Upload folder using huggingface_hub

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text

adapter_config.json

@@ -0,0 +1,22 @@
+{
+  "config": {
+    "architecture": "prefix_tuning",
+    "bottleneck_size": 512,
+    "cross_prefix": true,
+    "dropout": 0.0,
+    "encoder_prefix": true,
+    "flat": false,
+    "leave_out": [],
+    "non_linearity": "tanh",
+    "prefix_length": 30,
+    "shared_gating": true,
+    "use_gating": false
+  },
+  "config_id": "648bf22f5afeaaa6",
+  "hidden_size": 768,
+  "model_class": "DebertaV2ForMaskedLM",
+  "model_name": "microsoft/mdeberta-v3-base",
+  "model_type": "deberta-v2",
+  "name": "cc100",
+  "version": "adapters.1.0.0"
+}

added_tokens.json

@@ -0,0 +1,3 @@
+{
+  "[MASK]": 250101
+}

head_config.json

@@ -0,0 +1,14 @@
+{
+  "config": null,
+  "hidden_size": 768,
+  "label2id": {
+    "LABEL_0": 0,
+    "LABEL_1": 1
+  },
+  "model_class": "DebertaV2ForMaskedLM",
+  "model_name": "microsoft/mdeberta-v3-base",
+  "model_type": "deberta-v2",
+  "name": null,
+  "num_labels": 2,
+  "version": "adapters.1.0.0"
+}

optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:83c9019fd2974ffc2087f02444977646a5c275fc08ce0813d32642584b775822
+size 83724291

pytorch_adapter.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ed4d2535ec788ce4d8bd2ce26dc66b40e2f549e99a3f992a5f4c6b90350a4450
+size 39503700

pytorch_model_head.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3a5de2d70086d60f46125df1486ff32351589bd247225b1e2191a1523673a8ca
+size 774447414

rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:daec9089091a3cfc2b7dbb5bd45a215431d1c31e549676ccd1f5ff1eb7eeafd4
+size 14244

scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b1779cc7b17c6f5ed33c2a8a77c4ae41b3c2418eee1200bcb469135e32da88b2
+size 1064

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
+  }
+}

spm.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:13c8d666d62a7bc4ac8f040aab68e942c861f93303156cc28f5c7e885d86d6e3
+size 4305025

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9ddbe35b768b22d6cd0d61a0a88ea3a188b7c00bbb7b825f9f247cf0b79c2365
+size 16350764

