Synthetic Data, Caching, Benchmarks, and GEval improvement

For deepeval's latest release v0.21.15, we release:

• Synthetic Data generation. Generate synthetic data from documents easily: https://docs.confident-ai.com/docs/evaluation-datasets-synthetic-data
• caching. If you're running 10k test cases and it fails at the 9999th test case, you no longer have to rerun the first 9999 test case as you can just read from cache using the -c flag: https://docs.confident-ai.com/docs/evaluation-introduction#cache
• repeats. If you want to repeat each test case for statistical significant, use the -r flag: https://docs.confident-ai.com/docs/evaluation-introduction#repeats
• LLM Benchmarks. Supporting popular benchmarks such as MMLU, HellaSwag, and BIG-BH so anyone can evaluate ANY model on research backed benchmarks in a few lines of code.
• G-Eval improvements. The G-Eval metric now supports using logprobs of tokens to find the weighted summed score.

Async Support for Prod

In deepeval v0.20.85:

• asynchronous support throughout deepeval, and no longer using threads. Users can also call individual metrics asynchronously: https://docs.confident-ai.com/docs/metrics-introduction#measuring-metrics-in-async
• improved the way in which you create a custom LLM for evaluation. You'll now have to implement an asynchronous generate() method to use deepeval's async features: https://docs.confident-ai.com/docs/metrics-introduction#using-a-custom-llm
• strict mode for all metrics!
• improve evaluate() function for more customizability: https://docs.confident-ai.com/docs/evaluation-introduction#evaluating-without-pytest

Conversational Metrics and Synthetic Data Generation

In DeepEval's latest release, there is now:

• conversational metrics: https://docs.confident-ai.com/docs/metrics-knowledge-retention. This metric evaluates whether your LLM is able to retain factual information presented to it throughout a conversation
• synthetic data generation. Generate evaluation datasets from scratch: https://docs.confident-ai.com/docs/evaluation-datasets#generate-an-evaluation-dataset

Production Stability

For the newest release, deepeval now is now stable for production use:

• reduced package size
• separated functionality of pytest vs deepeval test run command
• included coverage score for summarization
• fix contextual precision node error
• released docs for better transparency into metrics calculation
• allows users to configure RAGAS metrics for custom embedding models: https://docs.confident-ai.com/docs/metrics-ragas#example
• fixed bugs with checking for package updates

Hugging Face and LlamaIndex integration

For the latest release, DeepEval:

• Supports Hugging Face users by providing real-time evaluations during fine-tuning: https://docs.confident-ai.com/docs/integrations-huggingface
• Supports LlamaIndex users by allowing unit testing of LlamaIndex apps in CI/CD, and offer metrics in LlamaIndex's evaluators: https://docs.confident-ai.com/docs/integrations-llamaindex
• Improvements to accuracy and reliability in Faithfulness and Answer Relevancy
• Summarization Metric now offers explanation
• You can now use ANY LLM for evaluation: https://docs.confident-ai.com/docs/metrics-introduction#using-a-custom-llm

LLM-Evals now support all LangChain chatmodels

• LLM-Evals (LLM evaluated metrics) now support all of langchain's chat models.
• LLMTestCase now has execution_time and cost, useful for those looking to evaluate on these parameters
• minimum_score is now threshold instead, meaning you can now create custom metrics that either have a "minimum" or "maximum" threshold
• LLMEvalMetric is now GEval
• Llamaindex Tracing integration: (https://docs.llamaindex.ai/en/stable/module_guides/observability/observability.html#deepeval)

ALL RAG Metrics now offers score reasoning, and a lot more.

In this release:

• Faithfulness, Answer Relevancy, Contextual Relevancy, Contextual Precision, and Contextual Recall, all offer a reasoning for its given score.
• Azure OpenAI now supported via a single command in the CLI: https://docs.confident-ai.com/docs/metrics-introduction#using-azure-openai
• New Summarization Metric that uses the QAG framework for its implementation: https://docs.confident-ai.com/docs/metrics-summarization
• Pulling datasets from Confident AI now offers an intermediate step for additional data processing before evaluation: https://docs.confident-ai.com/docs/confident-ai-evaluate-datasets#pull-your-dataset-from-confident-ai
• Decoupled imports from transformers, sentence_transformers, and pandas to reduce package size

Lots of new features

Lots of new features this release:

1. JudgementalGPT now allows for different languages - useful for our APAC and European friends
2. RAGAS metrics now supports all OpenAI models - useful for those running into context length issues
3. LLMEvalMetric now returns a reasoning for its score
4. deepeval test run now has hooks that call on test run completion
5. evaluate now displays retrieval_context for RAG evaluation
6. RAGAS metric now displays metric breakdown for all its distinct metrics

Continuous Evaluation

Automatically integrated with Confident AI for continous evaluation throughout the lifetime of your LLM (app):

-log evaluation results and analyze metrics pass / fails
-compare and pick the optimal hyperparameters (eg. prompt templates, chunk size, models used, etc.) based on evaluation results
-debug evaluation results via LLM traces
-manage evaluation test cases / datasets in one place
-track events to identify live LLM responses in production
-add production events to existing evaluation datasets to strength evals over time

Continuous Evaluation

Automatically integrated with Confident AI for continous evaluation throughout the lifetime of your LLM (app):

-log evaluation results and analyze metrics pass / fails
-compare and pick the optimal hyperparameters (eg. prompt templates, chunk size, models used, etc.) based on evaluation results
-debug evaluation results via LLM traces
-manage evaluation test cases / datasets in one place
-track events to identify live LLM responses in production
-add production events to existing evaluation datasets to strength evals over time