diff options
| -rw-r--r-- | fine_tune/bert/demo.py | 81 | ||||
| -rw-r--r-- | fine_tune/bert/tutorials/00_bert_model_architecture_params.ipynb | 858 | ||||
| -rw-r--r-- | fine_tune/bert_parameters.py | 42 | ||||
| -rw-r--r-- | web/simu/gravity.html | 118 |
4 files changed, 1099 insertions, 0 deletions
diff --git a/fine_tune/bert/demo.py b/fine_tune/bert/demo.py new file mode 100644 index 0000000..c64f30f --- /dev/null +++ b/fine_tune/bert/demo.py @@ -0,0 +1,81 @@ + +import torch +import re +from transformers import BertTokenizer +import pandas as pd + + +if torch.cuda.is_available(): + device = torch.device("cuda") + print(f'There are {torch.cuda.device_count()} GPU(s) available.') + print('Device name:', torch.cuda.get_device_name(0)) + +else: + print('No GPU available, using the CPU instead.') + device = torch.device("cpu") + + +def text_preprocessing(text): + """ + - Remove entity mentions (eg. '@united') + - Correct errors (eg. '&' to '&') + @param text (str): a string to be processed. + @return text (Str): the processed string. + """ + # Remove '@name' + text = re.sub(r'(@.*?)[\s]', ' ', text) + + # Replace '&' with '&' + text = re.sub(r'&', '&', text) + + # Remove trailing whitespace + text = re.sub(r'\s+', ' ', text).strip() + + return text + + +model_name = 'bert-base-uncased' +tokenizer = BertTokenizer.from_pretrained('bert-base-uncased', do_lower_case=True) +# Specify `MAX_LEN` +MAX_LEN = 64 + +def preprocessing_for_bert(data): + """Perform required preprocessing steps for pretrained BERT. + @param data (np.array): Array of texts to be processed. + @return input_ids (torch.Tensor): Tensor of token ids to be fed to a model. + @return attention_masks (torch.Tensor): Tensor of indices specifying which + tokens should be attended to by the model. + """ + # Create empty lists to store outputs + input_ids = [] + attention_masks = [] + + # For every sentence... + for sent in data: + # `encode_plus` will: + # (1) Tokenize the sentence + # (2) Add the `[CLS]` and `[SEP]` token to the start and end + # (3) Truncate/Pad sentence to max length + # (4) Map tokens to their IDs + # (5) Create attention mask + # (6) Return a dictionary of outputs + encoded_sent = tokenizer.encode_plus( + text=text_preprocessing(sent), # Preprocess sentence + add_special_tokens=True, # Add `[CLS]` and `[SEP]` + max_length=MAX_LEN, # Max length to truncate/pad + pad_to_max_length=True, # Pad sentence to max length + # return_tensors='pt', # Return PyTorch tensor + return_attention_mask=True # Return attention mask + ) + + # Add the outputs to the lists + input_ids.append(encoded_sent.get('input_ids')) + attention_masks.append(encoded_sent.get('attention_mask')) + + # Convert lists to tensors + input_ids = torch.tensor(input_ids) + attention_masks = torch.tensor(attention_masks) + + return input_ids, attention_masks + + diff --git a/fine_tune/bert/tutorials/00_bert_model_architecture_params.ipynb b/fine_tune/bert/tutorials/00_bert_model_architecture_params.ipynb new file mode 100644 index 0000000..2c36bd2 --- /dev/null +++ b/fine_tune/bert/tutorials/00_bert_model_architecture_params.ipynb @@ -0,0 +1,858 @@ +{ + "cells": [ + { + "cell_type": "code", + "execution_count": 5, + "metadata": { + "ExecuteTime": { + "end_time": "2022-06-27T14:30:07.738410Z", + "start_time": "2022-06-27T14:30:07.735883Z" + } + }, + "outputs": [], + "source": [ + "from transformers import BertModel, BertForSequenceClassification" + ] + }, + { + "cell_type": "code", + "execution_count": 2, + "metadata": { + "ExecuteTime": { + "end_time": "2022-06-27T14:28:30.235712Z", + "start_time": "2022-06-27T14:28:30.233000Z" + } + }, + "outputs": [], + "source": [ + "model_name = 'bert-base-uncased'" + ] + }, + { + "cell_type": "code", + "execution_count": 6, + "metadata": { + "ExecuteTime": { + "end_time": "2022-06-27T14:30:35.901753Z", + "start_time": "2022-06-27T14:30:28.225646Z" + } + }, + "outputs": [ + { + "name": "stderr", + "output_type": "stream", + "text": [ + "Some weights of