tokenizer_config.json

@@ -0,0 +1,858 @@
+{
+  "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
+    },
+    "250001": {
+      "content": "▁<extra_id_99>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250002": {
+      "content": "▁<extra_id_98>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250003": {
+      "content": "▁<extra_id_97>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250004": {
+      "content": "▁<extra_id_96>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250005": {
+      "content": "▁<extra_id_95>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250006": {
+      "content": "▁<extra_id_94>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250007": {
+      "content": "▁<extra_id_93>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250008": {
+      "content": "▁<extra_id_92>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250009": {
+      "content": "▁<extra_id_91>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250010": {
+      "content": "▁<extra_id_90>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250011": {
+      "content": "▁<extra_id_89>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250012": {
+      "content": "▁<extra_id_88>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250013": {
+      "content": "▁<extra_id_87>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250014": {
+      "content": "▁<extra_id_86>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250015": {
+      "content": "▁<extra_id_85>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250016": {
+      "content": "▁<extra_id_84>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250017": {
+      "content": "▁<extra_id_83>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250018": {
+      "content": "▁<extra_id_82>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250019": {
+      "content": "▁<extra_id_81>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250020": {
+      "content": "▁<extra_id_80>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250021": {
+      "content": "▁<extra_id_79>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250022": {
+      "content": "▁<extra_id_78>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250023": {
+      "content": "▁<extra_id_77>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250024": {
+      "content": "▁<extra_id_76>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250025": {
+      "content": "▁<extra_id_75>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250026": {
+      "content": "▁<extra_id_74>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250027": {
+      "content": "▁<extra_id_73>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250028": {
+      "content": "▁<extra_id_72>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250029": {
+      "content": "▁<extra_id_71>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250030": {
+      "content": "▁<extra_id_70>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250031": {
+      "content": "▁<extra_id_69>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250032": {
+      "content": "▁<extra_id_68>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250033": {
+      "content": "▁<extra_id_67>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250034": {
+      "content": "▁<extra_id_66>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250035": {
+      "content": "▁<extra_id_65>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250036": {
+      "content": "▁<extra_id_64>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250037": {
+      "content": "▁<extra_id_63>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250038": {
+      "content": "▁<extra_id_62>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250039": {
+      "content": "▁<extra_id_61>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250040": {
+      "content": "▁<extra_id_60>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250041": {
+      "content": "▁<extra_id_59>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250042": {
+      "content": "▁<extra_id_58>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250043": {
+      "content": "▁<extra_id_57>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250044": {
+      "content": "▁<extra_id_56>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250045": {
+      "content": "▁<extra_id_55>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250046": {
+      "content": "▁<extra_id_54>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250047": {
+      "content": "▁<extra_id_53>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250048": {
+      "content": "▁<extra_id_52>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250049": {
+      "content": "▁<extra_id_51>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250050": {
+      "content": "▁<extra_id_50>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250051": {
+      "content": "▁<extra_id_49>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250052": {
+      "content": "▁<extra_id_48>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250053": {
+      "content": "▁<extra_id_47>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250054": {
+      "content": "▁<extra_id_46>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250055": {
+      "content": "▁<extra_id_45>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250056": {
+      "content": "▁<extra_id_44>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250057": {
+      "content": "▁<extra_id_43>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250058": {
+      "content": "▁<extra_id_42>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250059": {
+      "content": "▁<extra_id_41>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250060": {
+      "content": "▁<extra_id_40>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250061": {
+      "content": "▁<extra_id_39>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250062": {
+      "content": "▁<extra_id_38>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250063": {
+      "content": "▁<extra_id_37>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250064": {
+      "content": "▁<extra_id_36>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250065": {
+      "content": "▁<extra_id_35>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250066": {
+      "content": "▁<extra_id_34>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250067": {
+      "content": "▁<extra_id_33>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250068": {
+      "content": "▁<extra_id_32>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250069": {
+      "content": "▁<extra_id_31>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250070": {
+      "content": "▁<extra_id_30>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250071": {
+      "content": "▁<extra_id_29>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250072": {
+      "content": "▁<extra_id_28>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250073": {
+      "content": "▁<extra_id_27>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250074": {
+      "content": "▁<extra_id_26>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250075": {
+      "content": "▁<extra_id_25>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250076": {
+      "content": "▁<extra_id_24>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250077": {
+      "content": "▁<extra_id_23>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250078": {
+      "content": "▁<extra_id_22>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250079": {
+      "content": "▁<extra_id_21>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250080": {
+      "content": "▁<extra_id_20>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250081": {
+      "content": "▁<extra_id_19>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250082": {
+      "content": "▁<extra_id_18>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250083": {
+      "content": "▁<extra_id_17>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250084": {
+      "content": "▁<extra_id_16>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250085": {
+      "content": "▁<extra_id_15>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250086": {
+      "content": "▁<extra_id_14>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250087": {
+      "content": "▁<extra_id_13>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250088": {
+      "content": "▁<extra_id_12>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250089": {
+      "content": "▁<extra_id_11>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250090": {
+      "content": "▁<extra_id_10>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250091": {
+      "content": "▁<extra_id_9>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250092": {
+      "content": "▁<extra_id_8>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250093": {
+      "content": "▁<extra_id_7>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250094": {
+      "content": "▁<extra_id_6>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250095": {
+      "content": "▁<extra_id_5>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250096": {
+      "content": "▁<extra_id_4>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250097": {
+      "content": "▁<extra_id_3>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250098": {
+      "content": "▁<extra_id_2>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250099": {
+      "content": "▁<extra_id_1>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250100": {
+      "content": "▁<extra_id_0>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "250101": {
+      "content": "[MASK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "[CLS]",
+  "clean_up_tokenization_spaces": false,
+  "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"
+}

