| import pandas as pd | |
| from datasets import load_dataset | |
| huggingface_dataset_name = "knkarthick/dialogsum" | |
| dataset = load_dataset(huggingface_dataset_name) |
| def build_llama2_prompt(messages): | |
| startPrompt = "<s>[INST] " | |
| endPrompt = " [/INST]" | |
| conversation = [] | |
| for index, message in enumerate(messages): | |
| if message["role"] == "system" and index == 0: | |
| conversation.append(f"<<SYS>>\n{message['content']}\n<</SYS>>\n\n") | |
| elif message["role"] == "user": | |
| conversation.append(message["content"].strip()) | |
| else: |
It contains steps for installing CUDA. It further shows how one can install Tensorflow and Pytorch and use GPUs with them.Our motive is to get output as in the image below,
where nvidia-smi, nvcc --version, torch.cuda.is_available(),tf.test.gpu_devic_name() all are working. Let's go then. I will specifically install CUDA 10.1 where both Tensorflow and Pytorch works fine.
| train_df.review = train_df.review.str.lower() | |
| sentences = train_df.review.values | |
| # We need to add special tokens at the beginning and end of each sentence for BERT to work properly | |
| sentences = ["[CLS] " + sentence + " [SEP]" for sentence in sentences] | |
| labels = train_df.sentiment.values | |
| tokenizer = BertTokenizer.from_pretrained('bert-base-uncased') | |
| tokenized_texts = [tokenizer.tokenize(sent) for sent in sentences] | |
| print ("Tokenize the first sentence:") |
| ##combine train and test reviews | |
| lm_df = pd.concat([train_df[['review']],test_df[['review']]]) | |
| lm_df.review = lm_df.review.str.lower() | |
| tqdm.pandas() | |
| ## We need one sentence per line with space between two lines | |
| changed_text=lm_df.review.apply(lambda x:x+"\n"+"\n") | |
| open(os.path.join(directory_path,'data_lm.txt'), "w").write(''.join(changed_text)) | |
| ##We need to create data in Bert Format, below command will do that | |
| !python3 pregenerate_training_data.py --train_corpus data_lm.txt --bert_model bert-base-uncased --do_lower_case --output_dir training/ --epochs_to_generate 2 --max_seq_len 256 | |
| ##We would now use this data to finetune the model |