File size: 4,636 Bytes
cdd6bc5
58616a8
 
 
b3fb4dd
419474a
b3fb4dd
 
58616a8
cdd6bc5
58616a8
1379e6f
 
 
 
b3fb4dd
cdd6bc5
419474a
 
 
cdd6bc5
 
c9510a3
156fabe
c9510a3
58616a8
419474a
156fabe
c9510a3
cdd6bc5
 
58616a8
c9510a3
 
58616a8
c9510a3
cdd6bc5
58616a8
cdd6bc5
58616a8
 
 
 
 
 
 
419474a
 
 
58616a8
 
 
b3fb4dd
58616a8
 
 
b3fb4dd
58616a8
 
 
b3fb4dd
 
1379e6f
b3fb4dd
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
58616a8
 
b3fb4dd
58616a8
 
 
 
 
 
 
 
 
 
 
 
cdd6bc5
 
58616a8
c9510a3
58616a8
 
 
 
 
 
cdd6bc5
 
c9510a3
cdd6bc5
58616a8
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
from fastapi import APIRouter
from datetime import datetime
from datasets import load_dataset
from sklearn.metrics import accuracy_score
import numpy as np
import os
import torch
from torch.utils.data import DataLoader

from .utils.evaluation import AudioEvaluationRequest
from .utils.emissions import tracker, clean_emissions_data, get_space_info
from .utils.data import FFTDataset
from .utils.models import DualEncoder
from .utils.train import Trainer
from .utils.data_utils import collate_fn, Container
import yaml

from dotenv import load_dotenv
load_dotenv()

router = APIRouter()

DESCRIPTION = "Random Baseline"
ROUTE = "/audio"



@router.post(ROUTE, tags=["Audio Task"],
             description=DESCRIPTION)
async def evaluate_audio(request: AudioEvaluationRequest):
    """
    Evaluate audio classification for rainforest sound detection.
    
    Current Model: Random Baseline
    - Makes random predictions from the label space (0-1)
    - Used as a baseline for comparison
    """
    # Get space info
    username, space_url = get_space_info()

    # Define the label mapping
    LABEL_MAPPING = {
        "chainsaw": 0,
        "environment": 1
    }
    # Load and prepare the dataset
    # Because the dataset is gated, we need to use the HF_TOKEN environment variable to authenticate
    dataset = load_dataset(request.dataset_name,token=os.getenv("HF_TOKEN"))
    
    # Split dataset
    train_test = dataset["train"].train_test_split(test_size=request.test_size, seed=request.test_seed)
    test_dataset = train_test["test"]

    # Start tracking emissions
    tracker.start()
    tracker.start_task("inference")

    #--------------------------------------------------------------------------------------------
    # YOUR MODEL INFERENCE CODE HERE
    # Update the code below to replace the random baseline by your model inference within the inference pass where the energy consumption and emissions are tracked.
    #--------------------------------------------------------------------------------------------
    device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
    args_path = 'utils/config.yaml'
    data_args = Container(**yaml.safe_load(open(args_path, 'r'))['Data'])
    model_args = Container(**yaml.safe_load(open(args_path, 'r'))['CNNEncoder'])
    model_args_f = Container(**yaml.safe_load(open(args_path, 'r'))['CNNEncoder_f'])
    conformer_args = Container(**yaml.safe_load(open(args_path, 'r'))['Conformer'])

    test_dataset = FFTDataset(test_dataset)
    test_dl = DataLoader(test_dataset, batch_size=data_args.batch_size, collate_fn=collate_fn)

    model = DualEncoder(model_args, model_args_f, conformer_args)
    model = model.to(device)
    missing, unexpected = model.load_state_dict(torch.load(model_args.checkpoint_path))

    loss_fn = torch.nn.BCEWithLogitsLoss()
    optimizer = torch.optim.Adam(model.parameters(), lr=5e-4)
    trainer = Trainer(model=model, optimizer=optimizer,
                      criterion=loss_fn, output_dim=model_args.output_dim, scaler=None,
                      scheduler=None, train_dataloader=None,
                      val_dataloader=None, device=device,
                      exp_num='test', log_path=None,
                      range_update=None,
                      accumulation_step=1, max_iter=np.inf,
                      exp_name=f"frugal_cnnencoder_inference")
    predictions, acc = trainer.predict(test_dl, device=device)
    # Make random predictions (placeholder for actual model inference)
    true_labels = test_dataset["label"]

    #--------------------------------------------------------------------------------------------
    # YOUR MODEL INFERENCE STOPS HERE
    #--------------------------------------------------------------------------------------------   
    
    # Stop tracking emissions
    emissions_data = tracker.stop_task()
    
    # Calculate accuracy
    accuracy = accuracy_score(true_labels, predictions)
    
    # Prepare results dictionary
    results = {
        "username": username,
        "space_url": space_url,
        "submission_timestamp": datetime.now().isoformat(),
        "model_description": DESCRIPTION,
        "accuracy": float(accuracy),
        "energy_consumed_wh": emissions_data.energy_consumed * 1000,
        "emissions_gco2eq": emissions_data.emissions * 1000,
        "emissions_data": clean_emissions_data(emissions_data),
        "api_route": ROUTE,
        "dataset_config": {
            "dataset_name": request.dataset_name,
            "test_size": request.test_size,
            "test_seed": request.test_seed
        }
    }
    
    return results