trainer_state.json

@@ -0,0 +1,684 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 46875,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.032,
+      "grad_norm": 2.559859037399292,
+      "learning_rate": 4.9466666666666665e-05,
+      "loss": 8.6064,
+      "step": 500
+    },
+    {
+      "epoch": 0.064,
+      "grad_norm": 2.7166013717651367,
+      "learning_rate": 4.8933333333333335e-05,
+      "loss": 6.4365,
+      "step": 1000
+    },
+    {
+      "epoch": 0.096,
+      "grad_norm": 2.9375717639923096,
+      "learning_rate": 4.8400000000000004e-05,
+      "loss": 5.7166,
+      "step": 1500
+    },
+    {
+      "epoch": 0.128,
+      "grad_norm": 2.681896209716797,
+      "learning_rate": 4.7866666666666674e-05,
+      "loss": 5.3318,
+      "step": 2000
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 2.957535982131958,
+      "learning_rate": 4.7333333333333336e-05,
+      "loss": 5.0808,
+      "step": 2500
+    },
+    {
+      "epoch": 0.192,
+      "grad_norm": 2.7159996032714844,
+      "learning_rate": 4.6800000000000006e-05,
+      "loss": 4.8739,
+      "step": 3000
+    },
+    {
+      "epoch": 0.224,
+      "grad_norm": 2.7686867713928223,
+      "learning_rate": 4.626666666666667e-05,
+      "loss": 4.7194,
+      "step": 3500
+    },
+    {
+      "epoch": 0.256,
+      "grad_norm": 2.9994072914123535,
+      "learning_rate": 4.573333333333333e-05,
+      "loss": 4.5874,
+      "step": 4000
+    },
+    {
+      "epoch": 0.288,
+      "grad_norm": 3.0717546939849854,
+      "learning_rate": 4.52e-05,
+      "loss": 4.4778,
+      "step": 4500
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 2.996824264526367,
+      "learning_rate": 4.466666666666667e-05,
+      "loss": 4.3819,
+      "step": 5000
+    },
+    {
+      "epoch": 0.352,
+      "grad_norm": 2.8943400382995605,
+      "learning_rate": 4.413333333333334e-05,
+      "loss": 4.283,
+      "step": 5500
+    },
+    {
+      "epoch": 0.384,
+      "grad_norm": 3.34658145904541,
+      "learning_rate": 4.36e-05,
+      "loss": 4.2271,
+      "step": 6000
+    },
+    {
+      "epoch": 0.416,
+      "grad_norm": 3.0995116233825684,
+      "learning_rate": 4.3066666666666665e-05,
+      "loss": 4.153,
+      "step": 6500
+    },
+    {
+      "epoch": 0.448,
+      "grad_norm": 3.0321125984191895,
+      "learning_rate": 4.2533333333333335e-05,
+      "loss": 4.0983,
+      "step": 7000
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 3.119123697280884,
+      "learning_rate": 4.2e-05,
+      "loss": 4.0394,
+      "step": 7500
+    },
+    {
+      "epoch": 0.512,
+      "grad_norm": 3.162355422973633,
+      "learning_rate": 4.146666666666667e-05,
+      "loss": 4.0039,
+      "step": 8000
+    },
+    {
+      "epoch": 0.544,
+      "grad_norm": 3.0241446495056152,
+      "learning_rate": 4.093333333333334e-05,
+      "loss": 3.9588,
+      "step": 8500
+    },
+    {
+      "epoch": 0.576,
+      "grad_norm": 3.027881383895874,
+      "learning_rate": 4.0400000000000006e-05,
+      "loss": 3.932,
+      "step": 9000
+    },
+    {
+      "epoch": 0.608,
+      "grad_norm": 3.2551701068878174,
