Spaces:
Runtime error
Runtime error
Upload 2 files
Browse files- app.py +24 -0
- requirements.txt +3 -0
app.py
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import gradio as gr
|
| 2 |
+
from transformers import pipeline
|
| 3 |
+
|
| 4 |
+
pipe = pipeline(
|
| 5 |
+
"text-generation",
|
| 6 |
+
model="cmz1024/olmo3-190m-zh-nano",
|
| 7 |
+
model_kwargs={"attn_implementation": "sdpa"},
|
| 8 |
+
) # 替换你的模型
|
| 9 |
+
|
| 10 |
+
|
| 11 |
+
def predict(message):
|
| 12 |
+
output = pipe(
|
| 13 |
+
message,
|
| 14 |
+
max_new_tokens=256,
|
| 15 |
+
do_sample=True, # 开启采样(默认是贪心解码)
|
| 16 |
+
temperature=0.7, # 温度,越高越随机
|
| 17 |
+
top_k=50, # 只从概率最高的 k 个 token 中采样
|
| 18 |
+
top_p=0.9, # 核采样,累积概率截断
|
| 19 |
+
repetition_penalty=1.2, # 重复惩罚
|
| 20 |
+
)
|
| 21 |
+
return output[0]["generated_text"]
|
| 22 |
+
|
| 23 |
+
|
| 24 |
+
gr.Interface(fn=predict, inputs="text", outputs="text").launch()
|
requirements.txt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
transformers
|
| 2 |
+
torch
|
| 3 |
+
accelerate
|