Adafactor: Difference between revisions
| Line 90: | Line 90: | ||
<math>G_t = \begin{bmatrix} 0.3&-0.2&0.4\\ -0.5&0.6&-0.1\\0.2&-0.4 &0.3 \end{bmatrix}</math> | <math>G_t = \begin{bmatrix} 0.3&-0.2&0.4\\ -0.5&0.6&-0.1\\0.2&-0.4 &0.3 \end{bmatrix}</math> | ||
'''<big>Hyperparameters setup</big>''' | '''<big>Hyperparameters setup</big>''' | ||
| Line 104: | Line 102: | ||
<math>\hat{\beta}_{2t} = 1 - t^{-0.8}</math> (Second moment decay) | <math>\hat{\beta}_{2t} = 1 - t^{-0.8}</math> (Second moment decay) | ||
'''<big>Step 1: Learning Rate Scaling</big>''' | '''<big>Step 1: Learning Rate Scaling</big>''' | ||
| Line 136: | Line 132: | ||
<math>\alpha_1 = max(0.001,0.806)\cdot 0.01=0.00806</math> | <math>\alpha_1 = max(0.001,0.806)\cdot 0.01=0.00806</math> | ||
'''<big>Step 2: Compute <math>G^{2}_t</math> (Element-wise Square of Gradient)</big>''' | '''<big>Step 2: Compute <math>G^{2}_t</math> (Element-wise Square of Gradient)</big>''' | ||
Compute the squared value of each element in the gradient matrix '''<math>G_t</math>'''. | |||
<math>G^{2}_t = \begin{bmatrix} 0.3^2&(-0.2)^2&0.4^2\\ (-0.5)^2&0.6^2&(-0.1)^2\\0.2^2&(-0.4)^2 &0.3^2 \end{bmatrix}</math> | <math>G^{2}_t = \begin{bmatrix} 0.3^2&(-0.2)^2&0.4^2\\ (-0.5)^2&0.6^2&(-0.1)^2\\0.2^2&(-0.4)^2 &0.3^2 \end{bmatrix}</math> | ||
| Line 153: | Line 147: | ||
'''<big>Step 3: Find the moment estimate</big>''' | '''<big>Step 3: Find the moment estimate</big>''' | ||
Compute the exponential moving average of squared gradients to capture the variance or scale of gradients. | |||
'''Step 3.1: Compute row moments (<math>R_t</math>)''' | '''Step 3.1: Compute row moments (<math>R_t</math>)''' | ||
| Line 163: | Line 157: | ||
<math>R_t = \hat{\beta_{2t}} \cdot R_{t-1} + (1-\hat{\beta})\cdot (\tfrac{1}{m}\textstyle \sum_{j=1}^m \displaystyle G^{2}_t[i,j]+\epsilon_1) </math> | <math>R_t = \hat{\beta_{2t}} \cdot R_{t-1} + (1-\hat{\beta})\cdot (\tfrac{1}{m}\textstyle \sum_{j=1}^m \displaystyle G^{2}_t[i,j]+\epsilon_1) </math> | ||
Since <math>\hat{\beta}_{2t} = 1 - t^{-0.8}</math>, for first iteration: <math>\hat{\beta}_{21} = 0</math>. And because <math>\epsilon_1 </math> is too small, we ignore it. The update of '''<math> | Since <math>\hat{\beta}_{2t} = 1 - t^{-0.8}</math>, for first iteration: <math>\hat{\beta}_{21} = 0</math>. And because <math>\epsilon_1 </math> is too small, we can ignore it. The update of '''<math>R_t</math>''' is: | ||
<math>R_{1} = \tfrac{1}{m}\textstyle \sum_{j=1}^m \displaystyle G^{2}_t[i,j] </math> | <math>R_{1} = \tfrac{1}{m}\textstyle \sum_{j=1}^m \displaystyle G^{2}_t[i,j] </math> | ||
| Line 189: | Line 183: | ||
The Second Moment Estimate is calculated as the outer product of the row moments ('''<math>R_t</math>''') and column moments ('''<math>C_t</math>'''). | The Second Moment Estimate is calculated as the outer product of the row moments ('''<math>R_t</math>''') and column moments ('''<math>C_t</math>'''). | ||
<math> | <math>\hat{V}_t = R_t \otimes C_t</math> | ||
<math>\hat{V}_1 = \begin{bmatrix} 0.0967\\0.2067\\0.0967 \end{bmatrix} \otimes \begin{bmatrix} 0.1267&0.1867&0.0867\\ \end{bmatrix} </math> | |||
<math> | <math>\hat{V}_1 = \begin{bmatrix} 0.0122&0.0180&0.0084\\ 0.0262&0.0386&0.0179\\ 0.0122&0.0180&0.0084\end{bmatrix} </math> | ||
| Line 228: | Line 223: | ||
Compute RMS of '''<math>U_t </math>''' | |||
'''<small><math>RMS(U_t) = \sqrt{\tfrac{1}{9} \sum_{i=1}^9 U_t[i]^2} \approx 3.303 </math></small>''' | '''<small><math>RMS(U_t) = \sqrt{\tfrac{1}{9} \sum_{i=1}^9 U_t[i]^2} \approx 3.303 </math></small>''' | ||
| Line 241: | Line 236: | ||
'''<big>Step 4: Weight Update (</big>'''<math>X_1 </math>'''<big>)</big>''' | '''<big>Step 4: Weight Update (</big>'''<math>X_1 </math>'''<big>)</big>''' | ||
<math>X_1 = X_0 - \alpha \cdot \hat{U_t}</math> | |||
The result for first iteration | |||
<math>X_1 = \begin{bmatrix} 0.7 &-0.5& 0.9\\ -1.1 & 0.8& -1.6\\1.2&-0.7& 0.4 \end{bmatrix} - 0.00806 \cdot \begin{bmatrix} 0.965&-0.53&1.556 \\-1.1&1.088&-0.266\\0.664&-1.06&1.167 \end{bmatrix} </math> | |||
<math>X_1 = \begin{bmatrix} 0.692&-0.496&0.887 \\-1.091&0.791&-0.596\\ 1.195&-0.691&0.391\end{bmatrix} </math> | |||
Revision as of 16:23, 11 December 2024
Author: Aolei Cao (ac3237), Ziyang Li (zl986), Junjia Liang (jl4439) (ChemE 6800 Fall 2024)
Stewards: Nathan Preuss, Wei-Han Chen, Tianqi Xiao, Guoqing Hu
Introduction
Problem formulation
1. Objective
Minimize the loss function , where and is the weight vector to be optimized.
