update model with additional 1.8ish epochs training

.gitattributes

@@ -1,7 +1,6 @@
 *.7z filter=lfs diff=lfs merge=lfs -text
 *.arrow filter=lfs diff=lfs merge=lfs -text
 *.bin filter=lfs diff=lfs merge=lfs -text
-*.bin.* filter=lfs diff=lfs merge=lfs -text
 *.bz2 filter=lfs diff=lfs merge=lfs -text
 *.ftz filter=lfs diff=lfs merge=lfs -text
 *.gz filter=lfs diff=lfs merge=lfs -text
@@ -17,10 +16,11 @@
 *.pt filter=lfs diff=lfs merge=lfs -text
 *.pth filter=lfs diff=lfs merge=lfs -text
 *.rar filter=lfs diff=lfs merge=lfs -text
-saved_model/**/* filter=lfs diff=lfs merge=lfs -text 
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.tar.* filter=lfs diff=lfs merge=lfs -text
 *.tflite filter=lfs diff=lfs merge=lfs -text
 *.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
 *.xz filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zstandard filter=lfs diff=lfs merge=lfs -text

.gitignore

@@ -0,0 +1 @@
+checkpoint-*/

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "allenai/led-large-16384",
+  "_name_or_path": "pszemraj/led-large-book-summary-11E",
   "_num_labels": 3,
   "activation_dropout": 0.0,
   "activation_function": "gelu",
@@ -52,7 +52,7 @@
   "length_penalty": 0.8,
   "max_decoder_position_embeddings": 1024,
   "max_encoder_position_embeddings": 16384,
-  "max_length": 1024,
+  "max_length": 512,
   "min_length": 32,
   "model_type": "led",
   "no_repeat_ngram_size": 3,
@@ -62,7 +62,7 @@
   "pad_token_id": 1,
   "prefix": " ",
   "torch_dtype": "float32",
-  "transformers_version": "4.16.2",
+  "transformers_version": "4.19.2",
   "use_cache": false,
   "vocab_size": 50265
 }

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6fd9ddcf9d526f237707eb0e06287ebc4b8b122e3624971dedca457623c67c26
+oid sha256:cae26c66ffc396d3f1b515359fef3cc2b4a080043dbfa56f7a968450c8ad1b27
 size 1839482407

tokenizer.json

@@ -1,59 +1,74 @@
 {
   "version": "1.0",
-  "truncation": null,
-  "padding": null,
+  "truncation": {
+    "direction": "Right",
+    "max_length": 1024,
+    "strategy": "LongestFirst",
+    "stride": 0
+  },
+  "padding": {
+    "strategy": {
+      "Fixed": 1024
+    },
+    "direction": "Right",
+    "pad_to_multiple_of": null,
+    "pad_id": 1,
+    "pad_type_id": 0,
+    "pad_token": "<pad>"
+  },
   "added_tokens": [
     {
       "id": 0,
-      "special": true,
       "content": "<s>",
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
-      "normalized": true
+      "normalized": true,
+      "special": true
     },
     {
       "id": 1,
-      "special": true,
       "content": "<pad>",
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
-      "normalized": true
+      "normalized": true,
+      "special": true
     },
     {
       "id": 2,
-      "special": true,
       "content": "</s>",
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
-      "normalized": true
+      "normalized": true,
+      "special": true
     },
     {
       "id": 3,
-      "special": true,
       "content": "<unk>",
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
-      "normalized": true
+      "normalized": true,
+      "special": true
     },
     {
       "id": 50264,
-      "special": true,
       "content": "<mask>",
       "single_word": false,
       "lstrip": true,
       "rstrip": false,
-      "normalized": true
+      "normalized": true,
+      "special": true
     }
   ],
   "normalizer": null,
   "pre_tokenizer": {
     "type": "ByteLevel",
     "add_prefix_space": false,
-    "trim_offsets": true
+    "trim_offsets": true,
+    "use_regex": true
   },
   "post_processor": {
     "type": "RobertaProcessing",
@@ -71,7 +86,8 @@
   "decoder": {
     "type": "ByteLevel",
     "add_prefix_space": true,
-    "trim_offsets": true
+    "trim_offsets": true,
+    "use_regex": true
   },
   "model": {
     "type": "BPE",

tokenizer_config.json

@@ -1 +1 @@
-{"unk_token": {"content": "<unk>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "bos_token": {"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "eos_token": {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "add_prefix_space": false, "errors": "replace", "sep_token": {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "cls_token": {"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "pad_token": {"content": "<pad>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "mask_token": {"content": "<mask>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "trim_offsets": true, "model_max_length": 16384, "special_tokens_map_file": "/root/.cache/huggingface/transformers/2ad921573d53ebf0c0450d63a211e61d8e328324e84830c365abff01f2d115f1.cb2244924ab24d706b02fd7fcedaea4531566537687a539ebb94db511fd122a0", "name_or_path": "allenai/led-large-16384", "tokenizer_class": "LEDTokenizer"}
