Update app.py
Browse files
app.py
CHANGED
@@ -7,15 +7,8 @@ import re
|
|
7 |
|
8 |
# Model configurations
|
9 |
MODELS = {
|
10 |
-
"
|
11 |
-
"
|
12 |
-
"Athena-R3 7B": "Spestly/Athena-R3-7B",
|
13 |
-
"Athena-3 3B": "Spestly/Athena-3-3B",
|
14 |
-
"Athena-3 7B": "Spestly/Athena-3-7B",
|
15 |
-
"Athena-3 14B": "Spestly/Athena-3-14B",
|
16 |
-
"Athena-2 1.5B": "Spestly/Athena-2-1.5B",
|
17 |
-
"Athena-1 3B": "Spestly/Athena-1-3B",
|
18 |
-
"Athena-1 7B": "Spestly/Athena-1-7B"
|
19 |
}
|
20 |
|
21 |
@spaces.GPU
|
@@ -38,10 +31,9 @@ def generate_response(model_id, conversation, user_message, max_length=512, temp
|
|
38 |
# Build messages in proper chat format (OpenAI-style messages)
|
39 |
messages = []
|
40 |
system_prompt = (
|
41 |
-
"You are
|
42 |
"You provide clear, accurate, and concise responses to user questions. "
|
43 |
"You are knowledgeable across many domains and always aim to be respectful and helpful. "
|
44 |
-
"You are finetuned by Aayan Mishra"
|
45 |
)
|
46 |
messages.append({"role": "system", "content": system_prompt})
|
47 |
|
@@ -175,8 +167,8 @@ css = """
|
|
175 |
}
|
176 |
"""
|
177 |
|
178 |
-
with gr.Blocks(title="
|
179 |
-
gr.Markdown("# 🚀
|
180 |
gr.Markdown("*Powered by HuggingFace ZeroGPU*")
|
181 |
|
182 |
# State to keep track of the conversation for the model
|
@@ -197,7 +189,7 @@ with gr.Blocks(title="Athena Playground Chat", css=css, theme='NoCrypt/miku') as
|
|
197 |
model_choice = gr.Dropdown(
|
198 |
label="📱 Model",
|
199 |
choices=list(MODELS.keys()),
|
200 |
-
value="
|
201 |
info="Select which Athena model to use"
|
202 |
)
|
203 |
max_length = gr.Slider(
|
@@ -243,11 +235,6 @@ with gr.Blocks(title="Athena Playground Chat", css=css, theme='NoCrypt/miku') as
|
|
243 |
inputs=[user_input]
|
244 |
)
|
245 |
|
246 |
-
gr.Markdown("""
|
247 |
-
### About the Thinking Tags
|
248 |
-
Some Athena models (particularly R3X series) include reasoning in `<think></think>` tags.
|
249 |
-
Click "Show reasoning" to see the model's thought process behind its answers.
|
250 |
-
""")
|
251 |
|
252 |
if __name__ == "__main__":
|
253 |
demo.launch(debug=True) # Enable debug mode for better error reporting
|
|
|
7 |
|
8 |
# Model configurations
|
9 |
MODELS = {
|
10 |
+
"20B": "openai/gpt-oss-20b",
|
11 |
+
"120B": "openai/gpt-oss-20b",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
12 |
}
|
13 |
|
14 |
@spaces.GPU
|
|
|
31 |
# Build messages in proper chat format (OpenAI-style messages)
|
32 |
messages = []
|
33 |
system_prompt = (
|
34 |
+
"You are GPT, a helpful, harmless, and honest AI assistant. "
|
35 |
"You provide clear, accurate, and concise responses to user questions. "
|
36 |
"You are knowledgeable across many domains and always aim to be respectful and helpful. "
|
|
|
37 |
)
|
38 |
messages.append({"role": "system", "content": system_prompt})
|
39 |
|
|
|
167 |
}
|
168 |
"""
|
169 |
|
170 |
+
with gr.Blocks(title="GPT-OSS Playground Chat", css=css, theme='NoCrypt/miku') as demo:
|
171 |
+
gr.Markdown("# 🚀 GPT-OSS Playground Chat")
|
172 |
gr.Markdown("*Powered by HuggingFace ZeroGPU*")
|
173 |
|
174 |
# State to keep track of the conversation for the model
|
|
|
189 |
model_choice = gr.Dropdown(
|
190 |
label="📱 Model",
|
191 |
choices=list(MODELS.keys()),
|
192 |
+
value="20B",
|
193 |
info="Select which Athena model to use"
|
194 |
)
|
195 |
max_length = gr.Slider(
|
|
|
235 |
inputs=[user_input]
|
236 |
)
|
237 |
|
|
|
|
|
|
|
|
|
|
|
238 |
|
239 |
if __name__ == "__main__":
|
240 |
demo.launch(debug=True) # Enable debug mode for better error reporting
|