Training in progress, step 100, checkpoint

checkpoint-100/config.json

@@ -1,35 +1,36 @@
 {
-  "_name_or_path": "klue/roberta-large",
+  "_name_or_path": "EleutherAI/polyglot-ko-1.3b",
   "architectures": [
-    "RobertaForSequenceClassification"
+    "GPTNeoXForSequenceClassification"
   ],
-  "attention_probs_dropout_prob": 0.1,
+  "attention_dropout": 0.0,
   "bos_token_id": 0,
-  "classifier_dropout": null,
+  "classifier_dropout": 0.1,
   "eos_token_id": 2,
-  "gradient_checkpointing": false,
   "hidden_act": "gelu",
-  "hidden_dropout_prob": 0.1,
-  "hidden_size": 1024,
+  "hidden_dropout": 0.0,
+  "hidden_size": 2048,
   "id2label": {
     "0": "LABEL_0"
   },
   "initializer_range": 0.02,
-  "intermediate_size": 4096,
+  "intermediate_size": 8192,
   "label2id": {
     "LABEL_0": 0
   },
   "layer_norm_eps": 1e-05,
-  "max_position_embeddings": 514,
-  "model_type": "roberta",
+  "max_position_embeddings": 2048,
+  "model_type": "gpt_neox",
   "num_attention_heads": 16,
   "num_hidden_layers": 24,
   "pad_token_id": 2,
-  "position_embedding_type": "absolute",
-  "tokenizer_class": "BertTokenizer",
+  "rope_scaling": null,
+  "rotary_emb_base": 10000,
+  "rotary_pct": 0.5,
+  "tie_word_embeddings": false,
   "torch_dtype": "float32",
   "transformers_version": "4.34.1",
-  "type_vocab_size": 1,
   "use_cache": true,
-  "vocab_size": 32000
+  "use_parallel_residual": true,
+  "vocab_size": 30080
 }

checkpoint-100/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c40d7d6ac7c9f651b2166fa7ccf825f008a4b0f6cc15c384dd83ced12e5db793
-size 2693490234
+oid sha256:ade30226eb4e297d8decaec73c9550740f04a0911264dc394137a7a57b068dde
+size 10161845574

checkpoint-100/pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:59587ddb0c8585b1b8154e3d30ecf54a9e18dae1ce009bcdaee1aac965f85089
-size 1346765678
+oid sha256:d5a038be930d7e4b9c06380cca5c12b961950b685498eb1df7b7188f4af6c5ba
+size 5080945162

checkpoint-100/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9e06bbdec16b98761b8f1c3f2be37280e4c9a8c49585a0b40267fbbf7cd255b2
+oid sha256:2c8c1979a903366056f816dfa903e1707fbe02269ee67fd5bd4bb542c4208efb
 size 14308

checkpoint-100/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1eb9b965954fc6facf61811f017e0585a50011ee22f071a2fb1e6e116234d3d7
+oid sha256:d25eaeb8d0bf8a025a9171ec8459913c5e37aeef9ab9f3203b3512815e80a717
 size 1064

checkpoint-100/special_tokens_map.json

@@ -1,9 +1,11 @@
 {
-  "bos_token": "[CLS]",
-  "cls_token": "[CLS]",
-  "eos_token": "[SEP]",
-  "mask_token": "[MASK]",
-  "pad_token": "[PAD]",
-  "sep_token": "[SEP]",
-  "unk_token": "[UNK]"
+  "additional_special_tokens": [
+    "<|endoftext|>",
+    "<|sep|>",
+    "<|acc|>",
+    "<|tel|>",
+    "<|rrn|>"
+  ],
+  "eos_token": "<|endoftext|>",
+  "pad_token": "<|endoftext|>"
 }

checkpoint-100/tokenizer_config.json

@@ -1,7 +1,7 @@
 {
   "added_tokens_decoder": {
     "0": {
-      "content": "[CLS]",
+      "content": "<|unused0|>",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -9,7 +9,7 @@
       "special": true
     },
     "1": {
-      "content": "[PAD]",
+      "content": "<|unused1|>",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -17,7 +17,7 @@
       "special": true
     },
     "2": {
-      "content": "[SEP]",
+      "content": "<|endoftext|>",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -25,15 +25,31 @@
       "special": true
     },
     "3": {
-      "content": "[UNK]",
+      "content": "<|sep|>",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
       "single_word": false,
       "special": true
     },
-    "4": {
-      "content": "[MASK]",
+    "30000": {
+      "content": "<|acc|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "30001": {
+      "content": "<|tel|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "30002": {
+      "content": "<|rrn|>",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -41,19 +57,16 @@
       "special": true
     }
   },
-  "bos_token": "[CLS]",
+  "additional_special_tokens": [
+    "<|endoftext|>",
+    "<|sep|>",
+    "<|acc|>",
+    "<|tel|>",
+    "<|rrn|>"
+  ],
   "clean_up_tokenization_spaces": true,
-  "cls_token": "[CLS]",
-  "do_basic_tokenize": true,
-  "do_lower_case": false,
-  "eos_token": "[SEP]",
-  "mask_token": "[MASK]",
-  "model_max_length": 512,
-  "never_split": null,
-  "pad_token": "[PAD]",
-  "sep_token": "[SEP]",
-  "strip_accents": null,
-  "tokenize_chinese_chars": true,
-  "tokenizer_class": "BertTokenizer",
-  "unk_token": "[UNK]"
+  "eos_token": "<|endoftext|>",
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "<|endoftext|>",
+  "tokenizer_class": "PreTrainedTokenizerFast"
 }

