Views
No views yet
https://www.kaggle.com/Cornell-University/movie-dialog-corpus
https://www.kaggle.com/jef1056/discord-data
1from transformers import AutoTokenizer, AutoModelWithLMHead
2
3tokenizer = AutoTokenizer.from_pretrained("deepparag/DumBot")
4model = AutoModelWithLMHead.from_pretrained("deepparag/DumBot")
5# Let's chat for 4 lines
6for step in range(4):
7 # encode the new user input, add the eos_token and return a tensor in Pytorch
8 new_user_input_ids = tokenizer.encode(input(">> User:") + tokenizer.eos_token, return_tensors='pt')
9 # print(new_user_input_ids)
10 # append the new user input tokens to the chat history
11 bot_input_ids = torch.cat([chat_history_ids, new_user_input_ids], dim=-1) if step > 0 else new_user_input_ids
12 # generated a response while limiting the total chat history to 1000 tokens,
13 chat_history_ids = model.generate(
14 bot_input_ids, max_length=200,
15 pad_token_id=tokenizer.eos_token_id,
16 no_repeat_ngram_size=4,
17 do_sample=True,
18 top_k=100,
19 top_p=0.7,
20 temperature=0.8
21 )
22
23 # pretty print last ouput tokens from bot
24 print("DumBot: {}".format(tokenizer.decode(chat_history_ids[:, bot_input_ids.shape[-1]:][0], skip_special_tokens=True)))