Remove broken handler.py (needs WAV decoding fix)

handler.py

@@ -1,70 +0,0 @@
-from typing import Dict, Any
-import torch
-from transformers import WhisperProcessor, WhisperForConditionalGeneration
-
-class EndpointHandler:
-    def __init__(self, path=""):
-        """
-        Initialize the handler with the model
-        """
-        self.processor = WhisperProcessor.from_pretrained(path, language="ar", task="transcribe")
-        self.model = WhisperForConditionalGeneration.from_pretrained(
-            path,
-            torch_dtype=torch.float16
-        ).to("cuda" if torch.cuda.is_available() else "cpu")
-        self.model.eval()
-        
-    def __call__(self, data: Dict[str, Any]) -> Dict[str, Any]:
-        """
-        Process the input data and return predictions with word timestamps
-        
-        Args:
-            data: dictionary with 'inputs' key containing audio bytes
-        
-        Returns:
-            dictionary with 'text' and 'words' (with timestamps)
-        """
-        # Get audio from request
-        inputs = data.pop("inputs", data)
-        
-        # Process audio
-        audio_inputs = self.processor(
-            inputs,
-            sampling_rate=16000,
-            return_tensors="pt"
-        )
-        
-        input_features = audio_inputs.input_features.to(self.model.device)
-        
-        # Generate with timestamps
-        with torch.no_grad():
-            generated_ids = self.model.generate(
-                input_features,
-                language="ar",
-                task="transcribe",
-                max_new_tokens=448,
-                return_timestamps=True  # Enable word timestamps
-            )
-        
-        # Decode with word-level timestamps
-        result = self.processor.batch_decode(
-            generated_ids,
-            skip_special_tokens=True,
-            return_timestamps='word'  # Request word-level
-        )[0]
-        
-        # Format response
-        text = result.get('text', '')
-        words = []
-        
-        for chunk in result.get('chunks', []):
-            words.append({
-                "text": chunk['text'],
-                "start": chunk['timestamp'][0],
-                "end": chunk['timestamp'][1]
-            })
-        
-        return {
-            "text": text,
-            "words": words
-        }