the model checkpoint at bert-base-uncased were not used when initializing BertModel: ['cls.predictions.bias', 'cls.seq_relationship.weight', 'cls.predictions.transform.dense.bias', 'cls.seq_relationship.bias', 'cls.predictions.transform.LayerNorm.bias', 'cls.predictions.transform.LayerNorm.weight', 'cls.predictions.decoder.weight', 'cls.predictions.transform.dense.weight']\n", + "- This IS expected if you are initializing BertModel from the checkpoint of a model trained on another task or with another architecture (e.g. initializing a BertForSequenceClassification model from a BertForPreTraining model).\n", + "- This IS NOT expected if you are initializing BertModel from the checkpoint of a model that you expect to be exactly identical (initializing a BertForSequenceClassification model from a BertForSequenceClassification model).\n", + "Some weights of the model checkpoint at bert-base-uncased were not used when initializing BertForSequenceClassification: ['cls.predictions.bias', 'cls.seq_relationship.weight', 'cls.predictions.transform.dense.bias', 'cls.seq_relationship.bias', 'cls.predictions.transform.LayerNorm.bias', 'cls.predictions.transform.LayerNorm.weight', 'cls.predictions.decoder.weight', 'cls.predictions.transform.dense.weight']\n", + "- This IS expected if you are initializing BertForSequenceClassification from the checkpoint of a model trained on another task or with another architecture (e.g. initializing a BertForSequenceClassification model from a BertForPreTraining model).\n", + "- This IS NOT expected if you are initializing BertForSequenceClassification from the checkpoint of a model that you expect to be exactly identical (initializing a BertForSequenceClassification model from a BertForSequenceClassification model).\n", + "Some weights of BertForSequenceClassification were not initialized from the model checkpoint at bert-base-uncased and are newly initialized: ['classifier.bias', 'classifier.weight']\n", + "You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.\n" + ] + } + ], + "source": [ + "model = BertModel.from_pretrained(model_name)\n", + "cls_model = BertForSequenceClassification.from_pretrained(model_name)" + ] + }, + { + "cell_type": "code", + "execution_count": 4, + "metadata": { + "ExecuteTime": { + "end_time": "2022-06-27T14:28:43.695685Z", + "start_time": "2022-06-27T14:28:43.683251Z" + } + }, + "outputs": [ + { + "data": { + "text/plain": [ + "BertModel(\n", + " (embeddings): BertEmbeddings(\n", + " (word_embeddings): Embedding(30522, 768, padding_idx=0)\n", + " (position_embeddings): Embedding(512, 768)\n", + " (token_type_embeddings): Embedding(2, 768)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " (encoder): BertEncoder(\n", + " (layer): ModuleList(\n", + " (0): BertLayer(\n", + " (attention): BertAttention(\n", + " (self): BertSelfAttention(\n", + " (query): Linear(in_features=768, out_features=768, bias=True)\n", + " (key): Linear(in_features=768, out_features=768, bias=True)\n", + " (value): Linear(in_features=768, out_features=768, bias=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " (output): BertSelfOutput(\n", + " (dense): Linear(in_features=768, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (intermediate): BertIntermediate(\n", + " (dense): Linear(in_features=768, out_features=3072, bias=True)\n", + " )\n", + " (output): BertOutput(\n", + " (dense): Linear(in_features=3072, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (1): BertLayer(\n", + " (attention): BertAttention(\n", + " (self): BertSelfAttention(\n", + " (query): Linear(in_features=768, out_features=768, bias=True)\n", + " (key): Linear(in_features=768, out_features=768, bias=True)\n", + " (value): Linear(in_features=768, out_features=768, bias=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " (output): BertSelfOutput(\n", + " (dense): Linear(in_features=768, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (intermediate): BertIntermediate(\n", + " (dense): Linear(in_features=768, out