+{"errors": "replace", "bos_token": {"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "eos_token": {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "sep_token": {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "cls_token": {"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "unk_token": {"content": "<unk>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "pad_token": {"content": "<pad>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "mask_token": {"content": "<mask>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "add_prefix_space": false, "trim_offsets": true, "model_max_length": 16384, "special_tokens_map_file": "/root/.cache/huggingface/transformers/2ad921573d53ebf0c0450d63a211e61d8e328324e84830c365abff01f2d115f1.cb2244924ab24d706b02fd7fcedaea4531566537687a539ebb94db511fd122a0", "name_or_path": "pszemraj/led-large-book-summary-11E", "tokenizer_class": "LEDTokenizer"}

trainer_state.json

@@ -0,0 +1,1537 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.9994069974303224,
+  "global_step": 1264,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.631578947368421e-06,
+      "loss": 1.0727,
+      "step": 5
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 5.263157894736842e-06,
+      "loss": 0.8573,
+      "step": 10
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 7.894736842105265e-06,
+      "loss": 1.0991,
+      "step": 15
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.0526315789473684e-05,
+      "loss": 0.8893,
+      "step": 20
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.3157894736842108e-05,
+      "loss": 0.9549,
+      "step": 25
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.578947368421053e-05,
+      "loss": 0.9953,
+      "step": 30
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.8421052631578947e-05,
+      "loss": 1.0163,
+      "step": 35
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9999868674866755e-05,
+      "loss": 0.9562,
+      "step": 40
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9998391306730024e-05,
+      "loss": 1.0517,
+      "step": 45
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9995272657365104e-05,
+      "loss": 0.9458,
+      "step": 50
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9990513238712407e-05,
+      "loss": 0.9163,
+      "step": 55
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.998411383205207e-05,
+      "loss": 0.8653,
+      "step": 60
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9976075487875692e-05,
+      "loss": 1.063,
+      "step": 65
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9966399525713882e-05,
+      "loss": 0.8901,
+      "step": 70
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9955087533919662e-05,
+      "loss": 1.0556,
+      "step": 75
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.994214136940773e-05,
+      "loss": 0.9434,
+      "step": 80
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9927563157349646e-05,
+      "loss": 0.9989,
+      "step": 85
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9911355290824955e-05,
+      "loss": 0.9829,
+      "step": 90
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9893520430428378e-05,
+      "loss": 0.9988,
+      "step": 95
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.987406150383304e-05,
+      "loss": 1.0073,
+      "step": 100
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.98529817053099e-05,
+      "loss": 0.8863,
+      "step": 105
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.983028449520338e-05,
+      "loss": 0.8015,
+      "step": 110
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.980597359936335e-05,
+      "loss": 1.0477,
+      "step": 115
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.9780053008533486e-05,
+      "loss": 0.961,
+      "step": 120
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.97525269776962e-05,
+      "loss": 0.9476,
+      "step": 125
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.9723400025374168e-05,
+      "loss": 1.1,
+      "step": 130
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.969267693288855e-05,
+      "loss": 0.946,
+      "step": 135
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.9660362743574163e-05,
+      "loss": 0.946,
+      "step": 140
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.9626462761951583e-05,
+      "loss": 0.9368,
+      "step": 145
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.959098255285636e-05,
+      "loss": 1.0338,
+      "step": 150
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.9553927940525557e-05,
+      "loss": 1.0359,
+      "step": 155
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.9515305007641653e-05,
+      "loss": 1.0189,
+      "step": 160
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.9475120094334046e-05,
+      "loss": 1.0388,
+      "step": 165
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.9433379797138314e-05,
+      "loss": 1.0733,
+      "step": 170
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.939009096791333e-05,
+      "loss": 0.9984,
+      "step": 175
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.9345260712716517e-05,
+      "loss": 1.0221,
+      "step": 180
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.9298896390637363e-05,
+      "loss": 1.2205,
+      "step": 185
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.9251005612589382e-05,
+      "loss": 0.9938,
+      "step": 190
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.9201596240060737e-05,
+      "loss": 1.0059,