+      "learning_rate": 3.986666666666667e-05,
+      "loss": 3.8788,
+      "step": 9500
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 3.1692616939544678,
+      "learning_rate": 3.933333333333333e-05,
+      "loss": 3.8402,
+      "step": 10000
+    },
+    {
+      "epoch": 0.672,
+      "grad_norm": 3.4097533226013184,
+      "learning_rate": 3.88e-05,
+      "loss": 3.8155,
+      "step": 10500
+    },
+    {
+      "epoch": 0.704,
+      "grad_norm": 3.4331860542297363,
+      "learning_rate": 3.8266666666666664e-05,
+      "loss": 3.7781,
+      "step": 11000
+    },
+    {
+      "epoch": 0.736,
+      "grad_norm": 3.341395139694214,
+      "learning_rate": 3.773333333333334e-05,
+      "loss": 3.7592,
+      "step": 11500
+    },
+    {
+      "epoch": 0.768,
+      "grad_norm": 3.1860477924346924,
+      "learning_rate": 3.72e-05,
+      "loss": 3.7219,
+      "step": 12000
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 3.3436477184295654,
+      "learning_rate": 3.6666666666666666e-05,
+      "loss": 3.7017,
+      "step": 12500
+    },
+    {
+      "epoch": 0.832,
+      "grad_norm": 3.2480103969573975,
+      "learning_rate": 3.6133333333333335e-05,
+      "loss": 3.6795,
+      "step": 13000
+    },
+    {
+      "epoch": 0.864,
+      "grad_norm": 3.427396774291992,
+      "learning_rate": 3.56e-05,
+      "loss": 3.6672,
+      "step": 13500
+    },
+    {
+      "epoch": 0.896,
+      "grad_norm": 3.6126272678375244,
+      "learning_rate": 3.506666666666667e-05,
+      "loss": 3.6411,
+      "step": 14000
+    },
+    {
+      "epoch": 0.928,
+      "grad_norm": 3.2980282306671143,
+      "learning_rate": 3.453333333333334e-05,
+      "loss": 3.6083,
+      "step": 14500
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 3.086813449859619,
+      "learning_rate": 3.4000000000000007e-05,
+      "loss": 3.6038,
+      "step": 15000
+    },
+    {
+      "epoch": 0.992,
+      "grad_norm": 3.5595266819000244,
+      "learning_rate": 3.346666666666667e-05,
+      "loss": 3.5761,
+      "step": 15500
+    },
+    {
+      "epoch": 1.024,
+      "grad_norm": 3.3751089572906494,
+      "learning_rate": 3.293333333333333e-05,
+      "loss": 3.558,
+      "step": 16000
+    },
+    {
+      "epoch": 1.056,
+      "grad_norm": 3.356954336166382,
+      "learning_rate": 3.24e-05,
+      "loss": 3.537,
+      "step": 16500
+    },
+    {
+      "epoch": 1.088,
+      "grad_norm": 3.410510301589966,
+      "learning_rate": 3.1866666666666664e-05,
+      "loss": 3.5313,
+      "step": 17000
+    },
+    {
+      "epoch": 1.12,
+      "grad_norm": 3.2071657180786133,
+      "learning_rate": 3.1333333333333334e-05,
+      "loss": 3.5152,
+      "step": 17500
+    },
+    {
+      "epoch": 1.152,
+      "grad_norm": 3.232302188873291,
+      "learning_rate": 3.08e-05,
+      "loss": 3.5104,
+      "step": 18000
+    },
+    {
+      "epoch": 1.184,
+      "grad_norm": 3.7388410568237305,
+      "learning_rate": 3.0266666666666666e-05,
+      "loss": 3.4852,
+      "step": 18500
+    },
+    {
+      "epoch": 1.216,
+      "grad_norm": 3.546588659286499,
+      "learning_rate": 2.9733333333333336e-05,
+      "loss": 3.4776,
