      "eval_Qnli-dev_manhattan_recall": 0.8771186440677966,
+      "eval_Qnli-dev_max_accuracy": 0.689453125,
+      "eval_Qnli-dev_max_accuracy_threshold": 722.6270751953125,
+      "eval_Qnli-dev_max_ap": 0.7192898440461163,
+      "eval_Qnli-dev_max_f1": 0.6911519198664441,
+      "eval_Qnli-dev_max_f1_threshold": 881.6773681640625,
+      "eval_Qnli-dev_max_precision": 0.5702479338842975,
+      "eval_Qnli-dev_max_recall": 0.9110169491525424,
+      "eval_allNLI-dev_cosine_accuracy": 0.720703125,
+      "eval_allNLI-dev_cosine_accuracy_threshold": 0.8192055225372314,
+      "eval_allNLI-dev_cosine_ap": 0.6064453769503819,
+      "eval_allNLI-dev_cosine_f1": 0.6255924170616114,
+      "eval_allNLI-dev_cosine_f1_threshold": 0.6535474061965942,
+      "eval_allNLI-dev_cosine_precision": 0.5301204819277109,
+      "eval_allNLI-dev_cosine_recall": 0.7630057803468208,
+      "eval_allNLI-dev_dot_accuracy": 0.71875,
+      "eval_allNLI-dev_dot_accuracy_threshold": 818.2044067382812,
+      "eval_allNLI-dev_dot_ap": 0.6065404649281851,
+      "eval_allNLI-dev_dot_f1": 0.6232558139534884,
+      "eval_allNLI-dev_dot_f1_threshold": 617.462158203125,
+      "eval_allNLI-dev_dot_precision": 0.5214007782101168,
+      "eval_allNLI-dev_dot_recall": 0.7745664739884393,
+      "eval_allNLI-dev_euclidean_accuracy": 0.71484375,
+      "eval_allNLI-dev_euclidean_accuracy_threshold": 18.268875122070312,
+      "eval_allNLI-dev_euclidean_ap": 0.6020313861913525,
+      "eval_allNLI-dev_euclidean_f1": 0.6205128205128205,
+      "eval_allNLI-dev_euclidean_f1_threshold": 24.51853370666504,
+      "eval_allNLI-dev_euclidean_precision": 0.5576036866359447,
+      "eval_allNLI-dev_euclidean_recall": 0.6994219653179191,
+      "eval_allNLI-dev_manhattan_accuracy": 0.720703125,
+      "eval_allNLI-dev_manhattan_accuracy_threshold": 748.0535888671875,
+      "eval_allNLI-dev_manhattan_ap": 0.6018544507527778,
+      "eval_allNLI-dev_manhattan_f1": 0.62,
+      "eval_allNLI-dev_manhattan_f1_threshold": 774.2725219726562,
+      "eval_allNLI-dev_manhattan_precision": 0.5462555066079295,
+      "eval_allNLI-dev_manhattan_recall": 0.7167630057803468,
+      "eval_allNLI-dev_max_accuracy": 0.720703125,
+      "eval_allNLI-dev_max_accuracy_threshold": 818.2044067382812,
+      "eval_allNLI-dev_max_ap": 0.6065404649281851,
+      "eval_allNLI-dev_max_f1": 0.6255924170616114,
+      "eval_allNLI-dev_max_f1_threshold": 774.2725219726562,
+      "eval_allNLI-dev_max_precision": 0.5576036866359447,
+      "eval_allNLI-dev_max_recall": 0.7745664739884393,
+      "eval_sequential_score": 0.7192898440461163,
+      "eval_sts-test_pearson_cosine": 0.917151106536728,
+      "eval_sts-test_pearson_dot": 0.9064091595160165,
+      "eval_sts-test_pearson_euclidean": 0.9276686994987783,
+      "eval_sts-test_pearson_manhattan": 0.9272998345544423,
+      "eval_sts-test_pearson_max": 0.9276686994987783,
+      "eval_sts-test_spearman_cosine": 0.9217884511955483,
