best model for qa

config.json

@@ -0,0 +1,26 @@
+{
+  "_name_or_path": "/projects/SolarWindowsADSP/shu/models/revision/bert/batterybert-cased/",
+  "architectures": [
+    "BertForQuestionAnswering"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "classifier_dropout": null,
+  "gradient_checkpointing": false,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 768,
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "layer_norm_eps": 1e-12,
+  "max_position_embeddings": 512,
+  "model_type": "bert",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 12,
+  "pad_token_id": 0,
+  "position_embedding_type": "absolute",
+  "torch_dtype": "float32",
+  "transformers_version": "4.17.0.dev0",
+  "type_vocab_size": 2,
+  "use_cache": true,
+  "vocab_size": 28996
+}

optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b6167d07179e12ee2f4c8fd825c6a7aac973f726329edca099ef07191cc0eb86
+size 861894637

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1619a2cdf8533839b44ac017ead99e582e3b856ef8c9c6c23db4c20c6ba8ac7a
+size 430973431

rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ff2d5ce636730f70bfe373ee1c3e846fd9a36e0e67b31556217369cc407dcae7
+size 14657

rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ef3fc4a2d23f9ac92d4a26df5894d9c0551024856ac05f49b00c818c52982410
+size 14657

rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a25568c7ffa5fe9c9556ab19eadf561ce6b6e25806f863cfe0f47e8b7dfdf22c
+size 14657

rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e16be472446729ed097e98c56b1e4411d0b79f7c7aa63a0d9fa909fcac66dbc5
+size 14657

rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7600ccdea87c2c34bc50cbfc117e1b87028fe99b24ea36f17018bbca0dcf98ad
+size 14657

rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:504f99aa72a899d8fdbdd799e5fa64b624c9b0b85f2c3a28615f4d150c7f3e16
+size 14657

rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8c6309c06db648f512a8f8edda137fd2df67c22191e921d5d2cf60298d17f6e6
+size 14657

rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f813994bf1fbba78c215010bb2cd8220372cfabaff81b08ca1ef417457bd0576
+size 14657

scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9bd6a744238c609c30e3be91276089eebd5ddcaa94a27dce9e772fe8020be25c
+size 623

special_tokens_map.json

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

tokenizer_config.json

@@ -0,0 +1 @@
+{"do_lower_case": false, "unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]", "tokenize_chinese_chars": true, "strip_accents": null, "model_max_length": 512, "special_tokens_map_file": null, "name_or_path": "/projects/SolarWindowsADSP/shu/models/revision/bert/batterybert-cased/", "tokenizer_class": "BertTokenizer"}