+      "step": 195
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.9150676383823775e-05,
+      "loss": 1.0072,
+      "step": 200
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.9098254402603573e-05,
+      "loss": 0.8886,
+      "step": 205
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.904433890170583e-05,
+      "loss": 1.1496,
+      "step": 210
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.898893873160428e-05,
+      "loss": 0.8883,
+      "step": 215
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.893206298648781e-05,
+      "loss": 0.9564,
+      "step": 220
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.887372100276764e-05,
+      "loss": 0.9264,
+      "step": 225
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.8813922357544713e-05,
+      "loss": 1.1487,
+      "step": 230
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.875267686703754e-05,
+      "loss": 1.0543,
+      "step": 235
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.8689994584970843e-05,
+      "loss": 1.0098,
+      "step": 240
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.8625885800925193e-05,
+      "loss": 0.915,
+      "step": 245
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.8560361038647917e-05,
+      "loss": 1.0092,
+      "step": 250
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.8493431054325583e-05,
+      "loss": 1.0312,
+      "step": 255
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.8425106834818336e-05,
+      "loss": 1.0963,
+      "step": 260
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.8355399595856326e-05,
+      "loss": 0.8076,
+      "step": 265
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.8284320780198624e-05,
+      "loss": 0.8218,
+      "step": 270
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.8211882055754824e-05,
+      "loss": 0.9424,
+      "step": 275
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.8138095313669705e-05,
+      "loss": 1.0962,
+      "step": 280
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.8062972666371258e-05,
+      "loss": 1.0918,
+      "step": 285
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.798652644558236e-05,
+      "loss": 1.065,
+      "step": 290
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.790876920029647e-05,
+      "loss": 0.9168,
+      "step": 295
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.7829713694717665e-05,
+      "loss": 1.0295,
+      "step": 300
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.774937290616533e-05,
+      "loss": 0.9455,
+      "step": 305
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.7667760022943864e-05,
+      "loss": 1.0272,
+      "step": 310
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.7584888442177774e-05,
+      "loss": 1.127,
+      "step": 315
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.7500771767612473e-05,
+      "loss": 1.0332,
+      "step": 320
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.7415423807381162e-05,
+      "loss": 1.0603,
+      "step": 325
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.7328858571738157e-05,
+      "loss": 1.1075,
+      "step": 330
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.7241090270759055e-05,
+      "loss": 1.1279,
+      "step": 335
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.715213331200807e-05,
+      "loss": 1.1037,
+      "step": 340
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.7062002298172984e-05,
+      "loss": 0.9818,
+      "step": 345
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.697071202466803e-05,
+      "loss": 0.9672,
+      "step": 350
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.687827747720517e-05,
+      "loss": 0.8006,
+      "step": 355
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.6784713829334124e-05,
+      "loss": 0.9823,
+      "step": 360
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.6690036439951552e-05,
+      "loss": 1.0844,
+      "step": 365
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.6594260850779837e-05,
+      "loss": 1.0858,
+      "step": 370
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.6497402783815834e-05,
+      "loss": 0.9117,
+      "step": 375
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.6399478138750015e-05,
+      "loss": 0.9313,
+      "step": 380
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.63005029903565e-05,
+      "loss": 0.9776,
+      "step": 385
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.620049358585427e-05,
+      "loss": 1.1445,
+      "step": 390
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.609946634224015e-05,
+      "loss": 1.0567,
+      "step": 395
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.5997437843593856e-05,
+      "loss": 0.9138,
+      "step": 400
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.5894424838355654e-05,
+      "loss": 1.1003,
+      "step": 405
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.5790444236577028e-05,
+      "loss": 1.0624,
+      "step": 410
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.568551310714482e-05,
+      "loss": 0.8902,
