Update app.py

app.py

@@ -1,38 +1,26 @@
-from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline
-import gradio as gr
+from transformers import AutoModelForCausalLM, AutoTokenizer
+from huggingface_hub import HfApi, login
 import os
 
-# Load OpenHermes
-model_id = "BuddAi" if os.path.exists("BuddAi") else "CaptMetal/BuddAi"
-try:
-    tokenizer = AutoTokenizer.from_pretrained(
-        model_id,
-        trust_remote_code=True  # Important for custom tokenizers
-    )
-    model = AutoModelForCausalLM.from_pretrained(
-        model_id,
-        device_map="auto"
-    )
-except Exception as e:
-    raise ValueError(f"Failed to load model: {str(e)}. Please verify all files exist in {model_id}")
+# 1. Login (add HF_TOKEN in Space Settings -> Secrets)
+login(token=os.environ["HF_TOKEN"])
 
-# Set chat template (OpenHermes-specific)
-tokenizer.chat_template = """{% for message in messages %}
-{{'<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>'}}
-{% endfor %}"""
+# 2. Load and save model files
+model_id = "teknium/OpenHermes-2.5-Mistral-7B" 
+tokenizer = AutoTokenizer.from_pretrained(model_id)
+model = AutoModelForCausalLM.from_pretrained(model_id)
 
-chatbot = pipeline(
-    "text-generation",
-    model=model,
-    tokenizer=tokenizer,
-    temperature=0.7,
-    max_new_tokens=200
-)
+# Save files temporarily
+os.makedirs("BuddAi", exist_ok=True)
+model.save_pretrained("BuddAi", safe_serialization=True)
+tokenizer.save_pretrained("BuddAi")
 
-def respond(message, history):
-    messages = [{"role": "user", "content": message}]
-    prompt = tokenizer.apply_chat_template(messages, tokenize=False)
-    outputs = chatbot(prompt)
-    return outputs[0]["generated_text"][len(prompt):].split("<|im_end|>")[0]
+# 3. Upload to your model repo
+api = HfApi()
+api.upload_folder(
+    folder_path="BuddAi",
+    repo_id="CaptMetal/BuddAi",
+    repo_type="model"
+)
 
-gr.ChatInterface(respond).launch()
+print("Upload complete! Visit https://huggingface.co/CaptMetal/BuddAi")