+      "eval_sts-test_spearman_dot": 0.9049310090394994,
+      "eval_sts-test_spearman_euclidean": 0.9226553278799612,
+      "eval_sts-test_spearman_manhattan": 0.9221132558507271,
+      "eval_sts-test_spearman_max": 0.9226553278799612,
+      "eval_vitaminc-pairs_loss": 2.814775228500366,
+      "eval_vitaminc-pairs_runtime": 4.4847,
+      "eval_vitaminc-pairs_samples_per_second": 28.542,
+      "eval_vitaminc-pairs_steps_per_second": 0.223,
+      "step": 942
+    },
+    {
+      "epoch": 0.6023017902813299,
+      "eval_negation-triplets_loss": 0.5488928556442261,
+      "eval_negation-triplets_runtime": 3.2998,
+      "eval_negation-triplets_samples_per_second": 38.79,
+      "eval_negation-triplets_steps_per_second": 0.303,
+      "step": 942
+    },
+    {
+      "epoch": 0.6023017902813299,
+      "eval_scitail-pairs-pos_loss": 0.0049793594516813755,
+      "eval_scitail-pairs-pos_runtime": 2.6898,
+      "eval_scitail-pairs-pos_samples_per_second": 47.588,
+      "eval_scitail-pairs-pos_steps_per_second": 0.372,
+      "step": 942
+    },
+    {
+      "epoch": 0.6023017902813299,
+      "eval_scitail-pairs-qa_loss": 0.0024724495597183704,
+      "eval_scitail-pairs-qa_runtime": 2.0719,
+      "eval_scitail-pairs-qa_samples_per_second": 61.778,
+      "eval_scitail-pairs-qa_steps_per_second": 0.483,
+      "step": 942
+    },
+    {
+      "epoch": 0.6023017902813299,
+      "eval_xsum-pairs_loss": 0.01770014315843582,
+      "eval_xsum-pairs_runtime": 3.0541,
+      "eval_xsum-pairs_samples_per_second": 41.911,
+      "eval_xsum-pairs_steps_per_second": 0.327,
+      "step": 942
+    },
+    {
+      "epoch": 0.6023017902813299,
+      "eval_sciq_pairs_loss": 0.030735233798623085,
+      "eval_sciq_pairs_runtime": 4.5905,
+      "eval_sciq_pairs_samples_per_second": 27.883,
+      "eval_sciq_pairs_steps_per_second": 0.218,
+      "step": 942
+    },
+    {
+      "epoch": 0.6023017902813299,
+      "eval_qasc_pairs_loss": 0.04597622901201248,
+      "eval_qasc_pairs_runtime": 2.1506,
+      "eval_qasc_pairs_samples_per_second": 59.518,
+      "eval_qasc_pairs_steps_per_second": 0.465,
+      "step": 942
+    },
+    {
+      "epoch": 0.6023017902813299,
+      "eval_openbookqa_pairs_loss": 0.3580770492553711,
+      "eval_openbookqa_pairs_runtime": 2.2167,
+      "eval_openbookqa_pairs_samples_per_second": 57.743,
+      "eval_openbookqa_pairs_steps_per_second": 0.451,
+      "step": 942
+    },
+    {
+      "epoch": 0.6023017902813299,
+      "eval_msmarco_pairs_loss": 0.08442261815071106,
+      "eval_msmarco_pairs_runtime": 2.2308,
+      "eval_msmarco_pairs_samples_per_second": 57.377,
+      "eval_msmarco_pairs_steps_per_second": 0.448,
+      "step": 942
+    },
+    {
+      "epoch": 0.6023017902813299,
+      "eval_nq_pairs_loss": 0.12354034185409546,
+      "eval_nq_pairs_runtime": 3.1805,
+      "eval_nq_pairs_samples_per_second": 40.245,
+      "eval_nq_pairs_steps_per_second": 0.314,
+      "step": 942
+    },
+    {