+      "step": 415
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.557964867497929e-05,
+      "loss": 1.0329,
+      "step": 420
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.5472868318206566e-05,
+      "loss": 1.0023,
+      "step": 425
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.5365189565305957e-05,
+      "loss": 0.9393,
+      "step": 430
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.5256630092232567e-05,
+      "loss": 0.9747,
+      "step": 435
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.5147207719515692e-05,
+      "loss": 1.0078,
+      "step": 440
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.5036940409333533e-05,
+      "loss": 0.9576,
+      "step": 445
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.4925846262564592e-05,
+      "loss": 0.992,
+      "step": 450
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.4813943515816344e-05,
+      "loss": 1.0192,
+      "step": 455
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.4701250538431617e-05,
+      "loss": 1.0154,
+      "step": 460
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.4587785829473173e-05,
+      "loss": 1.1043,
+      "step": 465
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.4473568014687018e-05,
+      "loss": 0.8351,
+      "step": 470
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.4358615843444876e-05,
+      "loss": 1.0276,
+      "step": 475
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.4242948185666419e-05,
+      "loss": 0.9423,
+      "step": 480
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.4126584028721677e-05,
+      "loss": 0.9598,
+      "step": 485
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.4009542474314173e-05,
+      "loss": 1.1755,
+      "step": 490
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.3891842735345285e-05,
+      "loss": 1.1018,
+      "step": 495
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.3773504132760379e-05,
+      "loss": 1.0445,
+      "step": 500
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.3654546092377166e-05,
+      "loss": 0.9674,
+      "step": 505
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.3534988141696891e-05,
+      "loss": 1.0473,
+      "step": 510
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.3414849906698788e-05,
+      "loss": 0.9346,
+      "step": 515
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.3294151108618379e-05,
+      "loss": 0.9643,
+      "step": 520
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.3172911560710167e-05,
+      "loss": 1.113,
+      "step": 525
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.3051151164995188e-05,
+      "loss": 1.0155,
+      "step": 530
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.2928889908994003e-05,
+      "loss": 0.9675,
+      "step": 535
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.280614786244566e-05,
+      "loss": 0.8424,
+      "step": 540
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.2682945174013148e-05,
+      "loss": 1.1247,
+      "step": 545
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.2559302067975914e-05,
+      "loss": 0.9641,
+      "step": 550
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.243523884090995e-05,
+      "loss": 1.0329,
+      "step": 555
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.2310775858356017e-05,
+      "loss": 1.0533,
+      "step": 560
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.2185933551476545e-05,
+      "loss": 1.0216,
+      "step": 565
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.2060732413701773e-05,
+      "loss": 1.0346,
+      "step": 570
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.1935192997365666e-05,
+      "loss": 0.9749,
+      "step": 575
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.1809335910332136e-05,
+      "loss": 1.0901,
+      "step": 580
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.1683181812612186e-05,
+      "loss": 1.0043,
+      "step": 585
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.1556751412972462e-05,
+      "loss": 0.9162,
+      "step": 590
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.1430065465535827e-05,
+      "loss": 0.9228,
+      "step": 595
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.1303144766374476e-05,
+      "loss": 0.9568,
+      "step": 600
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.1176010150096158e-05,
+      "loss": 1.063,
+      "step": 605
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.104868248642408e-05,
+      "loss": 0.9938,
+      "step": 610
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.092118267677106e-05,
+      "loss": 1.1056,
+      "step": 615
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.0793531650808469e-05,
+      "loss": 0.9269,
+      "step": 620
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.0665750363030498e-05,
+      "loss": 1.0452,
+      "step": 625
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.0537859789314424e-05,
+      "loss": 0.8855,
+      "step": 630
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.0409880923477293e-05,
+      "loss": 1.0583,
