Rename app.py to proyect.toml

app.py

@@ -1,92 +0,0 @@
-import gradio as gr
-from huggingface_hub import hf_hub_download
-import time, traceback, os
-import torch
-import numpy as np
-from fastai.vision.all import *
-
-os.environ["OMP_NUM_THREADS"] = "1"
-torch.set_num_threads(1)
-
-def log(msg):
-    print(f"[DEBUG {time.strftime('%H:%M:%S')}] {msg}", flush=True)
-
-
-# --- CARGA DEL MODELO ---
-repo_id = "daniihc16/chest-xray-classifier"
-filename = "model.pkl"
-
-try:
-    log("Descargando...")
-    model_path = hf_hub_download(repo_id=repo_id, filename=filename)
-    
-    log("Cargando estructura...")
-    learn = load_learner(model_path)
-    model = learn.model.eval() # Extraemos el modelo puro de PyTorch
-    vocab = learn.dls.vocab    # Guardamos las etiquetas
-    
-    log(f"Modelo extraído. Vocabulario: {vocab}")
-    
-except Exception as e:
-    log(f"ERROR FATAL: {e}")
-    raise e
-
-# --- PREDICCIÓN MANUAL (SIN DATALOADERS) ---
-def predict_pure_pytorch(img):
-    start = time.time()
-    log("1. Petición recibida")
-    
-    try:
-        # 1. PREPROCESAMIENTO MANUAL
-        
-        if img.size != (224, 224):
-            img = img.resize((224, 224))
-
-        x = torch.tensor(np.array(img)).float()
-        
-        x = x.permute(2, 0, 1)
-        
-        x /= 255.0
-        mean = torch.tensor([0.485, 0.456, 0.406]).view(3, 1, 1)
-        std = torch.tensor([0.229, 0.224, 0.225]).view(3, 1, 1)
-        x = (x - mean) / std
-        
-        x = x.unsqueeze(0)
-        
-        log(f"   Tensor preparado: {x.shape}")
-
-        # 2. INFERENCIA
-        log("2. Ejecutando forward pass...")
-        with torch.no_grad():
-            out = model(x) # Inferencia directa, sin wrappers
-            probs = torch.softmax(out, dim=1) # Convertir logits a probabilidades
-            
-        probs_np = probs[0].numpy()
-        log(f"3. Probabilidades crudas: {probs_np}")
-        
-        # 3. RESULTADO
-        result = {vocab[i]: float(probs_np[i]) for i in range(len(vocab))}
-        log(f"   Tiempo: {time.time()-start:.3f}s")
-        
-        return result
-
-    except Exception as e:
-        log("ERROR EN PREDICCIÓN MANUAL")
-        traceback.print_exc()
-        return {f"Error: {str(e)}": 0.0}
-
-# --- INTERFAZ ---
-interface = gr.Interface(
-    fn=predict_pure_pytorch,
-    inputs=gr.Image(type="pil"),
-    outputs=gr.Label(),
-    title="Clasificador Radiografías",
-    description="",
-    examples=[
-        "person1_bacteria_1.jpeg",
-        "IM-0115-0001.jpeg"
-    ],
-)
-
-if __name__ == "__main__":
-    interface.launch(server_name="0.0.0.0", server_port=7860, ssr_mode=False)

proyect.toml

@@ -0,0 +1,3 @@
+[build-system]
+requires = ["setuptools>=40.8.0", "wheel", "python=3.12.12", "fastai=2.8.1", "fastcore=1.8.18"]
+build-backend = "setuptools.build_meta:__legacy__"