Update app.py

app.py

@@ -1,60 +1,46 @@
 import os, textwrap, torch, gradio as gr
 from transformers import (
-    pipeline,
     AutoTokenizer,
     AutoModelForCausalLM,
     BitsAndBytesConfig,
+    pipeline,
 )
 
-# ✅ 1.  Use the *Instruct* checkpoint
-MODEL_ID = os.getenv(
-    "MODEL_ID",
-    "mistralai/Mixtral-8x7B-Instruct-v0.3"       # correct model name
-)
-
-# ✅ 2.  Load in 4-bit so it fits on Hugging-Face ZeroGPU (<15 GB)
-bnb_cfg = BitsAndBytesConfig(load_in_4bit=True)
+MODEL_ID = "TheBloke/Mixtral-8x7B-Instruct-v0.1-GPTQ"   # ← 8 GB quantised
+bnb_cfg  = BitsAndBytesConfig(load_in_4bit=True)
 
-tok = AutoTokenizer.from_pretrained(MODEL_ID)
+tok = AutoTokenizer.from_pretrained(MODEL_ID, use_fast=True)
 
 model = AutoModelForCausalLM.from_pretrained(
     MODEL_ID,
-    quantization_config=bnb_cfg,          # 4-bit
     device_map="auto",
-    torch_dtype=torch.float16,            # ZeroGPU has a single T4/L4
-    trust_remote_code=True,               # required for Mixtral
+    trust_remote_code=True,
+    quantization_config=bnb_cfg,
 )
 
-# ✅ 3.  Use *text-generation* with an explicit prompt template
 prompt_tmpl = (
     "Summarise the following transcript in short in 1 or 2 paragraph and point wise and don't miss any key information cover all"
 )
 
-gen = pipeline(
-    "text-generation",
-    model=model,
-    tokenizer=tok,
-    max_new_tokens=256,
-    temperature=0.1,
-)
+gen = pipeline("text-generation", model=model, tokenizer=tok,
+               max_new_tokens=256, temperature=0.3)
 
-MAX_CHUNK = 6_000  # ≈4 k tokens
+MAX_CHUNK = 6_000   # ≈ 4 k tokens
 
-def summarize(txt):
+def summarize(txt: str) -> str:
     parts = textwrap.wrap(txt, MAX_CHUNK, break_long_words=False)
     partials = [
-        gen(prompt_tmpl.format(chunk=p))[0]["generated_text"].split("### Summary:")[-1].strip()
+        gen(prompt_tpl.format(chunk=p))[0]["generated_text"]
+        .split("### Summary:")[-1].strip()
         for p in parts
     ]
-    return gen(prompt_tmpl.format(chunk=" ".join(partials)))[0]["generated_text"]\
+    return gen(prompt_tpl.format(chunk=" ".join(partials)))[0]["generated_text"]\
              .split("### Summary:")[-1].strip()
 
-demo = gr.Interface(
-    fn=summarize,
-    inputs=gr.Textbox(lines=20, label="Transcript"),
-    outputs="text",
-    title="Mixtral-8×7B Transcript Summariser",
-)
+demo = gr.Interface(fn=summarize,
+                    inputs=gr.Textbox(lines=20, label="Transcript"),
+                    outputs="text",
+                    title="Free Transcript Summariser – Mixtral-8×7B")
 
 if __name__ == "__main__":
     demo.launch()