_features=3072, bias=True)\n", + " )\n", + " (output): BertOutput(\n", + " (dense): Linear(in_features=3072, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (2): BertLayer(\n", + " (attention): BertAttention(\n", + " (self): BertSelfAttention(\n", + " (query): Linear(in_features=768, out_features=768, bias=True)\n", + " (key): Linear(in_features=768, out_features=768, bias=True)\n", + " (value): Linear(in_features=768, out_features=768, bias=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " (output): BertSelfOutput(\n", + " (dense): Linear(in_features=768, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (intermediate): BertIntermediate(\n", + " (dense): Linear(in_features=768, out_features=3072, bias=True)\n", + " )\n", + " (output): BertOutput(\n", + " (dense): Linear(in_features=3072, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (3): BertLayer(\n", + " (attention): BertAttention(\n", + " (self): BertSelfAttention(\n", + " (query): Linear(in_features=768, out_features=768, bias=True)\n", + " (key): Linear(in_features=768, out_features=768, bias=True)\n", + " (value): Linear(in_features=768, out_features=768, bias=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " (output): BertSelfOutput(\n", + " (dense): Linear(in_features=768, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (intermediate): BertIntermediate(\n", + " (dense): Linear(in_features=768, out_features=3072, bias=True)\n", + " )\n", + " (output): BertOutput(\n", + " (dense): Linear(in_features=3072, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (4): BertLayer(\n", + " (attention): BertAttention(\n", + " (self): BertSelfAttention(\n", + " (query): Linear(in_features=768, out_features=768, bias=True)\n", + " (key): Linear(in_features=768, out_features=768, bias=True)\n", + " (value): Linear(in_features=768, out_features=768, bias=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " (output): BertSelfOutput(\n", + " (dense): Linear(in_features=768, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (intermediate): BertIntermediate(\n", + " (dense): Linear(in_features=768, out_features=3072, bias=True)\n", + " )\n", + " (output): BertOutput(\n", + " (dense): Linear(in_features=3072, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (5): BertLayer(\n", + " (attention): BertAttention(\n", + " (self): BertSelfAttention(\n", + " (query): Linear(in_features=768, out_features=768, bias=True)\n", + " (key): Linear(in_features=768, out_features=768, bias=True)\n", + " (value): Linear(in_features=768, out_features=768, bias=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " (output): BertSelfOutput(\n", + " (dense): Linear(in_features=768, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (intermediate): BertIntermediate(\n", + " (dense): Linear(in_features=768, out_features=3072, bias=True)\n", + " )\n", + " (output): BertOutput(\n", + " (dense): Linear(in_features=3072, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (6): BertLayer(\n", + " (attention): BertAttention(\n", + " (self): BertSelfAttention(\n", + " (query): Linear(in_features=768, out_features=768, bias=True)\n", + " (key): Linear(in_features=768, out_features=768, bias=True)\n", + " (value): Linear(in_features=768, out_features=768, bias=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " (output): BertSelfOutput(\n", + " (dense): Linear(in_features=768, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (intermediate): BertIntermediate(\n", + " (dense): Linear(in_features=768, out_features=3072, bias=True)\n", + " )\n", + " (output): BertOutput(\n", + " (dense): Linear(in_features=3072, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (7): BertLayer(\n", + " (attention): BertAttention(\n", + " (self): BertSelfAttention(\n", + " (query): Linear(in_features=768, out_features=768, bias=True)\n", + " (key): Linear(in_features=768, out_features=768, bias=True)\n", + " (value): Linear(in_features=768, out_features=768, bias=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " (output): BertSelfOutput(\n", + " (dense): Linear(in_features=768, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (intermediate): BertIntermediate(\n", + " (dense): Linear(in_features=768, out_features=3072, bias=True)\n", + " )\n", + " (output): BertOutput(\n", + " (dense): Linear(in_features=3072, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (8): BertLayer(\n", + " (attention): BertAttention(\n", + " (self): BertSelfAttention(\n", + " (query): Linear(in_features=768, out_features=768, bias=True)\n", + " (key): Linear(in_features=768, out_features=768, bias=True)\n", + " (value): Linear(in_features=768, out_features=768, bias=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " (output): BertSelfOutput(\n", + " (dense): Linear(in_features=768, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (intermediate): BertIntermediate(\n", + " (dense): Linear(in_features=768, out_features=3072, bias=True)\n", + " )\n", + " (output): BertOutput(\n", + " (dense): Linear(in_features=3072, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (9): BertLayer(\n", + " (attention): BertAttention(\n", + " (self): BertSelfAttention(\n", + " (query): Linear(in_features=768, out_features=768, bias=True)\n", + " (key): Linear(in_features=768, out_features=768, bias=True)\n", + " (value): Linear(in_features=768, out_features=768, bias=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " (output): BertSelfOutput(\n", + " (dense): Linear(in_features=768, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (intermediate): BertIntermediate(\n", + " (dense): Linear(in_features=768, out_features=3072, bias=True)\n", + " )\n", + " (output): BertOutput(\n", + " (dense): Linear(in_features=3072, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (10): BertLayer(\n", + " (attention): BertAttention(\n", + " (self): BertSelfAttention(\n", + " (query): Linear(in_features=768, out_features=768, bias=True)\n", + " (key): Linear(in_features=768, out_features=768, bias=True)\n", + " (value): Linear(in_features=768, out_features=768, bias=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " (output): BertSelfOutput(\n", + " (dense): Linear(in_features=768, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (intermediate): BertIntermediate(\n", + " (dense): Linear(in_features=768, out_features=3072, bias=True)\n", + " )\n", + " (output): BertOutput(\n", + " (dense): Linear(in_features=3072, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (11): BertLayer(\n", + " (attention): BertAttention(\n", + " (self): BertSelfAttention(\n", + " (query): Linear(in_features=768, out_features=768, bias=True)\n", + " (key): Linear(in_features=768, out_features=768, bias=True)\n", + " (value): Linear(in_features=768, out_features=768, bias=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " (output): BertSelfOutput(\n", + " (dense): Linear(in_features=768, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (intermediate): BertIntermediate(\n", + " (dense): Linear(in_features=768, out_features=3072, bias=True)\n", + " )\n", + " (output): BertOutput(\n", + " (dense): Linear(in_features=3072, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " )\n", + " )\n", + " (pooler): BertPooler(\n", + " (dense): Linear(in_features=768, out_features=768, bias=True)\n", + " (activation): Tanh()\n", + " )\n", + ")" + ] + }, + "execution_count": 4, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "model" + ] + }, + { + "cell_type": "code", + "execution_count": 7, + "metadata": { + "ExecuteTime": { + "end_time": "2022-06-27T14:30:47.519574Z", + "start_time": "2022-06-27T14:30:47.513082Z" + } + }, + "outputs": [ + { + "data": { + "text/plain": [ + "BertForSequenceClassification(\n", + " (bert): BertModel(\n", + " (embeddings): BertEmbeddings(\n", + " (word_embeddings): Embedding(30522, 768, padding_idx=0)\n", + " (position_embeddings): Embedding(512, 768)\n", + " (token_type_embeddings): Embedding(2, 