Add other files from fine-tunning
Browse filesDelete eval_nbest_predictions.json due to size limit
- README.md +20 -0
- all_results.json +22 -0
- eval_null_odds.json +0 -0
- eval_predictions.json +0 -0
- eval_results.json +17 -0
- train_results.json +8 -0
- trainer_state.json +289 -0
- training_args.bin +3 -0
- tune_qa.sh +16 -0
README.md
ADDED
|
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
**[`microsoft/BiomedNLP-PubMedBERT-base-uncased-abstract-fulltext`](https://huggingface.co/microsoft/BiomedNLP-PubMedBERT-base-uncased-abstract-fulltext)** fine-tuned on **[`SQuAD V2`](https://rajpurkar.github.io/SQuAD-explorer/)** using **[`run_qa.py`](https://github.com/huggingface/transformers/blob/master/examples/pytorch/question-answering/run_qa.py)**
|
| 2 |
+
|
| 3 |
+
Tunning script:
|
| 4 |
+
```bash
|
| 5 |
+
BASE_MODEL=microsoft/BiomedNLP-PubMedBERT-base-uncased-abstract-fulltext
|
| 6 |
+
OUTPUT_DIR=~/Documents/projects/tunned_models/ms_pubmed_bert_squadv2/
|
| 7 |
+
|
| 8 |
+
python run_qa.py \
|
| 9 |
+
--model_name_or_path $BASE_MODEL\
|
| 10 |
+
--dataset_name squad_v2 \
|
| 11 |
+
--do_train \
|
| 12 |
+
--do_eval \
|
| 13 |
+
--version_2_with_negative \
|
| 14 |
+
--per_device_train_batch_size 12 \
|
| 15 |
+
--learning_rate 3e-5 \
|
| 16 |
+
--num_train_epochs 2 \
|
| 17 |
+
--max_seq_length 384 \
|
| 18 |
+
--doc_stride 128 \
|
| 19 |
+
--output_dir $OUTPUT_DIR
|
| 20 |
+
```
|
all_results.json
ADDED
|
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.0,
|
| 3 |
+
"eval_HasAns_exact": 73.26248313090419,
|
| 4 |
+
"eval_HasAns_f1": 79.53074066003731,
|
| 5 |
+
"eval_HasAns_total": 5928,
|
| 6 |
+
"eval_NoAns_exact": 81.53069806560134,
|
| 7 |
+
"eval_NoAns_f1": 81.53069806560134,
|
| 8 |
+
"eval_NoAns_total": 5945,
|
| 9 |
+
"eval_best_exact": 77.4025098964036,
|
| 10 |
+
"eval_best_exact_thresh": 0.0,
|
| 11 |
+
"eval_best_f1": 80.53215115242146,
|
| 12 |
+
"eval_best_f1_thresh": 0.0,
|
| 13 |
+
"eval_exact": 77.4025098964036,
|
| 14 |
+
"eval_f1": 80.53215115242146,
|
| 15 |
+
"eval_samples": 12271,
|
| 16 |
+
"eval_total": 11873,
|
| 17 |
+
"train_loss": 0.9526093118189042,
|
| 18 |
+
"train_runtime": 484536.8861,
|
| 19 |
+
"train_samples": 132768,
|
| 20 |
+
"train_samples_per_second": 0.548,
|
| 21 |
+
"train_steps_per_second": 0.046
|
| 22 |
+
}
|
eval_null_odds.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
eval_predictions.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
eval_results.json
ADDED
|
@@ -0,0 +1,17 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.0,
|
| 3 |
+
"eval_HasAns_exact": 73.26248313090419,
|
| 4 |
+
"eval_HasAns_f1": 79.53074066003731,
|
| 5 |
+
"eval_HasAns_total": 5928,
|
| 6 |
+
"eval_NoAns_exact": 81.53069806560134,
|
| 7 |
+
"eval_NoAns_f1": 81.53069806560134,
|
| 8 |
+
"eval_NoAns_total": 5945,
|
| 9 |
+
"eval_best_exact": 77.4025098964036,
|
| 10 |
+
"eval_best_exact_thresh": 0.0,
|
| 11 |
+
"eval_best_f1": 80.53215115242146,
|
| 12 |
+
"eval_best_f1_thresh": 0.0,
|
| 13 |
+
"eval_exact": 77.4025098964036,
|
| 14 |
+
"eval_f1": 80.53215115242146,
|
| 15 |
+
"eval_samples": 12271,
|
| 16 |
+
"eval_total": 11873
|
| 17 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.0,
|
| 3 |
+
"train_loss": 0.9526093118189042,
|
| 4 |
+
"train_runtime": 484536.8861,
|
| 5 |
+
"train_samples": 132768,
|
| 6 |
+
"train_samples_per_second": 0.548,
|
| 7 |
+
"train_steps_per_second": 0.046
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,289 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.0,
|
| 5 |
+
"global_step": 22128,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.05,
|
| 12 |
+
"learning_rate": 2.9322125813449026e-05,
|
| 13 |
+
"loss": 2.223,
|
| 14 |
+
"step": 500
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.09,
|
| 18 |
+
"learning_rate": 2.8644251626898048e-05,
|
| 19 |
+
"loss": 1.5712,
|
| 20 |
+
"step": 1000
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.14,
|
| 24 |
+
"learning_rate": 2.7966377440347073e-05,
|
| 25 |
+
"loss": 1.4365,
|
| 26 |
+
"step": 1500