+      "epoch": 0.6023017902813299,
+      "eval_trivia_pairs_loss": 0.08584598451852798,
+      "eval_trivia_pairs_runtime": 3.2829,
+      "eval_trivia_pairs_samples_per_second": 38.989,
+      "eval_trivia_pairs_steps_per_second": 0.305,
+      "step": 942
+    },
+    {
+      "epoch": 0.6023017902813299,
+      "eval_gooaq_pairs_loss": 0.1563483476638794,
+      "eval_gooaq_pairs_runtime": 2.1458,
+      "eval_gooaq_pairs_samples_per_second": 59.651,
+      "eval_gooaq_pairs_steps_per_second": 0.466,
+      "step": 942
+    },
+    {
+      "epoch": 0.6023017902813299,
+      "eval_paws-pos_loss": 0.026773599907755852,
+      "eval_paws-pos_runtime": 2.2412,
+      "eval_paws-pos_samples_per_second": 57.112,
+      "eval_paws-pos_steps_per_second": 0.446,
+      "step": 942
+    },
+    {
+      "epoch": 0.6023017902813299,
+      "eval_global_dataset_loss": 0.15439902245998383,
+      "eval_global_dataset_runtime": 8.5471,
+      "eval_global_dataset_samples_per_second": 38.025,
+      "eval_global_dataset_steps_per_second": 0.351,
+      "step": 942
+    },
+    {
+      "epoch": 0.6035805626598465,
+      "grad_norm": 5.923606872558594,
+      "learning_rate": 1.9914802981895636e-05,
+      "loss": 0.126,
+      "step": 944
+    },
+    {
+      "epoch": 0.6138107416879796,
+      "grad_norm": 2.5291152000427246,
+      "learning_rate": 1.9998887778598045e-05,
+      "loss": 0.1416,
+      "step": 960
+    },
+    {
+      "epoch": 0.6240409207161125,
+      "grad_norm": 6.44370174407959,
+      "learning_rate": 1.9993945237540454e-05,
+      "loss": 0.157,
+      "step": 976
+    },
+    {
+      "epoch": 0.6342710997442456,
+      "grad_norm": 5.631255149841309,
+      "learning_rate": 1.9985051399986202e-05,
+      "loss": 0.076,
+      "step": 992
+    },
+    {
+      "epoch": 0.6445012787723785,
+      "grad_norm": 3.087242364883423,
+      "learning_rate": 1.9972210955346275e-05,
+      "loss": 0.0956,
+      "step": 1008
+    },
+    {
+      "epoch": 0.6547314578005116,
+      "grad_norm": 5.553103446960449,
+      "learning_rate": 1.995543067394022e-05,
+      "loss": 0.1297,
+      "step": 1024
+    },
+    {
+      "epoch": 0.6649616368286445,
+      "grad_norm": 4.881236553192139,
+      "learning_rate": 1.9934719403426386e-05,
+      "loss": 0.1673,
+      "step": 1040
+    },
+    {
+      "epoch": 0.6751918158567775,
+      "grad_norm": 9.578516006469727,
+      "learning_rate": 1.9910088064136855e-05,
+      "loss": 0.0801,
+      "step": 1056
+    },
+    {
+      "epoch": 0.6854219948849105,
+      "grad_norm": 2.1746671199798584,
+      "learning_rate": 1.9881549643319548e-05,
+      "loss": 0.1508,
+      "step": 1072
+    },
+    {
+      "epoch": 0.6956521739130435,
+      "grad_norm": 1.0966675281524658,
+      "learning_rate": 1.9849119188290472e-05,
+      "loss": 0.082,
+      "step": 1088
+    },
+    {
+      "epoch": 0.7026854219948849,
+      "eval_Qnli-dev_cosine_accuracy": 0.701171875,
