Skip to main content

No project description provided

Project description

ai-utils

Installation

pip install cyvidia-ai-utils

Cross validation

Huggingface Trainer API

from cyvidia_ai_utils import TransformerCrossValidationModel, cross_validate, EvaluationResult
from transformers import AutoTokenizer, AutoModelForSequenceClassification,TrainingArguments
from datasets import load_dataset, Dataset
from transformers import DataCollatorWithPadding,Trainer

folds = load_dataset("dipesh/Intent-Classification-small",split=[f"train[{k}%:{k+10}%]" for k in range(0, 100, 10)])
assert(isinstance(folds, list))

def create_model():
    model = AutoModelForSequenceClassification.from_pretrained("prajjwal1/bert-tiny", num_labels=21)
    tokenizer= AutoTokenizer.from_pretrained("prajjwal1/bert-tiny")

    return TransformerCrossValidationModel(
        model= model,
        tokenizer= tokenizer,
        training_args= TrainingArguments(
            output_dir=f'tests/test_models/{uuid.uuid4()}',
            learning_rate=2e-5,
            per_device_train_batch_size=16,
            per_device_eval_batch_size=16,
            num_train_epochs=0.001,
            weight_decay=0.01,
            evaluation_strategy="epoch",
            save_strategy="epoch",
            load_best_model_at_end=True,
        )
    )

# folds can either be a list of Dataset's or a dictionary of label to Dataset
folds_dict= {f"train[{k}%:{k+10}%]": fold for k, fold in enumerate(folds)}
results=cross_validate(create_model, folds_dict, target_id_column="label", input_text_column="text")



aggragated_result= EvaluationResult.aggregate(list(results.values()))

Custom Trainer

from cyvidia_ai_utils import CrossValidationModel

class MyCustomCrossValidationModel(CrossValidationModel):
    def get_label_for_id(self, id: int)-> str:
        # Implement

    def train(self, train_ds, val_ds)-> CrossValidationModel:
        # Implement

    def predict_values(self, values)-> Dict[str,Any]:
        # Implement


def create_model():
    return MyCustomCrossValidationModel()

results= cross_validate(create_model, folds, target_id_column="label", input_text_column="text")

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

cyvidia_ai_utils-1.0.4.tar.gz (4.1 kB view details)

Uploaded Source

Built Distribution

cyvidia_ai_utils-1.0.4-py3-none-any.whl (6.1 kB view details)

Uploaded Python 3

File details

Details for the file cyvidia_ai_utils-1.0.4.tar.gz.

File metadata

  • Download URL: cyvidia_ai_utils-1.0.4.tar.gz
  • Upload date:
  • Size: 4.1 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: poetry/1.7.1 CPython/3.12.1 Darwin/23.2.0

File hashes

Hashes for cyvidia_ai_utils-1.0.4.tar.gz
Algorithm Hash digest
SHA256 22bcee1bcd0f9289da56aaaa3a3ff305201ce6213ced8e0d616af2bd1ba3397b
MD5 5fef4fa180929eb2eed34fd5e62d8b05
BLAKE2b-256 dcfb33063cd12b3f4bb81bd061c54c2a2af34a67833c80ceaef070c00fc6f2d2

See more details on using hashes here.

File details

Details for the file cyvidia_ai_utils-1.0.4-py3-none-any.whl.

File metadata

File hashes

Hashes for cyvidia_ai_utils-1.0.4-py3-none-any.whl
Algorithm Hash digest
SHA256 3122ceb232079c5e8919bd06019e9b4c3799315d23cf602e51295cc046ab433d
MD5 7ff5a925ae83b5d1715ec154e0836456
BLAKE2b-256 09b82dec24bae5371118e20394ac26bbe768186d05707bd7d018bc722caaca03

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page