- Fastest Inference: mlc stands out as the fastest, prompting a need to assess its quality despite its impressive speed.
- Favorite Tool: CTranslate2 is the preferred choice due to its speed and user-friendliness, supported by excellent documentation. It lacks distributed inference unlike vLLM.
- vLLM Performance: vLLM is also fast but CTranslate outperforms it in speed. However, vLLM supports distributed inference, making it suitable for larger models.
- Text Generation Inference (TGI): An acceptable choice for deploying HuggingFace LLMs traditionally, but not as swift as vLLM. Offers features like telemetry and HF ecosystem integration. Note that TGI’s licensing has become more restrictive as of 7/28/2023, potentially limiting certain commercial uses.
Archives quotidiennes :
The History of Open-Source LLMs: Better Base Models (Part Two)
https://cameronrwolfe.substack.com/p/the-history-of-open-source-llms-better
- Value of Open-source LLM Research: Aims to democratize influential technology; despite initial struggles and criticism, open-source LLMs gained popularity and significance.
- Early Challenges: Initial open-source LLMs performed poorly and faced criticism, posing difficulties for advancement.
- Transformative Research Line: Focuses on enhancing open-source LLMs, leading to high-performing pre-trained models accessible to all.
- Significance of High-Performing Models: Creation of powerful, cost-effective pre-trained LLMs revolutionized research accessibility.
- Series Overview: Part two of a three-part series on open-source LLM history. The first part explored initial open-source LLM attempts.
- Study Focus: This overview delves into the most popular open-source base models, emphasizing pre-trained models not yet fine-tuned or aligned.
- Future Exploration: Subsequent installment will discuss fine-tuning and alignment of models for diverse practical applications.
Advanced Prompt Engineering
https://cameronrwolfe.substack.com/p/advanced-prompt-engineering
The emergence of large language models (LLMs) has revolutionized problem-solving approaches. In the past, tasks like document reformatting or sentence classification necessitated creating specific computer programs. LLMs have transformed this process, enabling tasks to be accomplished through textual prompts. For instance, reformatting documents can be achieved by instructing an LLM. This shift was exemplified by GPT-3’s ability to achieve accurate results with minimal guidance.
As LLM research progressed, more sophisticated techniques emerged beyond basic prompting methods like zero/few-shot learning. Instruction-following LLMs (e.g., InstructGPT, ChatGPT) prompted investigations into tackling complex tasks. The goal was to extend LLMs beyond simple problems, requiring them to comprehend intricate instructions and execute multi-step reasoning. However, such challenges demand advanced prompting strategies due to their complexity.
Practical Prompt Engineering
https://cameronrwolfe.substack.com/p/practical-prompt-engineering-part
- Prompt engineering: An empirical science focused on optimizing LLM (Large Language Model) performance through various prompting strategies.
- Aims to understand prompting mechanics and employs techniques to enhance LLM capabilities.
- Zero/few-shot learning: A fundamental technique where LLMs perform tasks with minimal or no training examples, showcasing their remarkable adaptability.
- Instruction prompting: Another vital technique involving explicit instructions in prompts to guide LLM behavior.
- Overview intends to impart practical insights and strategies for effective prompt engineering and LLM utilization.
- Provides actionable tricks and takeaways for prompt engineers and LLM practitioners to enhance their effectiveness.
The History of Open-Source LLMs: Early Days (Part One)
https://cameronrwolfe.substack.com/p/the-history-of-open-source-llms-early
- Language modeling research traces back to models like GPT, GPT-2, and pre-transformer methods such as ULMFit.
- GPT-3’s proposal marked the initial rise in popularity by showcasing impressive few-shot learning through self-supervised pre-training and in-context learning.
- The recognition of GPT-3 led to the creation of various large language models (LLMs), including InstructGPT and ChatGPT, sparking widespread interest in generative AI.
- Early LLMs often remained closed source, limiting researchers’ understanding and improvement of their workings.
- Open-source variants of popular language models began to emerge gradually, although they initially lagged behind proprietary models in performance.
- These early open-source models laid the groundwork for increased transparency in LLM research and inspired the development of more potent subsequent models like Falcon and LLaMA-21.
- The overview is part of a three-part series that delves into the history of open-source language models, exploring their beginnings, recent developments, and the application of imitation and alignment techniques to enhance their performance.
Large Transformer Model Inference Optimization
https://lilianweng.github.io/posts/2023-01-10-inference-optimization/#quantization
Large transformer models are mainstream nowadays, creating SoTA results for a variety of tasks. They are powerful but very expensive to train and use. The extremely high inference cost, in both time and memory, is a big bottleneck for adopting a powerful transformer for solving real-world tasks at scale.
Why is it hard to run inference for large transformer models? Besides the increasing size of SoTA models, there are two main factors contributing to the inference challenge (Pope et al. 2022):
- Large memory footprint. Both model parameters and intermediate states are needed in memory at inference time. For example,
- The KV cache should be stored in memory during decoding time; E.g. For a batch size of 512 and context length of 2048, the KV cache totals 3TB, that is 3x the model size (!).
- Inference cost from the attention mechanism scales quadratically with input sequence length.
- Low parallelizability. Inference generation is executed in an autoregressive fashion, making the decoding process hard to parallel.
In this post, we will look into several approaches for making transformer inference more efficient. Some are general network compression methods, while others are specific to transformer architecture.
Universal and Transferable Adversarial Attacks on Aligned Language Models
This research examines the safety of large language models (LLMs) such as ChatGPT, Bard, and Claude. It demonstrates the potential for automated creation of adversarial attacks, using character sequences added to user queries that manipulate the LLM into following harmful commands. Unlike traditional « jailbreaks, » these attacks are automated and can affect both open-source and closed-source chatbots. The study raises concerns about the effectiveness of mitigation measures and suggests that the challenges posed by adversarial behavior might persist due to the nature of deep learning models. The findings highlight the need for careful consideration of the safety implications as LLMs become more integrated into various applications.