+      "eval_Qnli-dev_cosine_accuracy_threshold": 0.6949167251586914,
+      "eval_Qnli-dev_cosine_ap": 0.7347660327067432,
+      "eval_Qnli-dev_cosine_f1": 0.6996336996336996,
+      "eval_Qnli-dev_cosine_f1_threshold": 0.6012573838233948,
+      "eval_Qnli-dev_cosine_precision": 0.6161290322580645,
+      "eval_Qnli-dev_cosine_recall": 0.809322033898305,
+      "eval_Qnli-dev_dot_accuracy": 0.681640625,
+      "eval_Qnli-dev_dot_accuracy_threshold": 601.9920654296875,
+      "eval_Qnli-dev_dot_ap": 0.6929187579311147,
+      "eval_Qnli-dev_dot_f1": 0.6769759450171821,
+      "eval_Qnli-dev_dot_f1_threshold": 541.3872680664062,
+      "eval_Qnli-dev_dot_precision": 0.569364161849711,
+      "eval_Qnli-dev_dot_recall": 0.8347457627118644,
+      "eval_Qnli-dev_euclidean_accuracy": 0.705078125,
+      "eval_Qnli-dev_euclidean_accuracy_threshold": 24.72307586669922,
+      "eval_Qnli-dev_euclidean_ap": 0.7433711154783218,
+      "eval_Qnli-dev_euclidean_f1": 0.6951871657754012,
+      "eval_Qnli-dev_euclidean_f1_threshold": 27.680713653564453,
+      "eval_Qnli-dev_euclidean_precision": 0.6,
+      "eval_Qnli-dev_euclidean_recall": 0.826271186440678,
+      "eval_Qnli-dev_manhattan_accuracy": 0.705078125,
+      "eval_Qnli-dev_manhattan_accuracy_threshold": 771.5457763671875,
+      "eval_Qnli-dev_manhattan_ap": 0.7417925203500755,
+      "eval_Qnli-dev_manhattan_f1": 0.7015706806282722,
+      "eval_Qnli-dev_manhattan_f1_threshold": 873.9434814453125,
+      "eval_Qnli-dev_manhattan_precision": 0.5964391691394659,
+      "eval_Qnli-dev_manhattan_recall": 0.8516949152542372,
+      "eval_Qnli-dev_max_accuracy": 0.705078125,
+      "eval_Qnli-dev_max_accuracy_threshold": 771.5457763671875,
+      "eval_Qnli-dev_max_ap": 0.7433711154783218,
+      "eval_Qnli-dev_max_f1": 0.7015706806282722,
+      "eval_Qnli-dev_max_f1_threshold": 873.9434814453125,
+      "eval_Qnli-dev_max_precision": 0.6161290322580645,
+      "eval_Qnli-dev_max_recall": 0.8516949152542372,
+      "eval_allNLI-dev_cosine_accuracy": 0.71484375,
+      "eval_allNLI-dev_cosine_accuracy_threshold": 0.7252209186553955,
+      "eval_allNLI-dev_cosine_ap": 0.5940767502246462,
+      "eval_allNLI-dev_cosine_f1": 0.6111111111111112,
+      "eval_allNLI-dev_cosine_f1_threshold": 0.6290402412414551,
+      "eval_allNLI-dev_cosine_precision": 0.4847457627118644,
+      "eval_allNLI-dev_cosine_recall": 0.8265895953757225,
+      "eval_allNLI-dev_dot_accuracy": 0.712890625,
+      "eval_allNLI-dev_dot_accuracy_threshold": 814.8841552734375,
+      "eval_allNLI-dev_dot_ap": 0.5890724491285201,
+      "eval_allNLI-dev_dot_f1": 0.6238938053097345,
+      "eval_allNLI-dev_dot_f1_threshold": 624.3211669921875,
+      "eval_allNLI-dev_dot_precision": 0.5053763440860215,
+      "eval_allNLI-dev_dot_recall": 0.815028901734104,
+      "eval_allNLI-dev_euclidean_accuracy": 0.712890625,
+      "eval_allNLI-dev_euclidean_accuracy_threshold": 21.297210693359375,
+      "eval_allNLI-dev_euclidean_ap": 0.5885663962477001,
