kingabzpro commited on
Commit
f754e38
1 Parent(s): 3223321

Create README.md

Browse files
Files changed (1) hide show
  1. README.md +43 -0
README.md ADDED
@@ -0,0 +1,43 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language: English
3
+ datasets:
4
+ - Andrada Olteanu Rickmorty-Scripts
5
+ tags:
6
+ - Transformers
7
+ - gpt2
8
+ - Chatbot
9
+ - Rick&Morty
10
+ license: apache-2.0
11
+ metrics:
12
+ - Perplexity
13
+ ---
14
+ ## Evaluation
15
+
16
+ ```python
17
+ tokenizer = AutoTokenizer.from_pretrained('kingabzpro/DialoGPT-small-Rick-Bot')
18
+ model = AutoModelWithLMHead.from_pretrained(kingabzpro/DialoGPT-small-Rick-Bot')
19
+ # Let's chat for 4 lines
20
+ for step in range(4):
21
+ # encode the new user input, add the eos_token and return a tensor in Pytorch
22
+ new_user_input_ids = tokenizer.encode(input(">> User:") + tokenizer.eos_token, return_tensors='pt')
23
+ # print(new_user_input_ids)
24
+
25
+ # append the new user input tokens to the chat history
26
+ bot_input_ids = torch.cat([chat_history_ids, new_user_input_ids], dim=-1) if step > 0 else new_user_input_ids
27
+
28
+ # generated a response while limiting the total chat history to 1000 tokens,
29
+ chat_history_ids = model.generate(
30
+ bot_input_ids, max_length=200,
31
+ pad_token_id=tokenizer.eos_token_id,
32
+ no_repeat_ngram_size=3,
33
+ do_sample=True,
34
+ top_k=100,
35
+ top_p=0.7,
36
+ temperature=0.8
37
+ )
38
+
39
+ # pretty print last ouput tokens from bot
40
+ print("RickBot: {}".format(tokenizer.decode(chat_history_ids[:, bot_input_ids.shape[-1]:][0], skip_special_tokens=True)))
41
+ ```
42
+
43
+ **Result**: perplexity = 8.53