Update app.py

app.py

@@ -1,21 +1,22 @@
-from llama_cpp import Llama
+from transformers import pipeline
 import os
 
-# Configure cache
-os.environ['GGUF_CACHE'] = '/tmp/gguf_cache'
-os.makedirs('/tmp/gguf_cache', exist_ok=True)
+# Cache setup
+os.environ['HF_HOME'] = '/tmp/cache'
 
-# Load GGUF model
-llm = Llama(
-    model_path="ninja-v1-nsfw-rp.gguf",
-    n_ctx=2048,  # Context window
-    n_threads=4  # CPU threads
+model = pipeline(
+    "text-generation",
+    model="gpt2",
+    device=-1  # Force CPU
 )
 
-def generate_text(prompt, max_tokens=560):
-    output = llm.create_chat_completion(
-        messages=[{"role": "user", "content": prompt}],
-        max_tokens=max_tokens,
-        temperature=0.7
+def generate_text(prompt, max_new_tokens=560, max_context=1080):
+    """Generate text with precise token control"""
+    output = model(
+        prompt,
+        max_new_tokens=max_new_tokens,  # Response tokens (560)
+        max_length=min(max_context, 1024),  # GPT-2's max context is 1024
+        truncation=True,
+        pad_token_id=50256  # Explicitly set to avoid warnings
     )
-    return output['choices'][0]['message']['content']
+    return output[0]["generated_text"]