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.18,
|
| 30 |
+
"learning_rate": 2.7288503253796095e-05,
|
| 31 |
+
"loss": 1.3255,
|
| 32 |
+
"step": 2000
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.23,
|
| 36 |
+
"learning_rate": 2.661062906724512e-05,
|
| 37 |
+
"loss": 1.2879,
|
| 38 |
+
"step": 2500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.27,
|
| 42 |
+
"learning_rate": 2.5932754880694143e-05,
|
| 43 |
+
"loss": 1.2262,
|
| 44 |
+
"step": 3000
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.32,
|
| 48 |
+
"learning_rate": 2.5254880694143168e-05,
|
| 49 |
+
"loss": 1.2299,
|
| 50 |
+
"step": 3500
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.36,
|
| 54 |
+
"learning_rate": 2.457700650759219e-05,
|
| 55 |
+
"loss": 1.1777,
|
| 56 |
+
"step": 4000
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.41,
|
| 60 |
+
"learning_rate": 2.3899132321041215e-05,
|
| 61 |
+
"loss": 1.1717,
|
| 62 |
+
"step": 4500
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.45,
|
| 66 |
+
"learning_rate": 2.322125813449024e-05,
|
| 67 |
+
"loss": 1.1411,
|
| 68 |
+
"step": 5000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.5,
|
| 72 |
+
"learning_rate": 2.2543383947939266e-05,
|
| 73 |
+
"loss": 1.1099,
|
| 74 |
+
"step": 5500
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.54,
|
| 78 |
+
"learning_rate": 2.1865509761388288e-05,
|
| 79 |
+
"loss": 1.1054,
|
| 80 |
+
"step": 6000
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.59,
|
| 84 |
+
"learning_rate": 2.1187635574837313e-05,
|
| 85 |
+
"loss": 1.0961,
|
| 86 |
+
"step": 6500
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.63,
|
| 90 |
+
"learning_rate": 2.0509761388286335e-05,
|
| 91 |
+
"loss": 1.0753,
|
| 92 |
+
"step": 7000
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.68,
|
| 96 |
+
"learning_rate": 1.983188720173536e-05,
|
| 97 |
+
"loss": 1.0311,
|
| 98 |
+
"step": 7500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.72,
|
| 102 |
+
"learning_rate": 1.9154013015184382e-05,
|
| 103 |
+
"loss": 1.0377,
|
| 104 |
+
"step": 8000
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.77,
|
| 108 |
+
"learning_rate": 1.8476138828633408e-05,
|
| 109 |
+
"loss": 1.0191,
|
| 110 |
+
"step": 8500
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.81,
|
| 114 |
+
"learning_rate": 1.779826464208243e-05,
|
| 115 |
+
"loss": 0.9882,
|
| 116 |
+
"step": 9000
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.86,
|
| 120 |
+
"learning_rate": 1.7120390455531455e-05,
|
| 121 |
+
"loss": 0.9928,
|
| 122 |
+
"step": 9500
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.9,
|
| 126 |
+
"learning_rate": 1.6442516268980477e-05,
|
| 127 |
+
"loss": 0.9701,
|
| 128 |
+
"step": 10000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.95,
|
| 132 |
+
"learning_rate": 1.5764642082429502e-05,
|
| 133 |
+
"loss": 0.9901,
|
| 134 |
+
"step": 10500
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.99,
|
| 138 |
+
"learning_rate": 1.5086767895878524e-05,
|
| 139 |
+
"loss": 0.9647,
|
| 140 |
+
"step": 11000
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 1.04,
|
| 144 |
+
"learning_rate": 1.4408893709327548e-05,
|
| 145 |
+
"loss": 0.784,
|
| 146 |
+
"step": 11500
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 1.08,
|
| 150 |
+
"learning_rate": 1.3731019522776571e-05,
|
| 151 |
+
"loss": 0.7568,
|
| 152 |
+
"step": 12000
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 1.13,
|
| 156 |
+
"learning_rate": 1.3053145336225597e-05,
|
| 157 |
+
"loss": 0.7582,
|
| 158 |
+
"step": 12500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 1.17,
|
| 162 |
+
"learning_rate": 1.237527114967462e-05,
|
| 163 |
+
"loss": 0.7732,
|
| 164 |
+
"step": 13000
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 1.22,
|
| 168 |
+
"learning_rate": 1.1697396963123646e-05,
|
| 169 |
+
"loss": 0.7331,
|
| 170 |
+
"step": 13500
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 1.27,
|
| 174 |
+
"learning_rate": 1.101952277657267e-05,
|
| 175 |
