Spaces:
Runtime error
Runtime error
Commit
·
bacca9b
1
Parent(s):
dcef39a
Update app.py
Browse files
app.py
CHANGED
@@ -305,24 +305,22 @@ async def text_to_speech_edge(text, language_code):
|
|
305 |
|
306 |
|
307 |
with gr.Blocks(title="ChatGLM2-6B-int4", theme=gr.themes.Soft(text_size="sm")) as demo:
|
308 |
-
|
|
|
|
|
|
|
|
|
|
|
309 |
gr.HTML(
|
310 |
-
"""<center><a href="https://huggingface.co/spaces/
|
311 |
)
|
312 |
|
|
|
313 |
with gr.Accordion("📒 相关信息", open=False):
|
314 |
-
_ = f"""
|
315 |
-
## {model_name}
|
316 |
-
Try to refresh the browser and try again when occasionally an error occurs.
|
317 |
-
With a GPU, a query takes from a few seconds to a few tens of seconds, dependent on the number of words/characters
|
318 |
-
the question and responses contain. The quality of the responses varies quite a bit it seems. Even the same
|
319 |
-
question with the same parameters, asked at different times, can result in quite different responses.
|
320 |
* Low temperature: responses will be more deterministic and focused; High temperature: responses more creative.
|
321 |
* Suggested temperatures -- translation: up to 0.3; chatting: > 0.4
|
322 |
* Top P controls dynamic vocabulary selection based on context.
|
323 |
-
For a table of example values for different scenarios, refer to [this](https://community.openai.com/t/cheat-sheet-mastering-temperature-and-top-p-in-chatgpt-api-a-few-tips-and-tricks-on-controlling-the-creativity-deterministic-output-of-prompt-responses/172683)
|
324 |
-
If the instance is not on a GPU (T4), it will be very slow. You can try to run the colab notebook [chatglm2-6b-4bit colab notebook](https://colab.research.google.com/drive/1WkF7kOjVCcBBatDHjaGkuJHnPdMWNtbW?usp=sharing) for a spin.
|
325 |
-
The T4 GPU is sponsored by a community GPU grant from Huggingface. Thanks a lot!
|
326 |
"""
|
327 |
gr.Markdown(dedent(_))
|
328 |
chatbot = gr.Chatbot()
|
|
|
305 |
|
306 |
|
307 |
with gr.Blocks(title="ChatGLM2-6B-int4", theme=gr.themes.Soft(text_size="sm")) as demo:
|
308 |
+
gr.HTML("<center>"
|
309 |
+
"<h1>🥳💕🎶 - ChatGLM2 + 声音克隆:和你喜欢的角色畅所欲言吧!</h1>"
|
310 |
+
"</center>")
|
311 |
+
gr.Markdown("### <center>💡 - 第二代ChatGLm大语言模型 + FreeVC变声,为您打造独一无二的沉浸式对话体验</center>")
|
312 |
+
gr.Markdown("### <center>🌊 - 更多精彩应用,敬请关注[滔滔AI](http://www.talktalkai.com);滔滔AI,为爱滔滔!💕</center>")
|
313 |
+
|
314 |
gr.HTML(
|
315 |
+
"""<center><a href="https://huggingface.co/spaces/kevinwang676/FreeVC?duplicate=true"><img src="https://bit.ly/3gLdBN6" alt="Duplicate Space"></a>点击此按钮即可复制该程序;切换到GPU环境后,就可以更快运行GLM2</center>"""
|
316 |
)
|
317 |
|
318 |
+
|
319 |
with gr.Accordion("📒 相关信息", open=False):
|
320 |
+
_ = f""" ChatGLM2 可选参数的信息:
|
|
|
|
|
|
|
|
|
|
|
321 |
* Low temperature: responses will be more deterministic and focused; High temperature: responses more creative.
|
322 |
* Suggested temperatures -- translation: up to 0.3; chatting: > 0.4
|
323 |
* Top P controls dynamic vocabulary selection based on context.
|
|
|
|
|
|
|
324 |
"""
|
325 |
gr.Markdown(dedent(_))
|
326 |
chatbot = gr.Chatbot()
|