Upload 2 files

app.py

@@ -0,0 +1,47 @@
+import gradio as gr
+import torch
+from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig
+from peft import PeftModel
+
+model_name = "HuggingFaceH4/zephyr-7b-beta"
+adapter_path = "zephyr_lora_adapter"
+
+tokenizer = AutoTokenizer.from_pretrained(adapter_path)
+
+bnb_config = BitsAndBytesConfig(
+    load_in_4bit=True,
+    bnb_4bit_use_double_quant=True,
+    bnb_4bit_quant_type="nf4",
+    bnb_4bit_compute_dtype=torch.float16
+)
+
+base_model = AutoModelForCausalLM.from_pretrained(
+    model_name,
+    quantization_config=bnb_config,
+    device_map="auto",
+    trust_remote_code=True
+)
+
+model = PeftModel.from_pretrained(base_model, adapter_path)
+model.eval()
+
+def solve_math(question, max_tokens=512):
+    prompt = f"<|user|>\n{question}\n<|assistant|>\n"
+    inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
+    with torch.no_grad():
+        outputs = model.generate(
+            **inputs,
+            max_new_tokens=max_tokens,
+            do_sample=False,
+            pad_token_id=tokenizer.pad_token_id or tokenizer.eos_token_id
+        )
+    decoded = tokenizer.decode(outputs[0], skip_special_tokens=True)
+    return decoded.split("<|assistant|>")[-1].strip()
+
+demo = gr.Interface(fn=solve_math,
+                    inputs=gr.Textbox(lines=5, label="Enter math problem"),
+                    outputs=gr.Textbox(label="Solution"),
+                    title="Math Solver (Zephyr Fine-Tuned)",
+                    description="This app uses a fine-tuned LLM to solve school-level math problems step by step.")
+
+demo.launch()

requirements.txt

@@ -0,0 +1,5 @@
+transformers
+torch
+gradio
+peft
+bitsandbytes