+      "step": 19000
+    },
+    {
+      "epoch": 1.248,
+      "grad_norm": 3.3913826942443848,
+      "learning_rate": 2.9199999999999998e-05,
+      "loss": 3.4615,
+      "step": 19500
+    },
+    {
+      "epoch": 1.28,
+      "grad_norm": 3.4872782230377197,
+      "learning_rate": 2.8666666666666668e-05,
+      "loss": 3.4588,
+      "step": 20000
+    },
+    {
+      "epoch": 1.312,
+      "grad_norm": 3.5155386924743652,
+      "learning_rate": 2.8133333333333334e-05,
+      "loss": 3.443,
+      "step": 20500
+    },
+    {
+      "epoch": 1.3439999999999999,
+      "grad_norm": 3.594855546951294,
+      "learning_rate": 2.7600000000000003e-05,
+      "loss": 3.4246,
+      "step": 21000
+    },
+    {
+      "epoch": 1.376,
+      "grad_norm": 3.613616466522217,
+      "learning_rate": 2.706666666666667e-05,
+      "loss": 3.4144,
+      "step": 21500
+    },
+    {
+      "epoch": 1.408,
+      "grad_norm": 3.264342784881592,
+      "learning_rate": 2.6533333333333332e-05,
+      "loss": 3.41,
+      "step": 22000
+    },
+    {
+      "epoch": 1.44,
+      "grad_norm": 3.3933544158935547,
+      "learning_rate": 2.6000000000000002e-05,
+      "loss": 3.4029,
+      "step": 22500
+    },
+    {
+      "epoch": 1.472,
+      "grad_norm": 3.900798797607422,
+      "learning_rate": 2.5466666666666668e-05,
+      "loss": 3.3901,
+      "step": 23000
+    },
+    {
+      "epoch": 1.504,
+      "grad_norm": 3.4809327125549316,
+      "learning_rate": 2.4933333333333334e-05,
+      "loss": 3.3738,
+      "step": 23500
+    },
+    {
+      "epoch": 1.536,
+      "grad_norm": 3.3931281566619873,
+      "learning_rate": 2.44e-05,
+      "loss": 3.3752,
+      "step": 24000
+    },
+    {
+      "epoch": 1.568,
+      "grad_norm": 3.4446539878845215,
+      "learning_rate": 2.3866666666666666e-05,
+      "loss": 3.3618,
+      "step": 24500
+    },
+    {
+      "epoch": 1.6,
+      "grad_norm": 3.256617546081543,
+      "learning_rate": 2.3333333333333336e-05,
+      "loss": 3.3559,
+      "step": 25000
+    },
+    {
+      "epoch": 1.6320000000000001,
+      "grad_norm": 3.423475503921509,
+      "learning_rate": 2.2800000000000002e-05,
+      "loss": 3.3445,
+      "step": 25500
+    },
+    {
+      "epoch": 1.6640000000000001,
+      "grad_norm": 4.00131893157959,
+      "learning_rate": 2.2266666666666668e-05,
+      "loss": 3.356,
+      "step": 26000
+    },
+    {
+      "epoch": 1.696,
+      "grad_norm": 3.4588286876678467,
+      "learning_rate": 2.1733333333333334e-05,
+      "loss": 3.3375,
+      "step": 26500
+    },
+    {
+      "epoch": 1.728,
+      "grad_norm": 3.8820817470550537,
+      "learning_rate": 2.12e-05,
+      "loss": 3.3301,
+      "step": 27000
+    },
+    {
+      "epoch": 1.76,
+      "grad_norm": 3.740394353866577,
+      "learning_rate": 2.0666666666666666e-05,
+      "loss": 3.3206,
+      "step": 27500
+    },
+    {
+      "epoch": 1.792,
+      "grad_norm": 3.536240339279175,
+      "learning_rate": 2.0133333333333336e-05,
+      "loss": 3.3305,
+      "step": 28000
+    },
+    {
+      "epoch": 1.8239999999999998,
+      "grad_norm": 3.527123212814331,
+      "learning_rate": 1.9600000000000002e-05,