checkpoint-100/trainer_state.json

@@ -1,7 +1,7 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 0.15241579027587257,
+  "epoch": 0.251414204902577,
   "eval_steps": 100,
   "global_step": 100,
   "is_hyper_param_search": false,
@@ -9,77 +9,77 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.02,
-      "learning_rate": 8.99484069051768e-06,
-      "loss": 0.672,
+      "epoch": 0.03,
+      "learning_rate": 8.985917667125256e-06,
+      "loss": 0.4081,
       "step": 10
     },
     {
-      "epoch": 0.03,
-      "learning_rate": 8.979374592503753e-06,
-      "loss": 0.393,
+      "epoch": 0.05,
+      "learning_rate": 8.943758807211778e-06,
+      "loss": 0.2059,
       "step": 20
     },
     {
-      "epoch": 0.05,
-      "learning_rate": 8.953637170129838e-06,
-      "loss": 0.1594,
+      "epoch": 0.08,
+      "learning_rate": 8.873787284747973e-06,
+      "loss": 0.1613,
       "step": 30
     },
     {
-      "epoch": 0.06,
-      "learning_rate": 8.917687439985847e-06,
-      "loss": 0.122,
+      "epoch": 0.1,
+      "learning_rate": 8.776441038520995e-06,
+      "loss": 0.1359,
       "step": 40
     },
     {
-      "epoch": 0.08,
-      "learning_rate": 8.87160783575339e-06,
-      "loss": 0.0956,
+      "epoch": 0.13,
+      "learning_rate": 8.652329340639053e-06,
+      "loss": 0.1794,
       "step": 50
     },
     {
-      "epoch": 0.09,
-      "learning_rate": 8.815504019183158e-06,
-      "loss": 0.102,
+      "epoch": 0.15,
+      "learning_rate": 8.502228983210244e-06,
+      "loss": 0.1471,
       "step": 60
     },
     {
-      "epoch": 0.11,
-      "learning_rate": 8.749504637809742e-06,
-      "loss": 0.0888,
+      "epoch": 0.18,
+      "learning_rate": 8.327079416544763e-06,
+      "loss": 0.1173,
       "step": 70
     },
     {
-      "epoch": 0.12,
-      "learning_rate": 8.673761029959426e-06,
-      "loss": 0.097,
+      "epoch": 0.2,
+      "learning_rate": 8.127976869309567e-06,
+      "loss": 0.1016,
       "step": 80
     },
     {
-      "epoch": 0.14,
-      "learning_rate": 8.588446877727417e-06,
-      "loss": 0.0968,
+      "epoch": 0.23,
+      "learning_rate": 7.906167487436374e-06,
+      "loss": 0.1486,
       "step": 90
     },
     {
-      "epoch": 0.15,
-      "learning_rate": 8.493757808720196e-06,
-      "loss": 0.0808,
+      "epoch": 0.25,
+      "learning_rate": 7.663039534725309e-06,
+      "loss": 0.1029,
       "step": 100
     },
     {
-      "epoch": 0.15,
-      "eval_accuracy": 0.0,
-      "eval_loss": 3.0685205274494365e-05,
-      "eval_runtime": 30.877,
-      "eval_samples_per_second": 141.659,
-      "eval_steps_per_second": 23.61,
+      "epoch": 0.25,
+      "eval_accuracy": 0.5,
+      "eval_loss": 1.1548028089336526e-09,
+      "eval_runtime": 31.6494,
+      "eval_samples_per_second": 89.354,
+      "eval_steps_per_second": 11.185,
       "step": 100
     }
   ],
   "logging_steps": 10,
-  "max_steps": 656,
+  "max_steps": 397,
   "num_train_epochs": 1,
   "save_steps": 100,
   "total_flos": 0.0,

checkpoint-100/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e1b0be5f447301b577e44061c504ec9398fb05913a538169784e7cfc7134c7d3
+oid sha256:19f57ba66f27fb038e41f4b640f06321151aa6bdef95a3c3a615bf73330d3d10
 size 4600