trainer_state.json

@@ -0,0 +1,304 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 4.0,
+  "global_step": 22184,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9549224666426253e-05,
+      "loss": 2.2682,
+      "step": 500
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.909844933285251e-05,
+      "loss": 1.4753,
+      "step": 1000
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.864767399927876e-05,
+      "loss": 1.3606,
+      "step": 1500
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.8196898665705013e-05,
+      "loss": 1.2586,
+      "step": 2000
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.7746123332131267e-05,
+      "loss": 1.2084,
+      "step": 2500
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.729534799855752e-05,
+      "loss": 1.1679,
+      "step": 3000
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.6844572664983772e-05,
+      "loss": 1.138,
+      "step": 3500
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.6393797331410027e-05,
+      "loss": 1.136,
+      "step": 4000
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.5943021997836278e-05,
+      "loss": 1.1001,
+      "step": 4500
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.5492246664262532e-05,
+      "loss": 1.0728,
+      "step": 5000
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.5041471330688787e-05,
+      "loss": 1.0669,
+      "step": 5500
+    },
+    {
+      "epoch": 1.0,
+      "eval_exact_match": 79.96215704824976,
+      "eval_f1": 87.68130943560381,
+      "step": 5546
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 1.4590695997115038e-05,
+      "loss": 0.8342,
+      "step": 6000
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 1.413992066354129e-05,
+      "loss": 0.8289,
+      "step": 6500
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 1.3689145329967547e-05,
+      "loss": 0.8292,
+      "step": 7000
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 1.3238369996393798e-05,
+      "loss": 0.8103,
+      "step": 7500
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 1.278759466282005e-05,
+      "loss": 0.8194,
+      "step": 8000
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 1.2336819329246306e-05,
+      "loss": 0.8103,
+      "step": 8500
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 1.1886043995672558e-05,
+      "loss": 0.8075,
+      "step": 9000
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 1.143526866209881e-05,
+      "loss": 0.7804,
+      "step": 9500
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 1.0984493328525063e-05,
+      "loss": 0.8155,
+      "step": 10000
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 1.0533717994951317e-05,
+      "loss": 0.795,
+      "step": 10500
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 1.008294266137757e-05,
+      "loss": 0.7755,
+      "step": 11000
+    },
+    {
+      "epoch": 2.0,
+      "eval_exact_match": 81.84484389782403,
+      "eval_f1": 89.0471703841863,
+      "step": 11092
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 9.632167327803823e-06,
+      "loss": 0.6198,
+      "step": 11500
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 9.181391994230077e-06,
+      "loss": 0.6012,
+      "step": 12000
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 8.73061666065633e-06,
+      "loss": 0.5688,
+      "step": 12500
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 8.279841327082583e-06,
+      "loss": 0.574,
+      "step": 13000
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 7.829065993508835e-06,
+      "loss": 0.5873,
+      "step": 13500
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 7.378290659935089e-06,
+      "loss": 0.5984,
+      "step": 14000
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 6.927515326361343e-06,
+      "loss": 0.5877,
+      "step": 14500
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 6.476739992787595e-06,
+      "loss": 0.5892,
+      "step": 15000
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 6.025964659213849e-06,
+      "loss": 0.5722,
+      "step": 15500
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 5.575189325640102e-06,
+      "loss": 0.5848,
+      "step": 16000
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 5.124413992066355e-06,
+      "loss": 0.568,
+      "step": 16500
+    },
+    {
+      "epoch": 3.0,
+      "eval_exact_match": 81.54210028382214,
+      "eval_f1": 88.9922400786148,
+      "step": 16638
+    },
+    {
+      "epoch": 3.07,
+      "learning_rate": 4.673638658492607e-06,
+      "loss": 0.4788,
+      "step": 17000
+    },
+    {
+      "epoch": 3.16,
+      "learning_rate": 4.222863324918861e-06,
+      "loss": 0.4398,
+      "step": 17500
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 3.7720879913451143e-06,
+      "loss": 0.4524,
+      "step": 18000
+    },
+    {
+      "epoch": 3.34,
+      "learning_rate": 3.3213126577713674e-06,
+      "loss": 0.4265,
+      "step": 18500
+    },
+    {
+      "epoch": 3.43,
+      "learning_rate": 2.87053732419762e-06,
+      "loss": 0.4424,
+      "step": 19000
+    },
+    {
+      "epoch": 3.52,
+      "learning_rate": 2.4197619906238733e-06,
+      "loss": 0.4456,
+      "step": 19500
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 1.9689866570501265e-06,
+      "loss": 0.4379,
+      "step": 20000
+    },
+    {
+      "epoch": 3.7,
+      "learning_rate": 1.5182113234763796e-06,
+      "loss": 0.4411,
+      "step": 20500
+    },
+    {
+      "epoch": 3.79,
+      "learning_rate": 1.0674359899026326e-06,
+      "loss": 0.4412,
+      "step": 21000
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 6.166606563288857e-07,
+      "loss": 0.4424,
+      "step": 21500
+    },
+    {
+      "epoch": 3.97,
+      "learning_rate": 1.6588532275513888e-07,
+      "loss": 0.4241,
+      "step": 22000
+    },
+    {
+      "epoch": 4.0,
+      "eval_exact_match": 81.54210028382214,
+      "eval_f1": 89.15597314634459,
+      "step": 22184
+    }
+  ],
+  "max_steps": 22184,
+  "num_train_epochs": 4,
+  "total_flos": 6.955928551581286e+16,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2a63234e1f0e4224cb9feef740c3db5ea28f6b24e417fcd54ef8cf1d66762f4c
+size 3183