768)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " (encoder): BertEncoder(\n", + " (layer): ModuleList(\n", + " (0): BertLayer(\n", + " (attention): BertAttention(\n", + " (self): BertSelfAttention(\n", + " (query): Linear(in_features=768, out_features=768, bias=True)\n", + " (key): Linear(in_features=768, out_features=768, bias=True)\n", + " (value): Linear(in_features=768, out_features=768, bias=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " (output): BertSelfOutput(\n", + " (dense): Linear(in_features=768, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (intermediate): BertIntermediate(\n", + " (dense): Linear(in_features=768, out_features=3072, bias=True)\n", + " )\n", + " (output): BertOutput(\n", + " (dense): Linear(in_features=3072, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (1): BertLayer(\n", + " (attention): BertAttention(\n", + " (self): BertSelfAttention(\n", + " (query): Linear(in_features=768, out_features=768, bias=True)\n", + " (key): Linear(in_features=768, out_features=768, bias=True)\n", + " (value): Linear(in_features=768, out_features=768, bias=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " (output): BertSelfOutput(\n", + " (dense): Linear(in_features=768, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (intermediate): BertIntermediate(\n", + " (dense): Linear(in_features=768, out_features=3072, bias=True)\n", + " )\n", + " (output): BertOutput(\n", + " (dense): Linear(in_features=3072, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (2): BertLayer(\n", + " (attention): BertAttention(\n", + " (self): BertSelfAttention(\n", + " (query): Linear(in_features=768, out_features=768, bias=True)\n", + " (key): Linear(in_features=768, out_features=768, bias=True)\n", + " (value): Linear(in_features=768, out_features=768, bias=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " (output): BertSelfOutput(\n", + " (dense): Linear(in_features=768, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (intermediate): BertIntermediate(\n", + " (dense): Linear(in_features=768, out_features=3072, bias=True)\n", + " )\n", + " (output): BertOutput(\n", + " (dense): Linear(in_features=3072, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (3): BertLayer(\n", + " (attention): BertAttention(\n", + " (self): BertSelfAttention(\n", + " (query): Linear(in_features=768, out_features=768, bias=True)\n", + " (key): Linear(in_features=768, out_features=768, bias=True)\n", + " (value): Linear(in_features=768, out_features=768, bias=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " (output): BertSelfOutput(\n", + " (dense): Linear(in_features=768, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (intermediate): BertIntermediate(\n", + " (dense): Linear(in_features=768, out_features=3072, bias=True)\n", + " )\n", + " (output): BertOutput(\n", + " (dense): Linear(in_features=3072, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (4): BertLayer(\n", + " (attention): BertAttention(\n", + " (self): BertSelfAttention(\n", + " (query): Linear(in_features=768, out_features=768, bias=True)\n", + " (key): Linear(in_features=768, out_features=768, bias=True)\n", + " (value): Linear(in_features=768, out_features=768, bias=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " (output): BertSelfOutput(\n", + " (dense): Linear(in_features=768, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (intermediate): BertIntermediate(\n", + " (dense): Linear(in_features=768, out_features=3072, bias=True)\n", + " )\n", + " (output): BertOutput(\n", + " (dense): Linear(in_features=3072, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (5): BertLayer(\n", + " (attention): BertAttention(\n", + " (self): BertSelfAttention(\n", + " (query): Linear(in_features=768, out_features=768, bias=True)\n", + " (key): Linear(in_features=768, out_features=768, bias=True)\n", + " (value): Linear(in_features=768, out_features=768, bias=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " (output): BertSelfOutput(\n", + " (dense): Linear(in_features=768, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (intermediate): BertIntermediate(\n", + " (dense): Linear(in_features=768, out_features=3072, bias=True)\n", + " )\n", + " (output): BertOutput(\n", + " (dense): Linear(in_features=3072, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (6): BertLayer(\n", + " (attention): BertAttention(\n", + " (self): BertSelfAttention(\n", + " (query): Linear(in_features=768, out_features=768, bias=True)\n", + " (key): Linear(in_features=768, out_features=768, bias=True)\n", + " (value): Linear(in_features=768, out_features=768, bias=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " (output): BertSelfOutput(\n", + " (dense): Linear(in_features=768, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (intermediate): BertIntermediate(\n", + " (dense): Linear(in_features=768, out_features=3072, bias=True)\n", + " )\n", + " (output): BertOutput(\n", + " (dense): Linear(in_features=3072, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (7): BertLayer(\n", + " (attention): BertAttention(\n", + " (self): BertSelfAttention(\n", + " (query): Linear(in_features=768, out_features=768, bias=True)\n", + " (key): Linear(in_features=768, out_features=768, bias=True)\n", + " (value): Linear(in_features=768, out_features=768, bias=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " (output): BertSelfOutput(\n", + " (dense): Linear(in_features=768, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (intermediate): BertIntermediate(\n", + " (dense): Linear(in_features=768, out_features=3072, bias=True)\n", + " )\n", + " (output): BertOutput(\n", + " (dense): Linear(in_features=3072, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (8): BertLayer(\n", + " (attention): BertAttention(\n", + " (self): BertSelfAttention(\n", + " (query): Linear(in_features=768, out_features=768, bias=True)\n", + " (key): Linear(in_features=768, out_features=768, bias=True)\n", + " (value): Linear(in_features=768, out_features=768, bias=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " (output): BertSelfOutput(\n", + " (dense): Linear(in_features=768, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (intermediate): BertIntermediate(\n", + " (dense): Linear(in_features=768, out_features=3072, bias=True)\n", + " )\n", + " (output): BertOutput(\n", + " (dense): Linear(in_features=3072, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (9): BertLayer(\n", + " (attention): BertAttention(\n", + " (self): BertSelfAttention(\n", + " (query): Linear(in_features=768, out_features=768, bias=True)\n", + " (key): Linear(in_features=768, out_features=768, bias=True)\n", + " (value): Linear(in_features=768, out_features=768, bias=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " (output): BertSelfOutput(\n", + " (dense): Linear(in_features=768, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (intermediate): BertIntermediate(\n", + " (dense): Linear(in_features=768, out_features=3072, bias=True)\n", + " )\n", + " (output): BertOutput(\n", + " (dense): Linear(in_features=3072, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (10): BertLayer(\n", + " (attention): BertAttention(\n", + " (self): BertSelfAttention(\n", + " (query): Linear(in_features=768, out_features=768, bias=True)\n", + " (key): Linear(in_features=768, out_features=768, bias=True)\n", + " (value): Linear(in_features=768, out_features=768, bias=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " (output): BertSelfOutput(\n", + " (dense): Linear(in_features=768, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (intermediate): BertIntermediate(\n", + " (dense): Linear(in_features=768, out_features=3072, bias=True)\n", + " )\n", + " (output): BertOutput(\n", + " (dense): Linear(in_features=3072, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (11): BertLayer(\n", + " (attention): BertAttention(\n", + " (self): BertSelfAttention(\n", + " (query): Linear(in_features=768, out_features=768, bias=True)\n", + " (key): Linear(in_features=768, out_