Training in progress, epoch 5, checkpoint

checkpoint-235/config.json

@@ -0,0 +1,43 @@
+{
+  "add_cross_attention": false,
+  "architectures": [
+    "ElectraForSequenceClassification"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "bos_token_id": null,
+  "classifier_dropout": null,
+  "dtype": "float32",
+  "embedding_size": 128,
+  "eos_token_id": null,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 256,
+  "id2label": {
+    "0": "valid",
+    "1": "fallacy"
+  },
+  "initializer_range": 0.02,
+  "intermediate_size": 1024,
+  "is_decoder": false,
+  "label2id": {
+    "fallacy": 1,
+    "valid": 0
+  },
+  "layer_norm_eps": 1e-12,
+  "max_position_embeddings": 512,
+  "model_type": "electra",
+  "num_attention_heads": 4,
+  "num_hidden_layers": 12,
+  "pad_token_id": 0,
+  "position_embedding_type": "absolute",
+  "problem_type": "single_label_classification",
+  "summary_activation": "gelu",
+  "summary_last_dropout": 0.1,
+  "summary_type": "first",
+  "summary_use_proj": true,
+  "tie_word_embeddings": true,
+  "transformers_version": "5.9.0",
+  "type_vocab_size": 2,
+  "use_cache": false,
+  "vocab_size": 30522
+}

checkpoint-235/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0e74d60afcb36718d989af9c5824bb37f7abc35e87e5694d9a539c43839d15c2
+size 54221200

checkpoint-235/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4afc547d5ed9672a6ea581cbcefe90301f991dab29500acea180f20bf8c06eaf
+size 108567563

checkpoint-235/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0c7022145e75b26fb5453057f190dec34ae3ed32e656c97df434cf0d46721ea2
+size 14645

checkpoint-235/scaler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3ebf75bfdec9cd91636c8d83fe2e1ebad7190a36eafe90e3fd6e6f00a6e3b2ff
+size 1383

checkpoint-235/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:44fb94a88c15d58ffa1bf978fccff6134483f7613a240fb8f9d498a06f59dd4a
+size 1465

checkpoint-235/tokenizer_config.json

@@ -0,0 +1,15 @@
+{
+  "backend": "tokenizers",
+  "cls_token": "[CLS]",
+  "do_lower_case": true,
+  "is_local": false,
+  "local_files_only": false,
+  "mask_token": "[MASK]",
+  "model_max_length": 512,
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "strip_accents": null,
+  "tokenize_chinese_chars": true,
+  "tokenizer_class": "BertTokenizer",
+  "unk_token": "[UNK]"
+}

