Upload 14 files

README.txt

@@ -0,0 +1,22 @@
+Used [run.sh](https://huggingface.co/madlag/bert-large-uncased-whole-word-masking-finetuned-squadv2/blob/main/run.sh) used to train using transformers/example/question_answering code.
+
+Evaluation results : F1= 85.85 , a much better result than the original 81.9 from the BERT paper, due to the use of the "whole-word-masking" variation.
+```
+{
+    "HasAns_exact": 80.58367071524967,
+    "HasAns_f1": 86.64594807945029,
+    "HasAns_total": 5928,
+    "NoAns_exact": 85.06307821698907,
+    "NoAns_f1": 85.06307821698907,
+    "NoAns_total": 5945,
+    "best_exact": 82.82658131895899,
+    "best_exact_thresh": 0.0,
+    "best_f1": 85.85337995578023,
+    "best_f1_thresh": 0.0,
+    "epoch": 2.0,
+    "eval_samples": 12134,
+    "exact": 82.82658131895899,
+    "f1": 85.85337995578037,
+    "total": 11873
+}
+```

all_results.json

@@ -0,0 +1,28 @@
+{
+    "HasAns_exact": 80.58367071524967,
+    "HasAns_f1": 86.64594807945029,
+    "HasAns_total": 5928,
+    "NoAns_exact": 85.06307821698907,
+    "NoAns_f1": 85.06307821698907,
+    "NoAns_total": 5945,
+    "best_exact": 82.82658131895899,
+    "best_exact_thresh": 0.0,
+    "best_f1": 85.85337995578023,
+    "best_f1_thresh": 0.0,
+    "epoch": 2.0,
+    "eval_samples": 12134,
+    "exact": 82.82658131895899,
+    "f1": 85.85337995578037,
+    "init_mem_cpu_alloc_delta": -231366656,
+    "init_mem_cpu_peaked_delta": 1333211136,
+    "init_mem_gpu_alloc_delta": 1337192960,
+    "init_mem_gpu_peaked_delta": 0,
+    "total": 11873,
+    "train_mem_cpu_alloc_delta": 887136256,
+    "train_mem_cpu_peaked_delta": 380289024,
+    "train_mem_gpu_alloc_delta": 4029973504,
+    "train_mem_gpu_peaked_delta": 16453983232,
+    "train_runtime": 13072.0089,
+    "train_samples": 131754,
+    "train_samples_per_second": 1.68
+}

config.json

@@ -0,0 +1,24 @@
+{
+  "_name_or_path": "bert-large-uncased-whole-word-masking",
+  "architectures": [
+    "BertForQuestionAnswering"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "gradient_checkpointing": false,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 1024,
+  "initializer_range": 0.02,
+  "intermediate_size": 4096,
+  "layer_norm_eps": 1e-12,
+  "max_position_embeddings": 512,
+  "model_type": "bert",
+  "num_attention_heads": 16,
+  "num_hidden_layers": 24,
+  "pad_token_id": 0,
+  "position_embedding_type": "absolute",
+  "transformers_version": "4.5.1",
+  "type_vocab_size": 2,
+  "use_cache": true,
+  "vocab_size": 30522
+}

eval_results.json

@@ -0,0 +1,17 @@
+{
+    "HasAns_exact": 80.58367071524967,
+    "HasAns_f1": 86.64594807945029,
+    "HasAns_total": 5928,
+    "NoAns_exact": 85.06307821698907,
+    "NoAns_f1": 85.06307821698907,
+    "NoAns_total": 5945,
+    "best_exact": 82.82658131895899,
+    "best_exact_thresh": 0.0,
+    "best_f1": 85.85337995578023,
+    "best_f1_thresh": 0.0,
+    "epoch": 2.0,
+    "eval_samples": 12134,
+    "exact": 82.82658131895899,
+    "f1": 85.85337995578037,
+    "total": 11873
+}

