File size: 4,160 Bytes
4d6e8c2 1eab01a 4d6e8c2 822b9e4 c6e64ec 4d6e8c2 b1183e7 1c33274 70f5f26 1c33274 70f5f26 4d6e8c2 70f5f26 4d6e8c2 70f5f26 4d6e8c2 70f5f26 b1183e7 70f5f26 b1183e7 4d6e8c2 70f5f26 4d6e8c2 1c33274 4d6e8c2 |
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 |
from fastapi import APIRouter
from datetime import datetime
from datasets import load_dataset
from sklearn.metrics import accuracy_score
import random
import torch
from .utils.evaluation import TextEvaluationRequest
from .utils.emissions import tracker, clean_emissions_data, get_space_info
from transformers import AutoTokenizer,BertForSequenceClassification,AutoModelForSequenceClassification,Trainer, TrainingArguments,DataCollatorWithPadding
import numpy as np
router = APIRouter()
DESCRIPTION = "ModernBert Baseline"
ROUTE = "/text"
@router.post(ROUTE, tags=["Text Task"],
description=DESCRIPTION)
async def evaluate_text(request: TextEvaluationRequest):
"""
Evaluate text classification for climate disinformation detection.
Current Model: Random Baseline
- Makes random predictions from the label space (0-7)
- Used as a baseline for comparison
"""
# Get space info
username, space_url = get_space_info()
# Define the label mapping
LABEL_MAPPING = {
"0_not_relevant": 0,
"1_not_happening": 1,
"2_not_human": 2,
"3_not_bad": 3,
"4_solutions_harmful_unnecessary": 4,
"5_science_unreliable": 5,
"6_proponents_biased": 6,
"7_fossil_fuels_needed": 7
}
# Load and prepare the dataset
dataset = load_dataset(request.dataset_name)
# Convert string labels to integers
dataset = dataset.map(lambda x: {"label": LABEL_MAPPING[x["label"]]})
# 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.
#--------------------------------------------------------------------------------------------
# Make random predictions (placeholder for actual model inference)
true_labels = test_dataset["label"]
predictions = [random.randint(0, 7) for _ in range(len(true_labels))]
#--------------------------------------------------------------------------------------------
# YOUR MODEL INFERENCE STOPS HERE
#--------------------------------------------------------------------------------------------
## Model loading
model = AutoModelForSequenceClassification.from_pretrained("Rcarvalo/test_modernbert_finetuned_v2")
tokenizer = AutoTokenizer.from_pretrained("answerdotai/ModernBERT-base")
## Data prep
def preprocess_function(df):
return tokenizer(df["quote"], truncation=True)
tokenized_test = test_dataset.map(preprocess_function, batched=True)
## Modify inference model
training_args = torch.load("./tasks/utils/training_args.bin")
training_args.eval_strategy='no'
trainer = Trainer(
model=model,
args=training_args,
tokenizer=tokenizer
)
## prediction
preds = trainer.predict(tokenized_test)
predictions = np.array([np.argmax(x) for x in preds[0]])
# 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 |