Messing with configuration.

app.py

@@ -2,8 +2,8 @@ import spaces
 import gradio
 import json
 import torch
-from transformers import AutoTokenizer
 from optimum.onnxruntime import ORTModelForSequenceClassification
+from transformers import AutoTokenizer
 from optimum.pipelines import pipeline
 from fastapi import FastAPI
 from fastapi.middleware.cors import CORSMiddleware
@@ -29,22 +29,22 @@ print(f"CUDA device: {torch.cuda.get_device_name(torch.cuda.current_device())}")
 # "Xenova/deBERTa-v3-base-mnli" "MoritzLaurer/DeBERTa-v3-base-mnli" Still a bit slow and not great answers
 # "xenova/nli-deberta-v3-small" "cross-encoder/nli-deberta-v3-small" Was using this for a good while and it was...okay
 
-model_name = "MoritzLaurer/deberta-v3-base-zeroshot-v2.0"
-file_name = "onnx/model.onnx"
-tokenizer_name = "MoritzLaurer/deberta-v3-base-zeroshot-v2.0"
+# model_name = "MoritzLaurer/deberta-v3-base-zeroshot-v2.0"
+# file_name = "onnx/model.onnx"
+# tokenizer_name = "MoritzLaurer/deberta-v3-base-zeroshot-v2.0"
 
-model = ORTModelForSequenceClassification.from_pretrained(model_name, export=True, provider="CUDAExecutionProvider")
-tokenizer = AutoTokenizer.from_pretrained(tokenizer_name, model_max_length=512)
+# model = ORTModelForSequenceClassification.from_pretrained(model_name, export=True, provider="CUDAExecutionProvider")
+# tokenizer = AutoTokenizer.from_pretrained(tokenizer_name, model_max_length=512)
 
-# model = ORTModelForSequenceClassification.from_pretrained(
-#     "distilbert-base-uncased-finetuned-sst-2-english",
-#     export=True,
-#     provider="CUDAExecutionProvider",
-# )
+model = ORTModelForSequenceClassification.from_pretrained(
+    "philschmid/tiny-bert-sst2-distilled",
+    export=True,
+    provider="CUDAExecutionProvider",
+)
 
-# tokenizer = AutoTokenizer.from_pretrained("distilbert-base-uncased-finetuned-sst-2-english")
+tokenizer = AutoTokenizer.from_pretrained("philschmid/tiny-bert-sst2-distilled")
 
-classifier = pipeline(task="zero-shot-classification", model=model, tokenizer=tokenizer, device="cuda:0")
+# classifier = pipeline(task="zero-shot-classification", model=model, tokenizer=tokenizer, device="cuda:0")
 
 def classify(data_string, request: gradio.Request):
     if request:
@@ -56,9 +56,10 @@ def classify(data_string, request: gradio.Request):
     # else:
     return zero_shot_classification(data)
 
-@spaces.GPU()
+# @spaces.GPU()
 def zero_shot_classification(data):
-    results = classifier(data['sequence'], candidate_labels=data['candidate_labels'], hypothesis_template=data['hypothesis_template'], multi_label=data['multi_label'])
+    results = []
+    # classifier(data['sequence'], candidate_labels=data['candidate_labels'], hypothesis_template=data['hypothesis_template'], multi_label=data['multi_label'])
     response_string = json.dumps(results)
     return response_string