+      "eval_allNLI-dev_euclidean_f1": 0.6053169734151329,
+      "eval_allNLI-dev_euclidean_f1_threshold": 27.594894409179688,
+      "eval_allNLI-dev_euclidean_precision": 0.46835443037974683,
+      "eval_allNLI-dev_euclidean_recall": 0.8554913294797688,
+      "eval_allNLI-dev_manhattan_accuracy": 0.7109375,
+      "eval_allNLI-dev_manhattan_accuracy_threshold": 655.72802734375,
+      "eval_allNLI-dev_manhattan_ap": 0.5903691578356358,
+      "eval_allNLI-dev_manhattan_f1": 0.6157024793388429,
+      "eval_allNLI-dev_manhattan_f1_threshold": 858.11962890625,
+      "eval_allNLI-dev_manhattan_precision": 0.4790996784565916,
+      "eval_allNLI-dev_manhattan_recall": 0.861271676300578,
+      "eval_allNLI-dev_max_accuracy": 0.71484375,
+      "eval_allNLI-dev_max_accuracy_threshold": 814.8841552734375,
+      "eval_allNLI-dev_max_ap": 0.5940767502246462,
+      "eval_allNLI-dev_max_f1": 0.6238938053097345,
+      "eval_allNLI-dev_max_f1_threshold": 858.11962890625,
+      "eval_allNLI-dev_max_precision": 0.5053763440860215,
+      "eval_allNLI-dev_max_recall": 0.861271676300578,
+      "eval_sequential_score": 0.7433711154783218,
+      "eval_sts-test_pearson_cosine": 0.9045402318307136,
+      "eval_sts-test_pearson_dot": 0.8756384139446844,
+      "eval_sts-test_pearson_euclidean": 0.9164304447060423,
+      "eval_sts-test_pearson_manhattan": 0.9157247908119455,
+      "eval_sts-test_pearson_max": 0.9164304447060423,
+      "eval_sts-test_spearman_cosine": 0.9084125676732273,
+      "eval_sts-test_spearman_dot": 0.8658350988039152,
+      "eval_sts-test_spearman_euclidean": 0.9116794097547996,
+      "eval_sts-test_spearman_manhattan": 0.9112685036727037,
+      "eval_sts-test_spearman_max": 0.9116794097547996,
+      "eval_vitaminc-pairs_loss": 4.510133266448975,
+      "eval_vitaminc-pairs_runtime": 4.4836,
+      "eval_vitaminc-pairs_samples_per_second": 28.548,
+      "eval_vitaminc-pairs_steps_per_second": 0.223,
+      "step": 1099
+    },
+    {
+      "epoch": 0.7026854219948849,
+      "eval_negation-triplets_loss": 0.6731557846069336,
+      "eval_negation-triplets_runtime": 3.3513,
+      "eval_negation-triplets_samples_per_second": 38.194,
+      "eval_negation-triplets_steps_per_second": 0.298,
+      "step": 1099
+    },
+    {
+      "epoch": 0.7026854219948849,
+      "eval_scitail-pairs-pos_loss": 0.026047488674521446,
+      "eval_scitail-pairs-pos_runtime": 2.666,
+      "eval_scitail-pairs-pos_samples_per_second": 48.012,
+      "eval_scitail-pairs-pos_steps_per_second": 0.375,
+      "step": 1099
+    },
+    {
+      "epoch": 0.7026854219948849,
+      "eval_scitail-pairs-qa_loss": 6.3366285758093e-05,
+      "eval_scitail-pairs-qa_runtime": 2.1132,
+      "eval_scitail-pairs-qa_samples_per_second": 60.571,
+      "eval_scitail-pairs-qa_steps_per_second": 0.473,
+      "step": 1099
+    },
+    {
+      "epoch": 0.7026854219948849,
+      "eval_xsum-pairs_loss": 0.010141758248209953,
+      "eval_xsum-pairs_runtime": 3.077,