+      "loss": 3.3162,
+      "step": 28500
+    },
+    {
+      "epoch": 1.8559999999999999,
+      "grad_norm": 3.5807790756225586,
+      "learning_rate": 1.9066666666666668e-05,
+      "loss": 3.3117,
+      "step": 29000
+    },
+    {
+      "epoch": 1.888,
+      "grad_norm": 3.438460111618042,
+      "learning_rate": 1.8533333333333334e-05,
+      "loss": 3.2941,
+      "step": 29500
+    },
+    {
+      "epoch": 1.92,
+      "grad_norm": 3.564744472503662,
+      "learning_rate": 1.8e-05,
+      "loss": 3.2991,
+      "step": 30000
+    },
+    {
+      "epoch": 1.952,
+      "grad_norm": 3.620878219604492,
+      "learning_rate": 1.7466666666666667e-05,
+      "loss": 3.2915,
+      "step": 30500
+    },
+    {
+      "epoch": 1.984,
+      "grad_norm": 3.5213491916656494,
+      "learning_rate": 1.6933333333333333e-05,
+      "loss": 3.2906,
+      "step": 31000
+    },
+    {
+      "epoch": 2.016,
+      "grad_norm": 3.39905047416687,
+      "learning_rate": 1.6400000000000002e-05,
+      "loss": 3.2781,
+      "step": 31500
+    },
+    {
+      "epoch": 2.048,
+      "grad_norm": 3.6184778213500977,
+      "learning_rate": 1.586666666666667e-05,
+      "loss": 3.2858,
+      "step": 32000
+    },
+    {
+      "epoch": 2.08,
+      "grad_norm": 4.695396900177002,
+      "learning_rate": 1.5333333333333334e-05,
+      "loss": 3.2691,
+      "step": 32500
+    },
+    {
+      "epoch": 2.112,
+      "grad_norm": 3.8069801330566406,
+      "learning_rate": 1.48e-05,
+      "loss": 3.2781,
+      "step": 33000
+    },
+    {
+      "epoch": 2.144,
+      "grad_norm": 3.849910259246826,
+      "learning_rate": 1.4266666666666667e-05,
+      "loss": 3.265,
+      "step": 33500
+    },
+    {
+      "epoch": 2.176,
+      "grad_norm": 3.518418550491333,
+      "learning_rate": 1.3733333333333335e-05,
+      "loss": 3.2553,
+      "step": 34000
+    },
+    {
+      "epoch": 2.208,
+      "grad_norm": 3.714787006378174,
+      "learning_rate": 1.32e-05,
+      "loss": 3.2659,
+      "step": 34500
+    },
+    {
+      "epoch": 2.24,
+      "grad_norm": 3.631002426147461,
+      "learning_rate": 1.2666666666666668e-05,
+      "loss": 3.2528,
+      "step": 35000
+    },
+    {
+      "epoch": 2.2720000000000002,
+      "grad_norm": 3.701296091079712,
+      "learning_rate": 1.2133333333333335e-05,
+      "loss": 3.2478,
+      "step": 35500
+    },
+    {
+      "epoch": 2.304,
+      "grad_norm": 3.5912673473358154,
+      "learning_rate": 1.16e-05,
+      "loss": 3.2542,
+      "step": 36000
+    },
+    {
+      "epoch": 2.336,
+      "grad_norm": 3.645458221435547,
+      "learning_rate": 1.1066666666666667e-05,
+      "loss": 3.2516,
+      "step": 36500
+    },
+    {
+      "epoch": 2.368,
+      "grad_norm": 3.7062928676605225,
+      "learning_rate": 1.0533333333333335e-05,
+      "loss": 3.2345,
+      "step": 37000
+    },
+    {
+      "epoch": 2.4,
+      "grad_norm": 3.8677589893341064,
+      "learning_rate": 1e-05,
+      "loss": 3.2316,
+      "step": 37500
+    },
+    {
+      "epoch": 2.432,
+      "grad_norm": 3.845417022705078,
+      "learning_rate": 9.466666666666667e-06,