+      "step": 635
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 1.028183477382971e-05,
+      "loss": 0.8533,
+      "step": 640
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 1.0153742359727226e-05,
+      "loss": 0.942,
+      "step": 645
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 1.0025624708119901e-05,
+      "loss": 1.0044,
+      "step": 650
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 9.897502850100648e-06,
+      "loss": 0.9652,
+      "step": 655
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 9.76939781745289e-06,
+      "loss": 1.0202,
+      "step": 660
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 9.641330639198083e-06,
+      "loss": 0.9401,
+      "step": 665
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 9.513322338143714e-06,
+      "loss": 0.8343,
+      "step": 670
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 9.385393927432307e-06,
+      "loss": 0.8904,
+      "step": 675
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 9.257566407092032e-06,
+      "loss": 0.9143,
+      "step": 680
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 9.129860760589441e-06,
+      "loss": 0.8408,
+      "step": 685
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 9.002297951384945e-06,
+      "loss": 0.9426,
+      "step": 690
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 8.874898919491564e-06,
+      "loss": 0.9476,
+      "step": 695
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 8.74768457803754e-06,
+      "loss": 0.9992,
+      "step": 700
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 8.62067580983333e-06,
+      "loss": 0.8491,
+      "step": 705
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 8.493893463943617e-06,
+      "loss": 0.8829,
+      "step": 710
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 8.367358352264834e-06,
+      "loss": 0.7933,
+      "step": 715
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 8.241091246108796e-06,
+      "loss": 0.9976,
+      "step": 720
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 8.115112872793006e-06,
+      "loss": 0.9947,
+      "step": 725
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 7.989443912238151e-06,
+      "loss": 0.9171,
+      "step": 730
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 7.864104993573422e-06,
+      "loss": 0.9348,
+      "step": 735
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 7.73911669175013e-06,
+      "loss": 0.9357,
+      "step": 740
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 7.614499524164251e-06,
+      "loss": 0.9273,
+      "step": 745
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 7.490273947288389e-06,
+      "loss": 0.8991,
+      "step": 750
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 7.366460353313762e-06,
+      "loss": 0.8373,
+      "step": 755
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 7.2430790668027274e-06,
+      "loss": 0.8316,
+      "step": 760
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 7.120150341352413e-06,
+      "loss": 0.7532,
+      "step": 765
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 6.99769435627e-06,
+      "loss": 0.8851,
+      "step": 770
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 6.875731213260193e-06,
+      "loss": 0.9678,
+      "step": 775
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 6.754280933125441e-06,
+      "loss": 0.9968,
+      "step": 780
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 6.633363452479431e-06,
+      "loss": 0.9126,
+      "step": 785
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 6.512998620474396e-06,
+      "loss": 0.8765,
+      "step": 790
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 6.393206195542791e-06,
+      "loss": 0.8761,
+      "step": 795
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 6.27400584215386e-06,
+      "loss": 0.99,
+      "step": 800
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 6.155417127585617e-06,
+      "loss": 0.7691,
+      "step": 805
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 6.037459518712796e-06,
+      "loss": 0.7794,
+      "step": 810
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 5.920152378811268e-06,
+      "loss": 0.9978,
+      "step": 815
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 5.803514964379482e-06,
+      "loss": 0.8165,
+      "step": 820
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 5.68756642197741e-06,
+      "loss": 0.8918,
+      "step": 825
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 5.572325785083563e-06,
+      "loss": 0.9572,
+      "step": 830
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 5.457811970970564e-06,
+      "loss": 0.9112,
+      "step": 835
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 5.3440437775997636e-06,
+      "loss": 0.9659,
+      "step": 840
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 5.231039880535511e-06,
+      "loss": 0.9516,
+      "step": 845
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 5.118818829879442e-06,
+      "loss": 1.076,
+      "step": 850
