multimodalart HF staff commited on
Commit
8caf605
β€’
1 Parent(s): 21ac753

Move performance metrics below

Browse files
Files changed (1) hide show
  1. app.py +10 -10
app.py CHANGED
@@ -95,15 +95,6 @@ def _launch_demo(args, model, tokenizer, config):
95
  ## Qwen-14B-Chat: A Large Language Model by Alibaba Cloud
96
  **Space created by [@artificialguybr](https://twitter.com/artificialguybr) based on QWEN Code. Thanks HF for GPU!**
97
  **Qwen is currently SOTA in the benchmarks for 14B models.**
98
-
99
- ### Performance Metrics:
100
- - **MMLU Accuracy**:
101
- - 0-shot: 64.6
102
- - 5-shot: 66.5
103
- - **HumanEval Pass@1**: 43.9
104
- - **GSM8K Accuracy**:
105
- - 0-shot: 60.1
106
- - 8-shot: 59.3
107
  """)
108
  chatbot = gr.Chatbot(label='Qwen-Chat', elem_classes="control-height", queue=True)
109
  query = gr.Textbox(lines=2, label='Input')
@@ -113,11 +104,20 @@ def _launch_demo(args, model, tokenizer, config):
113
  empty_btn = gr.Button("🧹 Clear History")
114
  submit_btn = gr.Button("πŸš€ Submit")
115
  regen_btn = gr.Button("πŸ€”οΈ Regenerate")
116
-
117
  submit_btn.click(predict, [query, chatbot, task_history], [chatbot], show_progress=True, queue=True) # Enable queue
118
  submit_btn.click(reset_user_input, [], [query])
119
  empty_btn.click(reset_state, [chatbot, task_history], outputs=[chatbot], show_progress=True)
120
  regen_btn.click(regenerate, [chatbot, task_history], [chatbot], show_progress=True, queue=True) # Enable queue
 
 
 
 
 
 
 
 
 
121
  demo.queue(max_size=20)
122
  demo.launch()
123
 
 
95
  ## Qwen-14B-Chat: A Large Language Model by Alibaba Cloud
96
  **Space created by [@artificialguybr](https://twitter.com/artificialguybr) based on QWEN Code. Thanks HF for GPU!**
97
  **Qwen is currently SOTA in the benchmarks for 14B models.**
 
 
 
 
 
 
 
 
 
98
  """)
99
  chatbot = gr.Chatbot(label='Qwen-Chat', elem_classes="control-height", queue=True)
100
  query = gr.Textbox(lines=2, label='Input')
 
104
  empty_btn = gr.Button("🧹 Clear History")
105
  submit_btn = gr.Button("πŸš€ Submit")
106
  regen_btn = gr.Button("πŸ€”οΈ Regenerate")
107
+
108
  submit_btn.click(predict, [query, chatbot, task_history], [chatbot], show_progress=True, queue=True) # Enable queue
109
  submit_btn.click(reset_user_input, [], [query])
110
  empty_btn.click(reset_state, [chatbot, task_history], outputs=[chatbot], show_progress=True)
111
  regen_btn.click(regenerate, [chatbot, task_history], [chatbot], show_progress=True, queue=True) # Enable queue
112
+ gr.Markdown("""### Performance Metrics:
113
+ - **MMLU Accuracy**:
114
+ - 0-shot: 64.6
115
+ - 5-shot: 66.5
116
+ - **HumanEval Pass@1**: 43.9
117
+ - **GSM8K Accuracy**:
118
+ - 0-shot: 60.1
119
+ - 8-shot: 59.3
120
+ """)
121
  demo.queue(max_size=20)
122
  demo.launch()
123