Spaces:
Running
Running
BeveledCube
commited on
Update main.py
Browse files
main.py
CHANGED
@@ -9,9 +9,8 @@ from transformers import GPT2LMHeadModel, GPT2Tokenizer, AutoModelForCausalLM, A
|
|
9 |
import torch
|
10 |
|
11 |
app = FastAPI()
|
12 |
-
name = "
|
13 |
customGen = False
|
14 |
-
gpt2based = False
|
15 |
|
16 |
# microsoft/DialoGPT-small
|
17 |
# microsoft/DialoGPT-medium
|
@@ -38,53 +37,17 @@ def read_root():
|
|
38 |
def read_root(data: req):
|
39 |
print("Prompt:", data.prompt)
|
40 |
print("Length:", data.length)
|
41 |
-
|
42 |
-
if (name == "microsoft/DialoGPT-small" or name == "microsoft/DialoGPT-medium" or name == "microsoft/DialoGPT-large") and customGen == True:
|
43 |
-
# tokenizer = AutoTokenizer.from_pretrained("microsoft/DialoGPT-small")
|
44 |
-
# model = AutoModelForCausalLM.from_pretrained("microsoft/DialoGPT-small")
|
45 |
-
|
46 |
-
step = 1
|
47 |
-
|
48 |
-
# encode the new user input, add the eos_token and return a tensor in Pytorch
|
49 |
-
new_user_input_ids = tokenizer.encode(data.prompt + tokenizer.eos_token, return_tensors='pt')
|
50 |
-
|
51 |
-
# append the new user input tokens to the chat history
|
52 |
-
bot_input_ids = torch.cat(new_user_input_ids, dim=-1) if step > 0 else new_user_input_ids
|
53 |
|
54 |
-
|
55 |
-
chat_history_ids = model.generate(bot_input_ids, max_length=1000, pad_token_id=tokenizer.eos_token_id)
|
56 |
|
57 |
-
|
58 |
-
|
59 |
-
|
60 |
-
|
61 |
-
|
62 |
-
|
63 |
-
|
64 |
-
|
65 |
-
|
66 |
-
|
67 |
-
|
68 |
-
|
69 |
-
# Generate output using the model
|
70 |
-
output_ids = gpt2model.generate(input_ids, max_length=data.length, num_beams=5, no_repeat_ngram_size=2)
|
71 |
-
generated_text = gpt2tokenizer.decode(output_ids[0], skip_special_tokens=True)
|
72 |
-
|
73 |
-
answer_data = { "answer": generated_text }
|
74 |
-
print("Answer:", generated_text)
|
75 |
-
|
76 |
-
return answer_data
|
77 |
-
else:
|
78 |
-
input_text = data.prompt
|
79 |
-
|
80 |
-
# Tokenize the input text
|
81 |
-
input_ids = tokenizer.encode(input_text, return_tensors="pt")
|
82 |
-
|
83 |
-
# Generate output using the model
|
84 |
-
output_ids = model.generate(input_ids, max_length=data.length, num_beams=5, no_repeat_ngram_size=2)
|
85 |
-
generated_text = tokenizer.decode(output_ids[0], skip_special_tokens=True)
|
86 |
-
|
87 |
-
answer_data = { "answer": generated_text }
|
88 |
-
print("Answer:", generated_text)
|
89 |
-
|
90 |
-
return answer_data
|
|
|
9 |
import torch
|
10 |
|
11 |
app = FastAPI()
|
12 |
+
name = "microsoft/DialoGPT-medium"
|
13 |
customGen = False
|
|
|
14 |
|
15 |
# microsoft/DialoGPT-small
|
16 |
# microsoft/DialoGPT-medium
|
|
|
37 |
def read_root(data: req):
|
38 |
print("Prompt:", data.prompt)
|
39 |
print("Length:", data.length)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
40 |
|
41 |
+
input_text = data.prompt
|
|
|
42 |
|
43 |
+
# Tokenize the input text
|
44 |
+
input_ids = tokenizer.encode(input_text, return_tensors="pt")
|
45 |
+
|
46 |
+
# Generate output using the model
|
47 |
+
output_ids = model.generate(input_ids, max_length=data.length, num_beams=5, no_repeat_ngram_size=2)
|
48 |
+
generated_text = tokenizer.decode(output_ids[0], skip_special_tokens=True)
|
49 |
+
|
50 |
+
answer_data = { "answer": generated_text }
|
51 |
+
print("Answer:", generated_text)
|
52 |
+
|
53 |
+
return answer_data
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|