+      "eval_xsum-pairs_samples_per_second": 41.599,
+      "eval_xsum-pairs_steps_per_second": 0.325,
+      "step": 1099
+    },
+    {
+      "epoch": 0.7026854219948849,
+      "eval_sciq_pairs_loss": 0.03606099635362625,
+      "eval_sciq_pairs_runtime": 4.6134,
+      "eval_sciq_pairs_samples_per_second": 27.745,
+      "eval_sciq_pairs_steps_per_second": 0.217,
+      "step": 1099
+    },
+    {
+      "epoch": 0.7026854219948849,
+      "eval_qasc_pairs_loss": 0.0794580727815628,
+      "eval_qasc_pairs_runtime": 2.0891,
+      "eval_qasc_pairs_samples_per_second": 61.27,
+      "eval_qasc_pairs_steps_per_second": 0.479,
+      "step": 1099
+    },
+    {
+      "epoch": 0.7026854219948849,
+      "eval_openbookqa_pairs_loss": 0.4053290784358978,
+      "eval_openbookqa_pairs_runtime": 2.2366,
+      "eval_openbookqa_pairs_samples_per_second": 57.229,
+      "eval_openbookqa_pairs_steps_per_second": 0.447,
+      "step": 1099
+    },
+    {
+      "epoch": 0.7026854219948849,
+      "eval_msmarco_pairs_loss": 0.08789154142141342,
+      "eval_msmarco_pairs_runtime": 2.2396,
+      "eval_msmarco_pairs_samples_per_second": 57.153,
+      "eval_msmarco_pairs_steps_per_second": 0.447,
+      "step": 1099
+    },
+    {
+      "epoch": 0.7026854219948849,
+      "eval_nq_pairs_loss": 0.07979367673397064,
+      "eval_nq_pairs_runtime": 3.1969,
+      "eval_nq_pairs_samples_per_second": 40.039,
+      "eval_nq_pairs_steps_per_second": 0.313,
+      "step": 1099
+    },
+    {
+      "epoch": 0.7026854219948849,
+      "eval_trivia_pairs_loss": 0.08963163197040558,
+      "eval_trivia_pairs_runtime": 3.2784,
+      "eval_trivia_pairs_samples_per_second": 39.044,
+      "eval_trivia_pairs_steps_per_second": 0.305,
+      "step": 1099
+    },
+    {
+      "epoch": 0.7026854219948849,
+      "eval_gooaq_pairs_loss": 0.1373358964920044,
+      "eval_gooaq_pairs_runtime": 2.1632,
+      "eval_gooaq_pairs_samples_per_second": 59.172,
+      "eval_gooaq_pairs_steps_per_second": 0.462,
+      "step": 1099
+    },
+    {
+      "epoch": 0.7026854219948849,
+      "eval_paws-pos_loss": 0.025509441271424294,
+      "eval_paws-pos_runtime": 2.2614,
+      "eval_paws-pos_samples_per_second": 56.603,
+      "eval_paws-pos_steps_per_second": 0.442,
+      "step": 1099
+    },
+    {
+      "epoch": 0.7026854219948849,
+      "eval_global_dataset_loss": 0.3096943199634552,
+      "eval_global_dataset_runtime": 8.5451,
+      "eval_global_dataset_samples_per_second": 38.034,
+      "eval_global_dataset_steps_per_second": 0.351,
+      "step": 1099
+    }
+  ],
+  "logging_steps": 16,
+  "max_steps": 3128,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 2,
+  "save_steps": 157,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 0.0,
+  "train_batch_size": 64,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-1099/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ae5d47b05b44fe0990966324c4a681d83a347e47ed07c3ca3b6e4479e0e161a9
+size 5560