Update README.md

README.md

@@ -3,7 +3,7 @@ license: apache-2.0
 ---
 ONNX format of voxreality/whisper-small-el-adapters model
 
-Run on CPU
+Model inference example:
 
 ```python
 from optimum.onnxruntime import ORTModelForSpeechSeq2Seq
@@ -14,48 +14,13 @@ from datasets import Audio, load_dataset
 model_id = "voxreality/whisper-small-el-adapters-onnx"
 processor = WhisperProcessor.from_pretrained(model_id)
 model = ORTModelForSpeechSeq2Seq.from_pretrained(model_id)
-# model.save_pretrained("whisper-small-el-finetune_onnx")
-forced_decoder_ids = processor.get_decoder_prompt_ids(language="greek", task="transcribe")
-# load streaming dataset and read first audio sample
-ds = load_dataset("mozilla-foundation/common_voice_11_0", "el", split="test", streaming=True)
-ds = ds.cast_column("audio", Audio(sampling_rate=16_000))
-input_speech = next(iter(ds))["audio"]
-input_features = processor(input_speech["array"], sampling_rate=input_speech["sampling_rate"], return_tensors="pt").input_features
-
-# Generate token ids (without 'use_adapters', as it is not needed)
-predicted_ids = model.generate(input_features, forced_decoder_ids=forced_decoder_ids)
-
-# Decode token ids to text
-transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True)
-
-print(transcription)
-```
-
-Run on GPU
-
-```python
-from optimum.onnxruntime import ORTModelForSpeechSeq2Seq
-from transformers import WhisperProcessor
-from datasets import Audio, load_dataset
-
-model_id = "voxreality/whisper-small-el-adapters-onnx"  # ONNX model path
-
-# Load model with CUDA provider for GPU inference
-model = ORTModelForSpeechSeq2Seq.from_pretrained(model_id, provider="CUDAExecutionProvider")
-
-# Load processor
-processor = WhisperProcessor.from_pretrained("voxreality/whisper-small-el-adapters")
-
-# Get decoder prompt ids
 forced_decoder_ids = processor.get_decoder_prompt_ids(language="greek", task="transcribe")
 
 # Load streaming dataset and read first audio sample
 ds = load_dataset("mozilla-foundation/common_voice_11_0", "el", split="test", streaming=True)
 ds = ds.cast_column("audio", Audio(sampling_rate=16_000))
 input_speech = next(iter(ds))["audio"]
-
-# Process input features
-input_features = processor(input_speech["array"], sampling_rate=input_speech["sampling_rate"], return_tensors="pt").input_features.to('cuda')  # Move inputs to GPU
+input_features = processor(input_speech["array"], sampling_rate=input_speech["sampling_rate"], return_tensors="pt").input_features
 
 # Generate token ids (without 'use_adapters', as it is not needed)
 predicted_ids = model.generate(input_features, forced_decoder_ids=forced_decoder_ids)