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 5.0073990472254075e-06,
+      "loss": 0.8136,
+      "step": 855
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 4.8967988226354945e-06,
+      "loss": 0.9711,
+      "step": 860
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 4.787036311637609e-06,
+      "loss": 0.8879,
+      "step": 865
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 4.678129532245189e-06,
+      "loss": 1.0423,
+      "step": 870
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 4.570096361999445e-06,
+      "loss": 0.892,
+      "step": 875
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 4.462954535034692e-06,
+      "loss": 0.9188,
+      "step": 880
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 4.356721639167202e-06,
+      "loss": 0.8706,
+      "step": 885
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 4.251415113008096e-06,
+      "loss": 0.8926,
+      "step": 890
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 4.147052243100706e-06,
+      "loss": 0.9126,
+      "step": 895
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 4.043650161082913e-06,
+      "loss": 0.9201,
+      "step": 900
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 3.941225840874925e-06,
+      "loss": 0.8758,
+      "step": 905
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 3.839796095892905e-06,
+      "loss": 0.8818,
+      "step": 910
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 3.7393775762889963e-06,
+      "loss": 0.9018,
+      "step": 915
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 3.639986766218112e-06,
+      "loss": 0.8346,
+      "step": 920
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 3.541639981131996e-06,
+      "loss": 0.9619,
+      "step": 925
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 3.4443533651009474e-06,
+      "loss": 1.0202,
+      "step": 930
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 3.348142888163726e-06,
+      "loss": 0.9186,
+      "step": 935
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 3.2530243437059773e-06,
+      "loss": 0.9326,
+      "step": 940
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 3.1590133458676787e-06,
+      "loss": 0.8347,
+      "step": 945
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 3.066125326980027e-06,
+      "loss": 0.8914,
+      "step": 950
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 2.9743755350321213e-06,
+      "loss": 0.8869,
+      "step": 955
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.8837790311679625e-06,
+      "loss": 0.9833,
+      "step": 960
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.7943506872140844e-06,
+      "loss": 0.8893,
+      "step": 965
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.7061051832382836e-06,
+      "loss": 0.8616,
+      "step": 970
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 2.6190570051398035e-06,
+      "loss": 0.9909,
+      "step": 975
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.5332204422714368e-06,
+      "loss": 0.9515,
+      "step": 980
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.4486095850938352e-06,
+      "loss": 1.0862,
+      "step": 985
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.365238322862511e-06,
+      "loss": 0.9123,
+      "step": 990
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.2831203413478555e-06,
+      "loss": 0.8965,
+      "step": 995
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 2.202269120588546e-06,
+      "loss": 0.9261,
+      "step": 1000
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 2.122697932678748e-06,
+      "loss": 0.9044,
+      "step": 1005
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 2.0444198395894332e-06,
+      "loss": 0.923,
+      "step": 1010
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 1.9674476910242055e-06,
+      "loss": 0.9581,
+      "step": 1015
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 1.891794122309949e-06,
+      "loss": 1.0578,
+      "step": 1020
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 1.8174715523227017e-06,
+      "loss": 0.886,
+      "step": 1025
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 1.7444921814490256e-06,
+      "loss": 0.8365,
+      "step": 1030
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 1.6728679895832622e-06,
+      "loss": 0.9913,
+      "step": 1035
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 1.6026107341609842e-06,
+      "loss": 0.8921,
+      "step": 1040
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 1.5337319482289503e-06,
+      "loss": 0.8323,
+      "step": 1045
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 1.4662429385519084e-06,
+      "loss": 0.8869,
+      "step": 1050
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 1.400154783756541e-06,
+      "loss": 0.9799,
+      "step": 1055
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 1.3354783325128561e-06,
+      "loss": 0.9416,
+      "step": 1060
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 1.2722242017533192e-06,
+      "loss": 0.9801,
+      "step": 1065
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 1.2104027749300574e-06,
