Update handler.py
Browse files- handler.py +7 -7
handler.py
CHANGED
|
@@ -1,18 +1,18 @@
|
|
| 1 |
from transformers import AutoTokenizer, AutoModelForSequenceClassification, pipeline
|
| 2 |
-
import torch
|
| 3 |
-
import os
|
| 4 |
|
| 5 |
class EndpointHandler:
|
| 6 |
def __init__(self, model_dir):
|
| 7 |
-
# Hugging Face passe le répertoire du modèle ici
|
| 8 |
self.model_dir = model_dir
|
| 9 |
self.model = AutoModelForSequenceClassification.from_pretrained(model_dir)
|
| 10 |
self.tokenizer = AutoTokenizer.from_pretrained(model_dir)
|
| 11 |
self.labels = ["presentation","projects","skills","education","contact","fallback"]
|
| 12 |
-
# Optionnel : créer un pipeline pour simplifier l'inférence
|
| 13 |
self.classifier = pipeline("text-classification", model=self.model, tokenizer=self.tokenizer)
|
| 14 |
|
| 15 |
-
def
|
| 16 |
-
|
| 17 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
| 18 |
return {"label": outputs[0]["label"], "score": float(outputs[0]["score"])}
|
|
|
|
| 1 |
from transformers import AutoTokenizer, AutoModelForSequenceClassification, pipeline
|
|
|
|
|
|
|
| 2 |
|
| 3 |
class EndpointHandler:
|
| 4 |
def __init__(self, model_dir):
|
|
|
|
| 5 |
self.model_dir = model_dir
|
| 6 |
self.model = AutoModelForSequenceClassification.from_pretrained(model_dir)
|
| 7 |
self.tokenizer = AutoTokenizer.from_pretrained(model_dir)
|
| 8 |
self.labels = ["presentation","projects","skills","education","contact","fallback"]
|
|
|
|
| 9 |
self.classifier = pipeline("text-classification", model=self.model, tokenizer=self.tokenizer)
|
| 10 |
|
| 11 |
+
def __call__(self, request):
|
| 12 |
+
"""
|
| 13 |
+
Hugging Face Default container attend __call__ comme point d'entrée.
|
| 14 |
+
`request` est le payload JSON reçu par l'endpoint.
|
| 15 |
+
"""
|
| 16 |
+
text = request.get("inputs", "")
|
| 17 |
+
outputs = self.classifier(text)
|
| 18 |
return {"label": outputs[0]["label"], "score": float(outputs[0]["score"])}
|