test use chronoboros instead

app.py

@@ -1,48 +1,38 @@
 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")
-
-
-def respond(
-    message,
-    history: list[tuple[str, str]],
-    system_message,
-    max_tokens,
-    temperature,
-    top_p,
-):
-    messages = [{"role": "system", "content": system_message}]
-
-    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,
+from transformers import AutoTokenizer, AutoModelForCausalLM
+
+# Load the model and tokenizer (you may need to adjust device_map or other settings depending on your hardware)
+tokenizer = AutoTokenizer.from_pretrained("TheBloke/Chronoboros-33B-GPTQ")
+model = AutoModelForCausalLM.from_pretrained("TheBloke/Chronoboros-33B-GPTQ", device_map="auto")
+
+def respond(message, history: list[tuple[str, str]], system_message, max_tokens, temperature, top_p):
+    # Build the prompt using conversation history
+    prompt = f"{system_message}\n"
+    for user_text, assistant_text in history:
+        if user_text:
+            prompt += f"User: {user_text}\n"
+        if assistant_text:
+            prompt += f"Assistant: {assistant_text}\n"
+    prompt += f"User: {message}\nAssistant: "
+
+    # Tokenize the prompt and generate a response
+    input_ids = tokenizer.encode(prompt, return_tensors="pt").to(model.device)
+    output_ids = model.generate(
+        input_ids,
+        max_new_tokens=max_tokens,
         temperature=temperature,
         top_p=top_p,
-    ):
-        token = message.choices[0].delta.content
+        do_sample=True,  # enable sampling for varied responses
+    )
+    # Get only the newly generated tokens (after the prompt)
+    new_tokens = output_ids[0][input_ids.shape[1]:]
 
-        response += token
-        yield response
+    # Simulate streaming by yielding partial responses token by token
+    for i in range(new_tokens.shape[0]):
+        current_response = tokenizer.decode(new_tokens[: i + 1], skip_special_tokens=True)
+        yield current_response
 
-
-"""
-For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
-"""
+# Configure the ChatInterface with additional inputs
 demo = gr.ChatInterface(
     respond,
     additional_inputs=[
@@ -59,6 +49,5 @@ demo = gr.ChatInterface(
     ],
 )
 
-
 if __name__ == "__main__":
     demo.launch()