+      "loss": 0.9614,
+      "step": 1070
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 1.150024200310348e-06,
+      "loss": 1.0792,
+      "step": 1075
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 1.0910983893107419e-06,
+      "loss": 0.8622,
+      "step": 1080
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 1.0336350148700668e-06,
+      "loss": 0.8627,
+      "step": 1085
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 9.776435098615578e-07,
+      "loss": 0.857,
+      "step": 1090
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 9.231330655444193e-07,
+      "loss": 1.0396,
+      "step": 1095
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 8.701126300550322e-07,
+      "loss": 1.0174,
+      "step": 1100
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 8.185909069380782e-07,
+      "loss": 1.031,
+      "step": 1105
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.685763537178093e-07,
+      "loss": 0.9254,
+      "step": 1110
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.200771805097206e-07,
+      "loss": 0.9294,
+      "step": 1115
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 6.731013486728044e-07,
+      "loss": 0.8641,
+      "step": 1120
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 6.276565695026671e-07,
+      "loss": 0.9669,
+      "step": 1125
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 5.837503029656888e-07,
+      "loss": 0.8447,
+      "step": 1130
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 5.413897564744253e-07,
+      "loss": 0.9566,
+      "step": 1135
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 5.005818837044885e-07,
+      "loss": 0.9457,
+      "step": 1140
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 4.613333834530631e-07,
+      "loss": 0.8523,
+      "step": 1145
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 4.23650698539273e-07,
+      "loss": 0.9696,
+      "step": 1150
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 3.8754001474655354e-07,
+      "loss": 0.7867,
+      "step": 1155
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 3.530072598072454e-07,
+      "loss": 0.9534,
+      "step": 1160
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 3.200581024295102e-07,
+      "loss": 0.9209,
+      "step": 1165
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 2.886979513667998e-07,
+      "loss": 0.8257,
+      "step": 1170
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 2.589319545299807e-07,
+      "loss": 0.9424,
+      "step": 1175
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 2.3076499814227992e-07,
+      "loss": 0.873,
+      "step": 1180
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 2.042017059371948e-07,
+      "loss": 0.843,
+      "step": 1185
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.7924643839947632e-07,
+      "loss": 0.9136,
+      "step": 1190
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 1.559032920493464e-07,
+      "loss": 1.0255,
+      "step": 1195
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.3417609877002691e-07,
+      "loss": 0.8384,
+      "step": 1200
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 1.1406842517872608e-07,
+      "loss": 0.7154,
+      "step": 1205
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 9.558357204115464e-08,
+      "loss": 0.8805,
+      "step": 1210
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 7.872457372969711e-08,
+      "loss": 0.7834,
+      "step": 1215
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.34941977253023e-08,
+      "loss": 0.8752,
+      "step": 1220
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 4.989494416318685e-08,
+      "loss": 0.9148,
+      "step": 1225
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 3.7929045422432364e-08,
+      "loss": 0.8612,
+      "step": 1230
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 2.7598465759526294e-08,
+      "loss": 0.9299,
+      "step": 1235
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 1.8904900985918796e-08,
+      "loss": 0.9918,
+      "step": 1240
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.184977818965205e-08,
+      "loss": 0.9008,
+      "step": 1245
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 6.434255501095443e-09,
+      "loss": 0.9461,
+      "step": 1250
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 2.659221902830966e-09,
+      "loss": 0.8181,
+      "step": 1255
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.252970837255067e-10,
+      "loss": 0.8643,
+      "step": 1260
+    },
+    {
+      "epoch": 2.0,
+      "step": 1264,
+      "total_flos": 3.882961297263821e+17,
+      "train_loss": 0.9575079066466682,
+      "train_runtime": 77706.2136,
+      "train_samples_per_second": 0.26,
+      "train_steps_per_second": 0.016
+    }
+  ],
+  "max_steps": 1264,
+  "num_train_epochs": 2,
+  "total_flos": 3.882961297263821e+17,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d2cfacbe013a88ea45672bbf16b7c9b6dedbe580c5f57c175c53b218c0dc6b3f
-size 4207
+oid sha256:c2c8d84c6ea850e743bbd2858bd2ac8b6913bd2628de22bb48d811f906d778c2
+size 4399