Spaces:
Paused
Paused
ranamhamoud
commited on
Commit
โข
9905ae2
1
Parent(s):
bda6d90
Update app.py
Browse files
app.py
CHANGED
@@ -56,7 +56,14 @@ def generate(
|
|
56 |
top_k: int = 50,
|
57 |
repetition_penalty: float = 1.2,
|
58 |
) -> Iterator[str]:
|
59 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
60 |
model = modelA
|
61 |
tokenizer = tokenizerA
|
62 |
enc = tokenizer(make_prompt(message), return_tensors="pt", padding=True, truncation=True)
|
@@ -66,13 +73,7 @@ def generate(
|
|
66 |
model = modelB
|
67 |
tokenizer = tokenizerB
|
68 |
input_ids = tokenizer.apply_chat_template(conversation, return_tensors="pt")
|
69 |
-
|
70 |
-
if system_prompt:
|
71 |
-
conversation.append({"role": "system", "content": system_prompt})
|
72 |
-
for user, assistant in chat_history:
|
73 |
-
conversation.extend([{"role": "user", "content": user}, {"role": "assistant", "content": assistant}])
|
74 |
-
conversation.append({"role": "user", "content": message})
|
75 |
-
|
76 |
if input_ids.shape[1] > MAX_INPUT_TOKEN_LENGTH:
|
77 |
input_ids = input_ids[:, -MAX_INPUT_TOKEN_LENGTH:]
|
78 |
gr.Warning(f"Trimmed input from conversation as it was longer than {MAX_INPUT_TOKEN_LENGTH} tokens.")
|
|
|
56 |
top_k: int = 50,
|
57 |
repetition_penalty: float = 1.2,
|
58 |
) -> Iterator[str]:
|
59 |
+
|
60 |
+
conversation = []
|
61 |
+
if system_prompt:
|
62 |
+
conversation.append({"role": "system", "content": system_prompt})
|
63 |
+
for user, assistant in chat_history:
|
64 |
+
conversation.extend([{"role": "user", "content": user}, {"role": "assistant", "content": assistant}])
|
65 |
+
conversation.append({"role": "user", "content": message})
|
66 |
+
if model == "A":
|
67 |
model = modelA
|
68 |
tokenizer = tokenizerA
|
69 |
enc = tokenizer(make_prompt(message), return_tensors="pt", padding=True, truncation=True)
|
|
|
73 |
model = modelB
|
74 |
tokenizer = tokenizerB
|
75 |
input_ids = tokenizer.apply_chat_template(conversation, return_tensors="pt")
|
76 |
+
|
|
|
|
|
|
|
|
|
|
|
|
|
77 |
if input_ids.shape[1] > MAX_INPUT_TOKEN_LENGTH:
|
78 |
input_ids = input_ids[:, -MAX_INPUT_TOKEN_LENGTH:]
|
79 |
gr.Warning(f"Trimmed input from conversation as it was longer than {MAX_INPUT_TOKEN_LENGTH} tokens.")
|