checkpoint-235/trainer_state.json

@@ -0,0 +1,250 @@
+{
+  "best_global_step": 235,
+  "best_metric": 0.8399359743897559,
+  "best_model_checkpoint": "models/electra-small-touche-enhanced-binary/trainer/checkpoint-235",
+  "epoch": 5.0,
+  "eval_steps": 500,
+  "global_step": 235,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.2127659574468085,
+      "grad_norm": 0.7465218901634216,
+      "learning_rate": 1.125e-05,
+      "loss": 0.694000244140625,
+      "step": 10
+    },
+    {
+      "epoch": 0.425531914893617,
+      "grad_norm": 1.2217258214950562,
+      "learning_rate": 2.3749999999999998e-05,
+      "loss": 0.6908843994140625,
+      "step": 20
+    },
+    {
+      "epoch": 0.6382978723404256,
+      "grad_norm": 0.6178316473960876,
+      "learning_rate": 2.928909952606635e-05,
+      "loss": 0.6861663818359375,
+      "step": 30
+    },
+    {
+      "epoch": 0.851063829787234,
+      "grad_norm": 2.0509798526763916,
+      "learning_rate": 2.7867298578199053e-05,
+      "loss": 0.6931976318359375,
+      "step": 40
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.65,
+      "eval_fallacy_f1": 0.7307692307692307,
+      "eval_loss": 0.6805615425109863,
+      "eval_macro_f1": 0.6153846153846154,
+      "eval_runtime": 0.2623,
+      "eval_samples_per_second": 762.415,
+      "eval_steps_per_second": 15.248,
+      "step": 47
+    },
+    {
+      "epoch": 1.0638297872340425,
+      "grad_norm": 0.7195119261741638,
+      "learning_rate": 2.6445497630331753e-05,
+      "loss": 0.6897735595703125,
+      "step": 50
+    },
+    {
+      "epoch": 1.2765957446808511,
+      "grad_norm": 1.8138525485992432,
+      "learning_rate": 2.5023696682464456e-05,
+      "loss": 0.6711883544921875,
+      "step": 60
+    },
+    {
+      "epoch": 1.4893617021276595,
+      "grad_norm": 1.5502103567123413,
+      "learning_rate": 2.360189573459716e-05,
+      "loss": 0.66123046875,
+      "step": 70
+    },
+    {
+      "epoch": 1.702127659574468,
+      "grad_norm": 1.2442517280578613,
+      "learning_rate": 2.2180094786729858e-05,
+      "loss": 0.6327056884765625,
+      "step": 80
+    },
+    {
+      "epoch": 1.9148936170212765,
+      "grad_norm": 3.835188150405884,
+      "learning_rate": 2.075829383886256e-05,
+      "loss": 0.615692138671875,
+      "step": 90
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.77,
+      "eval_fallacy_f1": 0.7604166666666666,
+      "eval_loss": 0.5881457328796387,
+      "eval_macro_f1": 0.7696314102564102,
+      "eval_runtime": 0.2664,
+      "eval_samples_per_second": 750.698,
+      "eval_steps_per_second": 15.014,
+      "step": 94
+    },
+    {
+      "epoch": 2.127659574468085,
+      "grad_norm": 1.659440279006958,
+      "learning_rate": 1.933649289099526e-05,
+      "loss": 0.6120407104492187,
+      "step": 100
+    },
+    {
+      "epoch": 2.3404255319148937,
+      "grad_norm": 1.9919922351837158,
+      "learning_rate": 1.791469194312796e-05,
+      "loss": 0.5680572509765625,
+      "step": 110
+    },
+    {
+      "epoch": 2.5531914893617023,
+      "grad_norm": 2.1114556789398193,
+      "learning_rate": 1.6492890995260666e-05,
+      "loss": 0.5408950805664062,
+      "step": 120
+    },
+    {
+      "epoch": 2.7659574468085104,
+      "grad_norm": 2.3676552772521973,
+      "learning_rate": 1.5071090047393365e-05,
+      "loss": 0.49699249267578127,
+      "step": 130
+    },
+    {
+      "epoch": 2.978723404255319,
+      "grad_norm": 2.613635301589966,
+      "learning_rate": 1.3649289099526066e-05,
+      "loss": 0.4867919921875,
+      "step": 140
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.82,
+      "eval_fallacy_f1": 0.8181818181818182,
+      "eval_loss": 0.4984106421470642,
+      "eval_macro_f1": 0.81998199819982,
+      "eval_runtime": 0.5325,
+      "eval_samples_per_second": 375.585,
+      "eval_steps_per_second": 7.512,
+      "step": 141
+    },
+    {
+      "epoch": 3.1914893617021276,
+      "grad_norm": 5.378300189971924,
+      "learning_rate": 1.2227488151658767e-05,
+      "loss": 0.47115631103515626,
+      "step": 150
+    },
+    {
+      "epoch": 3.404255319148936,
+      "grad_norm": 2.245528221130371,
+      "learning_rate": 1.080568720379147e-05,
+      "loss": 0.43046722412109373,
+      "step": 160
+    },
+    {
+      "epoch": 3.617021276595745,
+      "grad_norm": 3.435818672180176,
+      "learning_rate": 9.383886255924171e-06,
+      "loss": 0.4158058166503906,
+      "step": 170
+    },
+    {
+      "epoch": 3.829787234042553,
+      "grad_norm": 3.552385091781616,
+      "learning_rate": 7.962085308056872e-06,
+      "loss": 0.4091644287109375,
+      "step": 180
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": 0.81,
+      "eval_fallacy_f1": 0.8207547169811321,
+      "eval_loss": 0.4597650170326233,
+      "eval_macro_f1": 0.809313528703332,
+      "eval_runtime": 0.2614,
+      "eval_samples_per_second": 765.065,
+      "eval_steps_per_second": 15.301,
+      "step": 188
+    },
+    {
+      "epoch": 4.042553191489362,
+      "grad_norm": 3.47029447555542,
+      "learning_rate": 6.5402843601895735e-06,
+      "loss": 0.40631790161132814,
+      "step": 190
+    },
+    {
+      "epoch": 4.25531914893617,
+      "grad_norm": 5.288370609283447,
+      "learning_rate": 5.1184834123222755e-06,
+      "loss": 0.3656455993652344,
+      "step": 200
+    },
+    {
+      "epoch": 4.468085106382979,
+      "grad_norm": 3.9861114025115967,
+      "learning_rate": 3.696682464454976e-06,
+      "loss": 0.3499275207519531,
+      "step": 210
+    },
+    {
+      "epoch": 4.680851063829787,
+      "grad_norm": 3.6823275089263916,
+      "learning_rate": 2.274881516587678e-06,
+      "loss": 0.38124618530273435,
+      "step": 220
+    },
+    {
+      "epoch": 4.8936170212765955,
+      "grad_norm": 2.9968783855438232,
+      "learning_rate": 8.530805687203791e-07,
+      "loss": 0.3715087890625,
+      "step": 230
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": 0.84,
+      "eval_fallacy_f1": 0.8431372549019608,
+      "eval_loss": 0.43988463282585144,
+      "eval_macro_f1": 0.8399359743897559,
+      "eval_runtime": 0.2604,
+      "eval_samples_per_second": 768.008,
+      "eval_steps_per_second": 15.36,
+      "step": 235
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 235,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 5,
+  "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": 58296167513328.0,
+  "train_batch_size": 16,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-235/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c6cc86806948eeade9111e096da8ecaa7ffd0b7f89647756f199fb88d9a3903b
+size 5393