flax_model.msgpack

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:086eb7284497813d2fa0e20c058a2bed213eaf3149765c9ce85ba4012bc02d60
+size 1336391324

gitattributes.txt

@@ -0,0 +1,17 @@
+*.bin.* filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tar.gz filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:afd0115a49f1d98662f68862670f761a35a905346a6cc82c645ba29a952c9bb2
+size 1336547639

run.sh

@@ -0,0 +1,15 @@
+python run_qa.py \
+  --model_name_or_path bert-large-uncased-whole-word-masking  \
+  --dataset_name squad_v2 \
+  --do_train \
+  --do_eval \
+  --save_steps 2500 \
+  --eval_steps 2500 \
+  --evaluation_strategy steps \
+  --per_device_train_batch_size 12 \
+  --learning_rate 3e-5 \
+  --num_train_epochs 2 \
+  --max_seq_length 384 \
+  --doc_stride 128 \
+  --output_dir /data_2to/devel_data/nn_pruning/output/teacher/squadv2-large-wwm \
+  --version_2_with_negative 1

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, "model_max_length": 512, "special_tokens_map_file": null, "name_or_path": "bert-large-uncased-whole-word-masking"}

train_results.json

@@ -0,0 +1,14 @@
+{
+    "epoch": 2.0,
+    "init_mem_cpu_alloc_delta": -231366656,
+    "init_mem_cpu_peaked_delta": 1333211136,
+    "init_mem_gpu_alloc_delta": 1337192960,
+    "init_mem_gpu_peaked_delta": 0,
+    "train_mem_cpu_alloc_delta": 887136256,
+    "train_mem_cpu_peaked_delta": 380289024,
+    "train_mem_gpu_alloc_delta": 4029973504,
+    "train_mem_gpu_peaked_delta": 16453983232,
+    "train_runtime": 13072.0089,
+    "train_samples": 131754,
+    "train_samples_per_second": 1.68
+}

