changed to use dclm model

used apple's new dclm 7b model

app.py

@@ -1,11 +1,8 @@
 import gradio as gr
-from huggingface_hub import InferenceClient
-
-"""
-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
-"""
-client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
+from transformers import AutoTokenizer, AutoModelForCausalLM
 
+tokenizer = AutoTokenizer.from_pretrained("apple/DCLM-Baseline-7B-8k")
+model = AutoModelForCausalLM.from_pretrained("apple/DCLM-Baseline-7B-8k")
 
 def respond(
     message,
@@ -25,23 +22,22 @@ def respond(
 
     messages.append({"role": "user", "content": message})
 
-    response = ""
+    prompt = "".join([f"{'[|Human|] ' if msg['role'] == 'user' else '[|AI|] '}{msg['content']}" for msg in messages])
 
-    for message in client.chat_completion(
-        messages,
-        max_tokens=max_tokens,
-        stream=True,
-        temperature=temperature,
-        top_p=top_p,
-    ):
-        token = message.choices[0].delta.content
+    inputs = tokenizer(prompt, return_tensors="pt")
+    gen_kwargs = {
+        "max_new_tokens": max_tokens,
+        "top_p": top_p,
+        "temperature": temperature,
+        "do_sample": True,
+        "repetition_penalty": 1.1
+    }
+    with torch.no_grad():
+        output = model.generate(inputs['input_ids'], **gen_kwargs)
+    response = tokenizer.decode(output[0].tolist(), skip_special_tokens=True)[len(prompt):]
 
-        response += token
-        yield response
+    yield response
 
-"""
-For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
-"""
 demo = gr.ChatInterface(
     respond,
     additional_inputs=[