add model

added_tokens.json

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

config.json

@@ -0,0 +1,41 @@
+{
+  "_name_or_path": "microsoft/deberta-v3-large",
+  "architectures": [
+    "DebertaV2ForSequenceClassification"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 1024,
+  "id2label": {
+    "0": "LABEL_0"
+  },
+  "initializer_range": 0.02,
+  "intermediate_size": 4096,
+  "label2id": {
+    "LABEL_0": 0
+  },
+  "layer_norm_eps": 1e-07,
+  "max_position_embeddings": 512,
+  "max_relative_positions": -1,
+  "model_type": "deberta-v2",
+  "norm_rel_ebd": "layer_norm",
+  "num_attention_heads": 16,
+  "num_hidden_layers": 24,
+  "pad_token_id": 0,
+  "pooler_dropout": 0,
+  "pooler_hidden_act": "gelu",
+  "pooler_hidden_size": 1024,
+  "pos_att_type": [
+    "p2c",
+    "c2p"
+  ],
+  "position_biased_input": false,
+  "position_buckets": 256,
+  "relative_attention": true,
+  "share_att_key": true,
+  "torch_dtype": "float32",
+  "transformers_version": "4.25.1",
+  "type_vocab_size": 0,
+  "vocab_size": 128100
+}

eval.jsonl

@@ -0,0 +1,6 @@
+{"eval_loss": 0.6821334362030029, "eval_rmse": 0.8259137272834778, "eval_spearman": 0.6366639193422842, "eval_pearson": 0.6731962999298167, "eval_runtime": 2.0034, "eval_samples_per_second": 120.795, "eval_steps_per_second": 15.474, "epoch": 1.0, "eval_set": "osu"}
+{"eval_loss": 1.3671470880508423, "eval_rmse": 1.1692506074905396, "eval_spearman": 0.46622255846621397, "eval_pearson": 0.4136858459894459, "eval_runtime": 13.114, "eval_samples_per_second": 139.012, "eval_steps_per_second": 17.386, "epoch": 1.0, "eval_set": "healthver"}
+{"eval_loss": 0.7539231181144714, "eval_rmse": 0.8682874441146851, "eval_spearman": 0.5488431129607819, "eval_pearson": 0.538707782127853, "eval_runtime": 0.8721, "eval_samples_per_second": 136.452, "eval_steps_per_second": 17.2, "epoch": 1.0, "eval_set": "deepset_1"}
+{"eval_loss": 0.49185943603515625, "eval_rmse": 0.7013269662857056, "eval_spearman": 0.6717814226965486, "eval_pearson": 0.691180339379398, "eval_runtime": 1.1756, "eval_samples_per_second": 132.7, "eval_steps_per_second": 17.013, "epoch": 1.0, "eval_set": "deepset_2"}
+{"eval_loss": 0.9618063569068909, "eval_rmse": 0.9807172417640686, "eval_spearman": 0.3984599642180852, "eval_pearson": 0.41185545800624546, "eval_runtime": 2.5757, "eval_samples_per_second": 134.33, "eval_steps_per_second": 17.082, "epoch": 1.0, "eval_set": "deepset_3"}
+{"eval_loss": 0.7669643759727478, "eval_rmse": 0.8757649064064026, "eval_spearman": 0.4542610937998787, "eval_pearson": 0.4986133277496825, "eval_runtime": 1.5198, "eval_samples_per_second": 135.546, "eval_steps_per_second": 17.108, "epoch": 1.0, "eval_set": "deepset_4"}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2085c6d3c342fc79ad74d556a5a86c45a92a27f914fe8e768c51d211c4ee92a7
+size 1740389291

special_tokens_map.json

@@ -0,0 +1,9 @@
+{
+  "bos_token": "[CLS]",
+  "cls_token": "[CLS]",
+  "eos_token": "[SEP]",
+  "mask_token": "[MASK]",
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "unk_token": "[UNK]"
+}

spm.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c679fbf93643d19aab7ee10c0b99e460bdbc02fedf34b92b05af343b4af586fd
+size 2464616

tokenizer_config.json

@@ -0,0 +1,17 @@
+{
+  "bos_token": "[CLS]",
+  "cls_token": "[CLS]",
+  "do_lower_case": false,
+  "eos_token": "[SEP]",
+  "mask_token": "[MASK]",
+  "model_max_length": 1000000000000000019884624838656,
+  "name_or_path": "microsoft/deberta-v3-large",
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "sp_model_kwargs": {},
+  "special_tokens_map_file": null,
+  "split_by_punct": false,
+  "tokenizer_class": "DebertaV2Tokenizer",
+  "unk_token": "[UNK]",
+  "vocab_type": "spm"
+}

training_cfg.json

@@ -0,0 +1,27 @@
+{
+  "model_name": "microsoft/deberta-v3-large",
+  "lr": 6e-06,
+  "epochs": 1,
+  "weight_decay": 0.01,
+  "train_bs": 3,
+  "eval_bs": 8,
+  "gradient_accumulation_steps": 8,
+  "gradient_checkpointing": false,
+  "eval_datasets": [
+    {"path": "./data/all_eval/osu_eval.csv", "name": "osu"},
+    {"path": "./data/all_eval/healthver_eval.csv", "name": "healthver"},
+    {"path": "./data/all_eval/deepset_1_eval.csv", "name": "deepset_1"},
+    {"path": "./data/all_eval/deepset_2_eval.csv", "name": "deepset_2"},
+    {"path": "./data/all_eval/deepset_3_eval.csv", "name": "deepset_3"},
+    {"path": "./data/all_eval/deepset_4_eval.csv", "name": "deepset_4"}
+  ],
+  "train_dataset_path": "./data/all_training/all_train_v2_pseudo.csv",
+  "ensemble_model_predictions": [
+        "mathislucka/deberta-large-hallucination-eval-v2",
+        "mathislucka/deberta-base-hallucination-eval-v2",
+        "models/albert-xxlarge-v2-optim-data-v1",
+        "models/deberta-base-v3-no-atomic-wfc-nq",
+        "models/deberta-v3-large-data-optim-v2"
+  ],
+  "half_precision": true
+}