transfer from [csebuetnlp/mT5_multilingual_XLSum] to [silver/chatglm-6b-int4-slim]

app.py

@@ -1,34 +1,35 @@
 import gradio as gr
 import re
-from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
+from transformers import AutoTokenizer, AutoModel
 
-MODEL_NAME = "csebuetnlp/mT5_multilingual_XLSum"
+MODEL_NAME = "silver/chatglm-6b-int4-slim"
 
-WHITESPACE_HANDLER = lambda k: re.sub('\s+', ' ', re.sub('\n+', ' ', k.strip()))
+tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME, trust_remote_code=True)
+model = AutoModel.from_pretrained(MODEL_NAME, trust_remote_code=True).half().cuda()
 
-tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
-model = AutoModelForSeq2SeqLM.from_pretrained(MODEL_NAME)
-
-def summarize(text):
-    text = WHITESPACE_HANDLER(text)
+def summarize(transcript, sentence_count):
+    history = []
     
-    inputs = tokenizer(text, 
-                max_length=1024, 
-                truncation=True,
-                return_tensors="pt")
+    prompt = f"""
+                视频脚本：{transcript}
+                我希望你作为一名专业的视频内容编辑，帮我用中文总结视频脚本的内容精华。请先用一句简短的话总结视频梗概。然后再请你将视频字幕文本进行总结（字幕中可能有错别字，如果你发现了错别字请改正）。请你以无序列表的方式返回，请注意不要超过{sentence_count}条哦，确保所有的句子都足够精简，清晰完整，祝你好运！
+            """
+
+    response, history = model.chat(tokenizer, prompt, history=history)
 
-    summary_ids = model.generate(inputs["input_ids"])
-    summary = tokenizer.batch_decode(summary_ids, 
-                                    skip_special_tokens=True, 
-                                    clean_up_tokenization_spaces=False)
-    return summary
+    return response
 
 
 demo = gr.Interface(fn = summarize,
                         inputs = [gr.inputs.Textbox(lines=10, 
-                                                    placeholder="Inpuy something...",
-                                                    label='Text here !!')],
-                        outputs = [gr.outputs.Textbox(
+                                                    placeholder="Input something...",
+                                                    label='Text here !!'),
+                                    gr.inputs.Slider(minimum=1,
+                                                     maximum=10,
+                                                     value=5,
+                                                     step=1,
+                                                     label='Sentence Count')],
+                        outputs = [gr.outputs.Textbox(lines=10,
                                                       label="Summary")],
                          
                         title = "🎈 Summarizer 🎈",

requirements.txt

@@ -2,4 +2,8 @@ gradio
 torch
 tqdm
 transformers
-sentencepiece
+sentencepiece
+protobuf==3.20.0 
+transformers==4.26.1 
+icetk 
+cpm_kernels