Selecting Large Language Model to Fine-tune via Rectified Scaling Law
- URL: http://arxiv.org/abs/2402.02314v3
- Date: Tue, 28 May 2024 16:16:42 GMT
- Title: Selecting Large Language Model to Fine-tune via Rectified Scaling Law
- Authors: Haowei Lin, Baizhou Huang, Haotian Ye, Qinyu Chen, Zihao Wang, Sujian Li, Jianzhu Ma, Xiaojun Wan, James Zou, Yitao Liang,
- Abstract summary: Given constrained resources, fine-tuning all models and making selections afterward is unrealistic.
We find that the fine-tuning scaling curve includes not just the well-known "power phase" but also the previously unobserved "pre-power phase"
By leveraging our law, we propose a novel LLM selection algorithm that selects the near-optimal model with hundreds of times less resource consumption.
- Score: 74.84096546112215
- License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
- Abstract: The ever-growing ecosystem of LLMs has posed a challenge in selecting the most appropriate pre-trained model to fine-tune amidst a sea of options. Given constrained resources, fine-tuning all models and making selections afterward is unrealistic. In this work, we formulate this resource-constrained selection task into predicting fine-tuning performance and illustrate its natural connection with Scaling Law. Unlike pre-training, we find that the fine-tuning scaling curve includes not just the well-known "power phase" but also the previously unobserved "pre-power phase". We also explain why existing Scaling Law fails to capture this phase transition phenomenon both theoretically and empirically. To address this, we introduce the concept of "pre-learned data size" into our Rectified Scaling Law, which overcomes theoretical limitations and fits experimental results much better. By leveraging our law, we propose a novel LLM selection algorithm that selects the near-optimal model with hundreds of times less resource consumption, while other methods may provide negatively correlated selection. The project page is available at rectified-scaling-law.github.io.
Related papers
- Why Has Predicting Downstream Capabilities of Frontier AI Models with Scale Remained Elusive? [26.04581530766348]
We show that modeling scaling behavior on widely used multiple-choice question-answering benchmarks is challenging.
We show that downstream performance is computed from negative log likelihoods via a sequence of transformations that progressively degrade the statistical relationship between performance and scale.
We empirically study how probability mass on the correct choice co-varies with probability mass on incorrect choices with increasing compute, suggesting that scaling laws for incorrect choices might be achievable.
arXiv Detail & Related papers (2024-06-06T17:46:56Z) - Temporal Scaling Law for Large Language Models [24.12384260752973]
We propose the novel concept of Temporal Scaling Law, studying how the test loss of an LLM evolves as the training steps scale up.
In contrast to modeling the test loss as a whole in a coarse-grained manner, we break it down and dive into the fine-grained test loss of each token position.
We derive the much more precise temporal scaling law by studying the temporal patterns of the parameters in the dynamic hyperbolic-law.
arXiv Detail & Related papers (2024-04-27T05:49:11Z) - A Dynamical Model of Neural Scaling Laws [79.59705237659547]
We analyze a random feature model trained with gradient descent as a solvable model of network training and generalization.
Our theory shows how the gap between training and test loss can gradually build up over time due to repeated reuse of data.
arXiv Detail & Related papers (2024-02-02T01:41:38Z) - Sparse is Enough in Fine-tuning Pre-trained Large Language Models [98.46493578509039]
We propose a gradient-based sparse fine-tuning algorithm, named Sparse Increment Fine-Tuning (SIFT)
We validate its effectiveness on a range of tasks including the GLUE Benchmark and Instruction-tuning.
arXiv Detail & Related papers (2023-12-19T06:06:30Z) - A Pseudo-Semantic Loss for Autoregressive Models with Logical
Constraints [87.08677547257733]
Neuro-symbolic AI bridges the gap between purely symbolic and neural approaches to learning.
We show how to maximize the likelihood of a symbolic constraint w.r.t the neural network's output distribution.
We also evaluate our approach on Sudoku and shortest-path prediction cast as autoregressive generation.
arXiv Detail & Related papers (2023-12-06T20:58:07Z) - nanoLM: an Affordable LLM Pre-training Benchmark via Accurate Loss Prediction across Scales [65.01417261415833]
We present an approach to predict the pre-training loss based on our observations that Maximal Update Parametrization (muP) enables accurate fitting of scaling laws.
With around 14% of the one-time pre-training cost, we can accurately forecast the loss for models up to 52B.
Our goal with nanoLM is to empower researchers with limited resources to reach meaningful conclusions on large models.
arXiv Detail & Related papers (2023-04-14T00:45:01Z) - Scaling Laws Beyond Backpropagation [64.0476282000118]
We study the ability of Direct Feedback Alignment to train causal decoder-only Transformers efficiently.
We find that DFA fails to offer more efficient scaling than backpropagation.
arXiv Detail & Related papers (2022-10-26T10:09:14Z) - Scaling Laws for Deep Learning [1.90365714903665]
In this thesis we take a systematic approach to address the algorithmic and methodological limitations at the root of these costs.
We first demonstrate that deep learning training and pruning are predictable and governed by scaling laws.
We then show through the exploration of a noiseless realizable case that DL is in fact dominated by error sources very far from the lower error limit.
arXiv Detail & Related papers (2021-08-17T15:37:05Z)
This list is automatically generated from the titles and abstracts of the papers in this site.
This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.