+      "loss": 3.2383,
+      "step": 38000
+    },
+    {
+      "epoch": 2.464,
+      "grad_norm": 4.2547712326049805,
+      "learning_rate": 8.933333333333333e-06,
+      "loss": 3.2306,
+      "step": 38500
+    },
+    {
+      "epoch": 2.496,
+      "grad_norm": 3.7692208290100098,
+      "learning_rate": 8.400000000000001e-06,
+      "loss": 3.233,
+      "step": 39000
+    },
+    {
+      "epoch": 2.528,
+      "grad_norm": 3.726915121078491,
+      "learning_rate": 7.866666666666667e-06,
+      "loss": 3.2362,
+      "step": 39500
+    },
+    {
+      "epoch": 2.56,
+      "grad_norm": 3.675123691558838,
+      "learning_rate": 7.333333333333334e-06,
+      "loss": 3.2379,
+      "step": 40000
+    },
+    {
+      "epoch": 2.592,
+      "grad_norm": 3.647224187850952,
+      "learning_rate": 6.800000000000001e-06,
+      "loss": 3.2274,
+      "step": 40500
+    },
+    {
+      "epoch": 2.624,
+      "grad_norm": 3.850755453109741,
+      "learning_rate": 6.266666666666666e-06,
+      "loss": 3.2224,
+      "step": 41000
+    },
+    {
+      "epoch": 2.656,
+      "grad_norm": 3.7878735065460205,
+      "learning_rate": 5.733333333333333e-06,
+      "loss": 3.2182,
+      "step": 41500
+    },
+    {
+      "epoch": 2.6879999999999997,
+      "grad_norm": 3.540040969848633,
+      "learning_rate": 5.2e-06,
+      "loss": 3.2141,
+      "step": 42000
+    },
+    {
+      "epoch": 2.7199999999999998,
+      "grad_norm": 3.7240407466888428,
+      "learning_rate": 4.666666666666667e-06,
+      "loss": 3.2198,
+      "step": 42500
+    },
+    {
+      "epoch": 2.752,
+      "grad_norm": 3.9615602493286133,
+      "learning_rate": 4.133333333333333e-06,
+      "loss": 3.2218,
+      "step": 43000
+    },
+    {
+      "epoch": 2.784,
+      "grad_norm": 3.593024730682373,
+      "learning_rate": 3.6e-06,
+      "loss": 3.2123,
+      "step": 43500
+    },
+    {
+      "epoch": 2.816,
+      "grad_norm": 3.817379951477051,
+      "learning_rate": 3.066666666666667e-06,
+      "loss": 3.2133,
+      "step": 44000
+    },
+    {
+      "epoch": 2.848,
+      "grad_norm": 3.950862169265747,
+      "learning_rate": 2.5333333333333334e-06,
+      "loss": 3.2082,
+      "step": 44500
+    },
+    {
+      "epoch": 2.88,
+      "grad_norm": 3.852062225341797,
+      "learning_rate": 2.0000000000000003e-06,
+      "loss": 3.2017,
+      "step": 45000
+    },
+    {
+      "epoch": 2.912,
+      "grad_norm": 3.842193126678467,
+      "learning_rate": 1.4666666666666667e-06,
+      "loss": 3.1989,
+      "step": 45500
+    },
+    {
+      "epoch": 2.944,
+      "grad_norm": 4.043825149536133,
+      "learning_rate": 9.333333333333334e-07,
+      "loss": 3.2116,
+      "step": 46000
+    },
+    {
+      "epoch": 2.976,
+      "grad_norm": 3.7213408946990967,
+      "learning_rate": 4.0000000000000003e-07,
+      "loss": 3.2063,
+      "step": 46500
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 46875,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 4.41216036864e+17,
+  "train_batch_size": 16,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:27bfce03009e611aae8209c59c1a44a600b30c43f22740ffd7cf5ad55d5630ce
+size 5176