trainer_state.json

@@ -0,0 +1,417 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 2.0,
+  "global_step": 21960,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.9316939890710385e-05,
+      "loss": 2.1044,
+      "step": 500
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 2.8633879781420765e-05,
+      "loss": 1.5614,
+      "step": 1000
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 2.795081967213115e-05,
+      "loss": 1.4288,
+      "step": 1500
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 2.7267759562841533e-05,
+      "loss": 1.3095,
+      "step": 2000
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 2.6584699453551913e-05,
+      "loss": 1.2225,
+      "step": 2500
+    },
+    {
+      "HasAns_exact": 71.89608636977059,
+      "HasAns_f1": 78.04168619043323,
+      "HasAns_total": 5928,
+      "NoAns_exact": 82.62405382674517,
+      "NoAns_f1": 82.62405382674517,
+      "NoAns_total": 5945,
+      "best_exact": 77.27617282910806,
+      "best_exact_thresh": 0.0,
+      "best_f1": 80.34457304277659,
+      "best_f1_thresh": 0.0,
+      "epoch": 0.23,
+      "exact": 77.26775035795502,
+      "f1": 80.33615057162366,
+      "step": 2500,
+      "total": 11873
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 2.5901639344262294e-05,
+      "loss": 1.2062,
+      "step": 3000
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 2.5218579234972678e-05,
+      "loss": 1.2019,
+      "step": 3500
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 2.453551912568306e-05,
+      "loss": 1.1856,
+      "step": 4000
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 2.3852459016393442e-05,
+      "loss": 1.1677,
+      "step": 4500
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 2.3169398907103826e-05,
+      "loss": 1.1343,
+      "step": 5000
+    },
+    {
+      "HasAns_exact": 74.44331983805668,
+      "HasAns_f1": 80.60592872513861,
+      "HasAns_total": 5928,
+      "NoAns_exact": 83.65012615643398,
+      "NoAns_f1": 83.65012615643398,
+      "NoAns_total": 5945,
+      "best_exact": 79.05331424239871,
+      "best_exact_thresh": 0.0,
+      "best_f1": 82.13020681231532,
+      "best_f1_thresh": 0.0,
+      "epoch": 0.46,
+      "exact": 79.05331424239871,
+      "f1": 82.13020681231545,
+      "step": 5000,
+      "total": 11873
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 2.248633879781421e-05,
+      "loss": 1.1489,
+      "step": 5500
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 2.180327868852459e-05,
+      "loss": 1.1349,
+      "step": 6000
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 2.1120218579234974e-05,
+      "loss": 1.1156,
+      "step": 6500
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 2.0437158469945358e-05,
+      "loss": 1.1066,
+      "step": 7000
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.975409836065574e-05,
+      "loss": 1.0877,
+      "step": 7500
+    },
+    {
+      "HasAns_exact": 80.61740890688259,
+      "HasAns_f1": 87.36413644383099,
+      "HasAns_total": 5928,
+      "NoAns_exact": 72.31286795626578,
+      "NoAns_f1": 72.31286795626578,
+      "NoAns_total": 5945,
+      "best_exact": 76.45919312726353,
+      "best_exact_thresh": 0.0,
+      "best_f1": 79.82772684570281,
+      "best_f1_thresh": 0.0,
+      "epoch": 0.68,
+      "exact": 76.45919312726353,
+      "f1": 79.82772684570291,
+      "step": 7500,
+      "total": 11873
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.907103825136612e-05,
+      "loss": 1.0836,
+      "step": 8000
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.8387978142076503e-05,
+      "loss": 1.0793,
+      "step": 8500
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.7704918032786887e-05,
+      "loss": 1.0722,
+      "step": 9000
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.7021857923497267e-05,
+      "loss": 1.0497,
+      "step": 9500
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.633879781420765e-05,
+      "loss": 0.8443,
+      "step": 10000
+    },
+    {
+      "HasAns_exact": 77.7327935222672,
+      "HasAns_f1": 84.30659896226393,
+      "HasAns_total": 5928,
+      "NoAns_exact": 81.19428090832632,
+      "NoAns_f1": 81.19428090832632,
+      "NoAns_total": 5945,
+      "best_exact": 79.4660153288975,
+      "best_exact_thresh": 0.0,
+      "best_f1": 82.7482117955276,
+      "best_f1_thresh": 0.0,
+      "epoch": 0.91,
+      "exact": 79.4660153288975,
+      "f1": 82.74821179552765,
+      "step": 10000,
+      "total": 11873
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.5655737704918035e-05,
+      "loss": 0.7997,
+      "step": 10500
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.4972677595628415e-05,
+      "loss": 0.7642,
+      "step": 11000
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 1.4289617486338798e-05,
+      "loss": 0.4993,