features=768, bias=True)\n", + " (value): Linear(in_features=768, out_features=768, bias=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " (output): BertSelfOutput(\n", + " (dense): Linear(in_features=768, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " (intermediate): BertIntermediate(\n", + " (dense): Linear(in_features=768, out_features=3072, bias=True)\n", + " )\n", + " (output): BertOutput(\n", + " (dense): Linear(in_features=3072, out_features=768, bias=True)\n", + " (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " )\n", + " )\n", + " )\n", + " )\n", + " (pooler): BertPooler(\n", + " (dense): Linear(in_features=768, out_features=768, bias=True)\n", + " (activation): Tanh()\n", + " )\n", + " )\n", + " (dropout): Dropout(p=0.1, inplace=False)\n", + " (classifier): Linear(in_features=768, out_features=2, bias=True)\n", + ")" + ] + }, + "execution_count": 7, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "cls_model" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "### summary\n", + "- bert: encoder of transformer\n", + " - transformer: encoder-decoder(seq2seq)\n", + "- bert\n", + " - embeddings\n", + " - word(token) embedding\n", + " - position embedding\n", + " - token type embedding\n", + " - encoder(12 layer)\n", + " - self attention (kqv)\n", + " - feed forward\n", + " - pooler" + ] + }, + { + "cell_type": "code", + "execution_count": 22, + "metadata": { + "ExecuteTime": { + "end_time": "2022-06-27T14:44:50.691120Z", + "start_time": "2022-06-27T14:44:50.685668Z" + } + }, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "embeddings.word_embeddings.weight -> torch.Size([30522, 768]) -> 23440896\n", + "embeddings.position_embeddings.weight -> torch.Size([512, 768]) -> 393216\n", + "embeddings.token_type_embeddings.weight -> torch.Size([2, 768]) -> 1536\n", + "embeddings.LayerNorm.weight -> torch.Size([768]) -> 768\n", + "embeddings.LayerNorm.bias -> torch.Size([768]) -> 768\n", + "encoder.layer.0.attention.self.query.weight -> torch.Size([768, 768]) -> 589824\n" + ] + } + ], + "source": [ + "# 参数量统计\n", + "total_params = 0\n", + "total_learnable_params = 0\n", + "total_embedding_params = 0\n", + "total_encoder_params = 0\n", + "total_pooler_params = 0\n", + "for name, param in model.named_parameters():\n", + " print(name, '->', param.shape, '->', param.numel())\n", + " if 'embedding' in name:\n", + " total_embedding_params += param.numel()\n", + " if 'encoder' in name:\n", + " total_encoder_params += param.numel()\n", + " if 'pooler' in name:\n", + " total_pooler_params += param.numel()\n", + " if param.requires_grad:\n", + " total_learnable_params += param.numel()\n", + " total_params += param.numel()" + ] + }, + { + "cell_type": "code", + "execution_count": 15, + "metadata": { + "ExecuteTime": { + "end_time": "2022-06-27T14:40:37.188753Z", + "start_time": "2022-06-27T14:40:37.184701Z" + } + }, + "outputs": [ + { + "data": { + "text/plain": [ + "109482240" + ] + }, + "execution_count": 15, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "total_params" + ] + }, + { + "cell_type": "code", + "execution_count": 16, + "metadata": { + "ExecuteTime": { + "end_time": "2022-06-27T14:40:44.983087Z", + "start_time": "2022-06-27T14:40:44.979805Z" + } + }, + "outputs": [ + { + "data": { + "text/plain": [ + "109482240" + ] + }, + "execution_count": 16, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "total_learnable_params" + ] + }, + { + "cell_type": "code", + "execution_count": 20, + "metadata": { + "ExecuteTime": { + "end_time": "2022-06-27T14:44:19.645024Z", + "start_time": "2022-06-27T14:44:19.641479Z" + } + }, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "0.21772649152958506\n", + "0.776879099295009\n", + "0.005394409175405983\n" + ] + } + ], + "source": [ + "params = [total_embedding_params, total_encoder_params, total_pooler_params]\n", + "for param in params:\n", + " print(param/sum(params))" + ] + } + ], + "metadata": { + "kernelspec": { + "display_name": "Python 