best model for qa

config.json

@@ -0,0 +1,26 @@
+{
+  "_name_or_path": "/projects/SolarWindowsADSP/shu/models/bert-base-uncased/",
+  "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": 30522
+}

optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f43a75852015858b4f106fcf3b97557dc1e44fed856987c34a1ce92aae7c50c9
+size 871270381

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f59d4c23109d477de53d64f22fdeb9b2a19d9563d6634bbef8e382ab016e2989
+size 435661303

rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c3ef130b5be4c23ac78d5c78b2c74ac6faeea04b06d20a94ecb5d477d4dc0ec5
+size 14657

rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4f2cac8605cdcba7c1f412fa353f4216e28b0c56e8385bb5e33aa0b696c419fa
+size 14657

rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:68bfe7b19fdbfe8a1347f0b6e44e861829c4aacdae1b6e4cb8e4fe2989bf4f36
+size 14657

rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:79cc1b0c72163eebc7ffdb3bf637cf2411c0c574309e12334041f4f325127fd8
+size 14657

rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e9bebb3c4fd0377bb49c4dda50fa3ed71b8ccbee13fc05dfc937ddf6e137f922
+size 14657

rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d86e4b0e4155b80c9a70fd251447e043e8eb7780fd5b187d9da1d1fffa48eee8
+size 14657

rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d2a115a3bf0cc9c0979421bd200df703e8202da9b6c05157e002dab7cc240105
+size 14657

rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:38fe06e41fd4720adca3374d9968468fee020bfc92bd2b70c2fed12b73f56458
+size 14657

scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9d76a2ae72c909de0bf20a045aa9629e9f08ec33067dbaa2c8defe8ed942edad
+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": true, "unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]", "tokenize_chinese_chars": true, "strip_accents": null, "special_tokens_map_file": null, "name_or_path": "/projects/SolarWindowsADSP/shu/models/bert-base-uncased/", "tokenizer_class": "BertTokenizer"}

trainer_state.json

@@ -0,0 +1,130 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "global_step": 8301,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.18,
+      "learning_rate": 2.8644741597397903e-05,
+      "loss": 2.0959,
+      "step": 500
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 2.728948319479581e-05,
+      "loss": 1.3446,
+      "step": 1000
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 2.5934224792193712e-05,
+      "loss": 1.2113,
+      "step": 1500
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 2.4578966389591615e-05,
+      "loss": 1.1509,
+      "step": 2000
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 2.3223707986989517e-05,
+      "loss": 1.0866,
+      "step": 2500
+    },
+    {
+      "epoch": 1.0,
+      "eval_exact_match": 79.10122989593188,
+      "eval_f1": 86.91234473539434,
+      "step": 2767
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 2.1868449584387424e-05,
+      "loss": 0.9377,
+      "step": 3000
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 2.051319118178533e-05,
+      "loss": 0.7975,
+      "step": 3500
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 1.9157932779183232e-05,
+      "loss": 0.814,
+      "step": 4000
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 1.7802674376581135e-05,
+      "loss": 0.7997,
+      "step": 4500
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 1.644741597397904e-05,
+      "loss": 0.8012,
+      "step": 5000
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 1.5092157571376944e-05,
+      "loss": 0.791,
+      "step": 5500
+    },
+    {
+      "epoch": 2.0,
+      "eval_exact_match": 79.78240302743615,
+      "eval_f1": 87.66969355486262,
+      "step": 5534
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 1.3736899168774846e-05,
+      "loss": 0.5813,
+      "step": 6000
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 1.238164076617275e-05,
+      "loss": 0.571,
+      "step": 6500
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 1.1026382363570653e-05,
+      "loss": 0.5749,
+      "step": 7000
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 9.671123960968558e-06,
+      "loss": 0.5654,
+      "step": 7500
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 8.315865558366462e-06,
+      "loss": 0.568,
+      "step": 8000
+    },
+    {
+      "epoch": 3.0,
+      "eval_exact_match": 80.9271523178808,
+      "eval_f1": 88.19682357031269,
+      "step": 8301
+    }
+  ],
+  "max_steps": 11068,
+  "num_train_epochs": 4,
+  "total_flos": 5.204717742535475e+16,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2c09975532a15ca0dbb6b1282a1f3b0496193a23e1f7547e78234b8c6ccbaa2b
+size 3119