+      "step": 11500
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 1.3606557377049181e-05,
+      "loss": 0.4945,
+      "step": 12000
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 1.2923497267759564e-05,
+      "loss": 0.5155,
+      "step": 12500
+    },
+    {
+      "HasAns_exact": 75.79284750337382,
+      "HasAns_f1": 82.13405464912947,
+      "HasAns_total": 5928,
+      "NoAns_exact": 86.67788057190917,
+      "NoAns_f1": 86.67788057190917,
+      "NoAns_total": 5945,
+      "best_exact": 81.24315674218816,
+      "best_exact_thresh": 0.0,
+      "best_f1": 84.4092205811537,
+      "best_f1_thresh": 0.0,
+      "epoch": 1.14,
+      "exact": 81.24315674218816,
+      "f1": 84.40922058115372,
+      "step": 12500,
+      "total": 11873
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 1.2240437158469946e-05,
+      "loss": 0.4964,
+      "step": 13000
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 1.1557377049180328e-05,
+      "loss": 0.4947,
+      "step": 13500
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 1.087431693989071e-05,
+      "loss": 0.5173,
+      "step": 14000
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 1.0191256830601094e-05,
+      "loss": 0.4718,
+      "step": 14500
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 9.508196721311476e-06,
+      "loss": 0.4768,
+      "step": 15000
+    },
+    {
+      "HasAns_exact": 77.83400809716599,
+      "HasAns_f1": 84.204371493218,
+      "HasAns_total": 5928,
+      "NoAns_exact": 85.23128679562657,
+      "NoAns_f1": 85.23128679562657,
+      "NoAns_total": 5945,
+      "best_exact": 81.53794323254444,
+      "best_exact_thresh": 0.0,
+      "best_f1": 84.71856432340547,
+      "best_f1_thresh": 0.0,
+      "epoch": 1.37,
+      "exact": 81.53794323254444,
+      "f1": 84.71856432340567,
+      "step": 15000,
+      "total": 11873
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 8.825136612021857e-06,
+      "loss": 0.4937,
+      "step": 15500
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 8.14207650273224e-06,
+      "loss": 0.4678,
+      "step": 16000
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 7.459016393442623e-06,
+      "loss": 0.4922,
+      "step": 16500
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 6.775956284153006e-06,
+      "loss": 0.4779,
+      "step": 17000
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 6.092896174863388e-06,
+      "loss": 0.4787,
+      "step": 17500
+    },
+    {
+      "HasAns_exact": 79.30161943319838,
+      "HasAns_f1": 85.07291240565793,
+      "HasAns_total": 5928,
+      "NoAns_exact": 85.5004205214466,
+      "NoAns_f1": 85.5004205214466,
+      "NoAns_total": 5945,
+      "best_exact": 82.40545776130716,
+      "best_exact_thresh": 0.0,
+      "best_f1": 85.28697252090775,
+      "best_f1_thresh": 0.0,
+      "epoch": 1.59,
+      "exact": 82.40545776130716,
+      "f1": 85.28697252090787,
+      "step": 17500,
+      "total": 11873
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 5.409836065573771e-06,
+      "loss": 0.4554,
+      "step": 18000
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 4.726775956284153e-06,
+      "loss": 0.4661,
+      "step": 18500
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 4.043715846994535e-06,
+      "loss": 0.449,
+      "step": 19000
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 3.3606557377049183e-06,
+      "loss": 0.4426,
+      "step": 19500
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 2.6775956284153005e-06,
+      "loss": 0.4732,
+      "step": 20000
+    },
+    {
+      "HasAns_exact": 81.29217273954116,
+      "HasAns_f1": 87.36118517518237,
+      "HasAns_total": 5928,
+      "NoAns_exact": 82.64087468460892,
+      "NoAns_f1": 82.64087468460892,
+      "NoAns_total": 5945,
+      "best_exact": 81.96748926134929,
+      "best_exact_thresh": 0.0,
+      "best_f1": 84.99765061218557,
+      "best_f1_thresh": 0.0,
+      "epoch": 1.82,
+      "exact": 81.96748926134929,
+      "f1": 84.99765061218568,
+      "step": 20000,
+      "total": 11873
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 1.994535519125683e-06,
+      "loss": 0.4269,
+      "step": 20500
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 1.3114754098360657e-06,
+      "loss": 0.4461,
+      "step": 21000
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 6.284153005464482e-07,
+      "loss": 0.434,
+      "step": 21500
+    },
+    {
+      "epoch": 2.0,
+      "step": 21960,
+      "total_flos": 2.0283616700399e+17,
+      "train_runtime": 13072.0089,
+      "train_samples_per_second": 1.68
+    }
+  ],
+  "max_steps": 21960,
+  "num_train_epochs": 2,
+  "total_flos": 2.0283616700399e+17,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d0dcc35c32983a4419cf8338e2803e51d726a45e398611409ecc0aa1a3bbfcf2
+size 2415