+
"loss": 0.7623,
|
| 176 |
+
"step": 14000
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 1.31,
|
| 180 |
+
"learning_rate": 1.0341648590021693e-05,
|
| 181 |
+
"loss": 0.7403,
|
| 182 |
+
"step": 14500
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 1.36,
|
| 186 |
+
"learning_rate": 9.663774403470717e-06,
|
| 187 |
+
"loss": 0.7096,
|
| 188 |
+
"step": 15000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 1.4,
|
| 192 |
+
"learning_rate": 8.98590021691974e-06,
|
| 193 |
+
"loss": 0.7049,
|
| 194 |
+
"step": 15500
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 1.45,
|
| 198 |
+
"learning_rate": 8.308026030368764e-06,
|
| 199 |
+
"loss": 0.7097,
|
| 200 |
+
"step": 16000
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 1.49,
|
| 204 |
+
"learning_rate": 7.630151843817788e-06,
|
| 205 |
+
"loss": 0.7169,
|
| 206 |
+
"step": 16500
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 1.54,
|
| 210 |
+
"learning_rate": 6.952277657266811e-06,
|
| 211 |
+
"loss": 0.7172,
|
| 212 |
+
"step": 17000
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 1.58,
|
| 216 |
+
"learning_rate": 6.274403470715835e-06,
|
| 217 |
+
"loss": 0.7446,
|
| 218 |
+
"step": 17500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 1.63,
|
| 222 |
+
"learning_rate": 5.5965292841648585e-06,
|
| 223 |
+
"loss": 0.6982,
|
| 224 |
+
"step": 18000
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 1.67,
|
| 228 |
+
"learning_rate": 4.918655097613883e-06,
|
| 229 |
+
"loss": 0.6878,
|
| 230 |
+
"step": 18500
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 1.72,
|
| 234 |
+
"learning_rate": 4.240780911062907e-06,
|
| 235 |
+
"loss": 0.6856,
|
| 236 |
+
"step": 19000
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 1.76,
|
| 240 |
+
"learning_rate": 3.5629067245119307e-06,
|
| 241 |
+
"loss": 0.6951,
|
| 242 |
+
"step": 19500
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 1.81,
|
| 246 |
+
"learning_rate": 2.8850325379609543e-06,
|
| 247 |
+
"loss": 0.6837,
|
| 248 |
+
"step": 20000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 1.85,
|
| 252 |
+
"learning_rate": 2.2071583514099784e-06,
|
| 253 |
+
"loss": 0.6826,
|
| 254 |
+
"step": 20500
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 1.9,
|
| 258 |
+
"learning_rate": 1.529284164859002e-06,
|
| 259 |
+
"loss": 0.6854,
|
| 260 |
+
"step": 21000
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 1.94,
|
| 264 |
+
"learning_rate": 8.514099783080261e-07,
|
| 265 |
+
"loss": 0.6944,
|
| 266 |
+
"step": 21500
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 1.99,
|
| 270 |
+
"learning_rate": 1.735357917570499e-07,
|
| 271 |
+
"loss": 0.6932,
|
| 272 |
+
"step": 22000
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 2.0,
|
| 276 |
+
"step": 22128,
|
| 277 |
+
"total_flos": 6.662030063085158e+16,
|
| 278 |
+
"train_loss": 0.9526093118189042,
|
| 279 |
+
"train_runtime": 484536.8861,
|
| 280 |
+
"train_samples_per_second": 0.548,
|
| 281 |
+
"train_steps_per_second": 0.046
|
| 282 |
+
}
|
| 283 |
+
],
|
| 284 |
+
"max_steps": 22128,
|
| 285 |
+
"num_train_epochs": 2,
|
| 286 |
+
"total_flos": 6.662030063085158e+16,
|
| 287 |
+
"trial_name": null,
|
| 288 |
+
"trial_params": null
|
| 289 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0890f8a5efa1bdca596e7c726964dc8817011914a614ad6704f0500520ac917a
|
| 3 |
+
size 2735
|
tune_qa.sh
ADDED
|
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
BASE_MODEL=microsoft/BiomedNLP-PubMedBERT-base-uncased-abstract-fulltext
|
| 2 |
+
OUTPUT_DIR=~/Documents/projects/tunned_models/ms_pubmed_bert_squadv2/
|
| 3 |
+
|
| 4 |
+
python run_qa.py \
|
| 5 |
+
--model_name_or_path $BASE_MODEL\
|
| 6 |
+
--dataset_name squad_v2 \
|
| 7 |
+
--do_train \
|
| 8 |
+
--do_eval \
|
| 9 |
+
--version_2_with_negative \
|
| 10 |
+
--per_device_train_batch_size 12 \
|
| 11 |
+
--learning_rate 3e-5 \
|
| 12 |
+
--num_train_epochs 2 \
|
| 13 |
+
--max_seq_length 384 \
|
| 14 |
+
--doc_stride 128 \
|
| 15 |
+
--output_dir $OUTPUT_DIR
|
| 16 |
+
|