Spaces:
Runtime error
Runtime error
Update app.py
Browse files
app.py
CHANGED
@@ -27,10 +27,11 @@ def get_model():
|
|
27 |
#model = AutoModelForCausalLM.from_pretrained("facebook/opt-125m")
|
28 |
tokenizer = AutoTokenizer.from_pretrained("BigSalmon/InformalToFormalLincoln63Paraphrase")
|
29 |
model = AutoModelForCausalLM.from_pretrained("BigSalmon/InformalToFormalLincoln68Paraphrase")
|
|
|
30 |
#model = AutoModelForCausalLM.from_pretrained("BigSalmon/InformalToFormalLincoln63Paraphrase")
|
31 |
-
return model, tokenizer
|
32 |
|
33 |
-
model, tokenizer = get_model()
|
34 |
|
35 |
st.text('''For Prompt Templates: https://huggingface.co/BigSalmon/InformalToFormalLincoln63Paraphrase''')
|
36 |
|
@@ -65,10 +66,41 @@ def run_generate(text, bad_words):
|
|
65 |
e = e.replace(text, "")
|
66 |
yo.append(e)
|
67 |
return yo
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
68 |
with st.form(key='my_form'):
|
69 |
text = st.text_area(label='Enter sentence', value=first)
|
70 |
submit_button = st.form_submit_button(label='Submit')
|
71 |
submit_button2 = st.form_submit_button(label='Submit Log Probs')
|
|
|
|
|
|
|
|
|
72 |
if submit_button:
|
73 |
translated_text = run_generate(text, bad_words)
|
74 |
st.write(translated_text if translated_text else "No translation found")
|
@@ -84,4 +116,19 @@ with st.form(key='my_form'):
|
|
84 |
probabilities = torch.nn.functional.softmax(logits)
|
85 |
best_logits, best_indices = logits.topk(logs_outputs)
|
86 |
best_words = [tokenizer.decode([idx.item()]) for idx in best_indices]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
87 |
st.write(best_words)
|
|
|
27 |
#model = AutoModelForCausalLM.from_pretrained("facebook/opt-125m")
|
28 |
tokenizer = AutoTokenizer.from_pretrained("BigSalmon/InformalToFormalLincoln63Paraphrase")
|
29 |
model = AutoModelForCausalLM.from_pretrained("BigSalmon/InformalToFormalLincoln68Paraphrase")
|
30 |
+
model2 = AutoModelForCausalLM.from_pretrained("BigSalmon/InformalToFormalLincoln63Paraphrase")
|
31 |
#model = AutoModelForCausalLM.from_pretrained("BigSalmon/InformalToFormalLincoln63Paraphrase")
|
32 |
+
return model, model2, tokenizer
|
33 |
|
34 |
+
model, model2, tokenizer = get_model()
|
35 |
|
36 |
st.text('''For Prompt Templates: https://huggingface.co/BigSalmon/InformalToFormalLincoln63Paraphrase''')
|
37 |
|
|
|
66 |
e = e.replace(text, "")
|
67 |
yo.append(e)
|
68 |
return yo
|
69 |
+
|
70 |
+
def run_generate2(text, bad_words):
|
71 |
+
yo = []
|
72 |
+
input_ids = tokenizer.encode(text, return_tensors='pt')
|
73 |
+
res = len(tokenizer.encode(text))
|
74 |
+
bad_words = bad_words.split()
|
75 |
+
bad_word_ids = []
|
76 |
+
for bad_word in bad_words:
|
77 |
+
bad_word = " " + bad_word
|
78 |
+
ids = tokenizer(bad_word).input_ids
|
79 |
+
bad_word_ids.append(ids)
|
80 |
+
sample_outputs = model2.generate(
|
81 |
+
input_ids,
|
82 |
+
do_sample=True,
|
83 |
+
max_length= res + lengths,
|
84 |
+
min_length = res + lengths,
|
85 |
+
top_k=50,
|
86 |
+
temperature=temp,
|
87 |
+
num_return_sequences=number_of_outputs,
|
88 |
+
bad_words_ids=bad_word_ids
|
89 |
+
)
|
90 |
+
for i in range(number_of_outputs):
|
91 |
+
e = tokenizer.decode(sample_outputs[i])
|
92 |
+
e = e.replace(text, "")
|
93 |
+
yo.append(e)
|
94 |
+
return yo
|
95 |
+
|
96 |
with st.form(key='my_form'):
|
97 |
text = st.text_area(label='Enter sentence', value=first)
|
98 |
submit_button = st.form_submit_button(label='Submit')
|
99 |
submit_button2 = st.form_submit_button(label='Submit Log Probs')
|
100 |
+
|
101 |
+
submit_button3 = st.form_submit_button(label='Submit Other Model')
|
102 |
+
submit_button4 = st.form_submit_button(label='Submit Log Probs Other Model')
|
103 |
+
|
104 |
if submit_button:
|
105 |
translated_text = run_generate(text, bad_words)
|
106 |
st.write(translated_text if translated_text else "No translation found")
|
|
|
116 |
probabilities = torch.nn.functional.softmax(logits)
|
117 |
best_logits, best_indices = logits.topk(logs_outputs)
|
118 |
best_words = [tokenizer.decode([idx.item()]) for idx in best_indices]
|
119 |
+
st.write(best_words)
|
120 |
+
if submit_button3:
|
121 |
+
translated_text = run_generate2(text, bad_words)
|
122 |
+
st.write(translated_text if translated_text else "No translation found")
|
123 |
+
if submit_button4:
|
124 |
+
text2 = str(text)
|
125 |
+
print(text2)
|
126 |
+
text3 = tokenizer.encode(text2)
|
127 |
+
myinput, past_key_values = torch.tensor([text3]), None
|
128 |
+
myinput = myinput
|
129 |
+
logits, past_key_values = model2(myinput, past_key_values = past_key_values, return_dict=False)
|
130 |
+
logits = logits[0,-1]
|
131 |
+
probabilities = torch.nn.functional.softmax(logits)
|
132 |
+
best_logits, best_indices = logits.topk(logs_outputs)
|
133 |
+
best_words = [tokenizer.decode([idx.item()]) for idx in best_indices]
|
134 |
st.write(best_words)
|