3", + "language": "python", + "name": "python3" + }, + "language_info": { + "codemirror_mode": { + "name": "ipython", + "version": 3 + }, + "file_extension": ".py", + "mimetype": "text/x-python", + "name": "python", + "nbconvert_exporter": "python", + "pygments_lexer": "ipython3", + "version": "3.6.8" + } + }, + "nbformat": 4, + "nbformat_minor": 2 +} diff --git a/fine_tune/bert_parameters.py b/fine_tune/bert_parameters.py new file mode 100644 index 0000000..98e2b15 --- /dev/null +++ b/fine_tune/bert_parameters.py @@ -0,0 +1,42 @@ + +from transformers import BertModel, BertForSequenceClassification +from collections import defaultdict +import matplotlib.pyplot as plt + + +model_name = 'bert-base-uncased' + +model = BertModel.from_pretrained(model_name) +cls_model = BertForSequenceClassification.from_pretrained(model_name) + +total_params = 0 +total_learnable_params = 0 +total_embedding_params = 0 +total_encoder_params = 0 +total_pooler_params = 0 + +params_dict = defaultdict(float) + +for name, para in model.named_parameters(): + print(name, para.shape, para.numel()) + if para.requires_grad: + total_learnable_params += para.numel() + total_params += para.numel() + if 'embedding' in name: + params_dict['embedding'] += para.numel() + total_embedding_params += para.numel() + if 'encoder' in name: + layer_index = name.split('.')[2] + params_dict['encoder({})'.format(layer_index)] += para.numel() + total_encoder_params += para.numel() + if 'pooler' in name: + params_dict['pooler'] += para.numel() + total_pooler_params += para.numel() + + +print(total_params) +print(total_learnable_params) +print(params_dict) +print(total_embedding_params) +print(total_encoder_params) +print(total_pooler_params)
\ No newline at end of file diff --git a/web/simu/gravity.html b/web/simu/gravity.html new file mode 100644 index 0000000..8ea6df5 --- /dev/null +++ b/web/simu/gravity.html @@ -0,0 +1,118 @@ +<html lang=en> +<head> +<meta charset=utf-8> + <title>Javascript gravity</title> + </head> + <body onload="init()"> + + +<script> + var canvas, ctx, container; + canvas = document.createElement( 'canvas' ); + ctx = canvas.getContext("2d"); + var ball; + var message = "gravity simulator"; + + + // Velocity x + var vx = 5.0; + // Velocity y - randomly set + var vy; + + var gravity = 0.5; + var bounce = 0.7; + var xFriction = 0.1; + + function init(){ + setupCanvas(); + vy = (Math.random() * -15) + -5; + ball = {x:canvas.width / 2, y:100, radius:20, status: 0, color:"red"}; + + }//end init method + + function draw() { + ctx.clearRect(0,0,canvas.width, canvas.height); + //display some text + ctx.fillStyle = "blue"; + ctx.font = "20px Arial"; + ctx.fillText(message, 20,20); + + + + //draw cirlce + ctx.beginPath(); + ctx.arc(ball.x, ball.y, ball.radius, 0, Math.PI*2, false); + ctx.fillStyle = ball.color; + ctx.fill(); + ctx.closePath(); + + ballMovement(); + + } + + setInterval(draw, 1000/35); + + + + + function ballMovement(){ + ball.x += vx; + ball.y += vy; + vy += gravity; + + //If either wall is hit, change direction on x axis + if (ball.x + ball.radius > canvas.width || ball.x - ball.radius < 0){ + vx *= -1; + } + + // Ball hits the floor + if (ball.y + ball.radius > canvas.height){// || + + // Re-positioning on the base + ball.y = canvas.height - ball.radius; + //bounce the ball + vy *= -bounce; + //do this otherwise, ball never stops bouncing + if(vy<0 && vy>-2.1) + vy=0; + //do this otherwise ball never stops on xaxis + if(Math.abs(vx)<1.1) + vx=0; + + xF(); + + } + + + } + + function xF(){ + if(vx>0) + vx = vx - xFriction; + if(vx<0) + vx = vx + xFriction; + } + + + + + + + function setupCanvas() {//setup canvas + + + container = document.createElement( 'div' ); + container.className = "container"; + + canvas.width = window.innerWidth; + canvas.height = window.innerHeight; + document.body.appendChild( container ); + container.appendChild(canvas); + + ctx.strokeStyle = "#ffffff"; + ctx.lineWidth =2; +} + + </script> + </body> +</html>
\ No newline at end of file |