Update app.py

app.py

@@ -1,14 +1,31 @@
 import gradio as gr
-from huggingface_hub import InferenceClient
-import os
+from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline
+from peft import PeftModel
+import torch
 
 """
 For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
 """
 
-api_key = os.environ.get("ACESS_TOKEN")
-print(api_key)
-client = InferenceClient("asdc/Mistral-7B-multilingual-temporal-expression-normalization", token=api_key)
+# Set your model and adapter paths
+BASE_MODEL = "mistralai/Mistral-7B-v0.1"
+PEFT_ADAPTER = "asdc/Mistral-7B-multilingual-temporal-expression-normalization"
+
+tokenizer = AutoTokenizer.from_pretrained(BASE_MODEL)
+base_model = AutoModelForCausalLM.from_pretrained(
+    BASE_MODEL,
+    torch_dtype=torch.float16,
+    device_map="auto"
+)
+
+model = PeftModel.from_pretrained(base_model, PEFT_ADAPTER)
+
+pipe = pipeline(
+    "text-generation",
+    model=model,
+    tokenizer=tokenizer,
+    device_map="auto"
+)
 
 
 def respond(
@@ -19,29 +36,24 @@ def respond(
     temperature,
     top_p,
 ):
-    messages = [{"role": "system", "content": system_message}]
+    prompt = system_message + "\n"
+    for user, assistant in history:
+        if user:
+            prompt += f"User: {user}\n"
+        if assistant:
+            prompt += f"Assistant: {assistant}\n"
+    prompt += f"User: {message}\nAssistant:"
 
-    for val in history:
-        if val[0]:
-            messages.append({"role": "user", "content": val[0]})
-        if val[1]:
-            messages.append({"role": "assistant", "content": val[1]})
-
-    messages.append({"role": "user", "content": message})
-
-    response = ""
-
-    for message in client.chat_completion(
-        messages,
-        max_tokens=max_tokens,
-        stream=True,
+    outputs = pipe(
+        prompt,
+        max_new_tokens=max_tokens,
         temperature=temperature,
         top_p=top_p,
-    ):
-        token = message.choices[0].delta.content
-
-        response += token
-        yield response
+        do_sample=True,
+        pad_token_id=tokenizer.eos_token_id,
+    )
+    response = outputs[0]["generated_text"][len(prompt):]
+    yield response
 
 
 """