2. Parameters
- Gradient:
- Second moment estimate:
- Where:
- is the running average of the squared gradient.
- is the corrected decay parameter.
- is a regularization constant.
- Step size:
- Where:
- is the relative step size.
- is a regularization constant.
- is the root mean square, defined as:
3. Algorithms
Adafactor for Weighted Vectors
Inputs:
- Initial point:
- Relative step sizes: for to
- Second moment decay: for to , with
- Regularization constants:
- Clipping threshold:
Algorithm:
- For to :
- Compute adaptive step size:
- Compute gradient:
- Update second moment estimate:
- Compute normalized gradient:
- Apply clipping:
- Update parameter:
- End for
Adafactor for Weighted Matrices
Inputs:
- Initial point:
- Relative step sizes: for to
- Second moment decay: for to , with
- Regularization constants:
- Clipping threshold:
Algorithm:
- For to :
- Compute adaptive step size:
- Compute gradient:
- Update row-wise second moment:
- Update column-wise second moment:
- Update overall second moment estimate:
- Compute normalized gradient:
- Apply clipping:
- Update parameter:
- End for
4. Proposed Hyperparameters for Adafactor
- Regularization constant 1:
- Regularization constant 2:
- Clipping threshold:
- Relative step size:
- Second moment decay:
Numerical Examples
Step-by-step instructions for determining the result of the first iteration.
Problem setup
Initial weights ():
Initial gradient ():
Gradient of the loss function with respect to X
Hyperparameters setup
(Minimum learning rate scaling factor))
(Regularization constant)
(Clipping threshold)
(Relative step size)
(Second moment decay)
Step 1: Learning Rate Scaling
Define the relative step size
Failed to parse (Conversion error. Server ("https://wikimedia.org/api/rest_") reported: "Cannot get mml. Server problem."): {\displaystyle \rho _{1}=\min(10^{-2},1/{\sqrt {1}})=10^{-2}}
Step 1.1: Root Mean Square(RMS) calculation for
Root Mean Square(RMS) calculation for
RMS formula
Failed to parse (Conversion error. Server ("https://wikimedia.org/api/rest_") reported: "Cannot get mml. Server problem."): {\displaystyle RMS(X_{0})={\sqrt {{\tfrac {1}{n}}\textstyle \sum _{i=1}^{n}\displaystyle X_{0}[i]^{2}}}}
Substitute the initial weights
Step 1.2: Find the Learning Rate Scaling ():
Learning rate formula
Failed to parse (Conversion error. Server ("https://wikimedia.org/api/rest_") reported: "Cannot get mml. Server problem."): {\displaystyle \alpha _{1}=max(\epsilon _{2},RMS(X_{0}))\cdot p_{1}}
Substitute the RMS
Step 2: Compute (Element-wise Square of Gradient)
Compute the squared value of each element in the gradient matrix .
Step 3: Find the moment estimate
Compute the exponential moving average of squared gradients to capture the variance or scale of gradients.
Step 3.1: Compute row moments ()
This equation computes the row-wise second moments ( ) as an exponential moving average of past moments () and the current row-wise mean of squared gradients ( Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle G^{2}_t} ), with a balance controlled by (Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle \hat{\beta}_{2t}} ).
For Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle G^{2}_t=\mathbb{R}^{m\times n} }
Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle R_t = \hat{\beta_{2t}} \cdot R_{t-1} + (1-\hat{\beta})\cdot (\tfrac{1}{m}\textstyle \sum_{j=1}^m \displaystyle G^{2}_t[i,j]+\epsilon_1) }
Since Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle \hat{\beta}_{2t} = 1 - t^{-0.8}} , for first iteration: Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle \hat{\beta}_{21} = 0} . And because Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle \epsilon_1 } is too small, we can ignore it. The update of Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle R_t} is:
Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle R_{1} = \tfrac{1}{m}\textstyle \sum_{j=1}^m \displaystyle G^{2}_t[i,j] }
Row-wise mean (Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle R_t} ):
Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle R_1 = \begin{bmatrix} \tfrac{0.09+0.04+0.16}{3} \\ \tfrac{0.25+0.36+0.01}{3}\\\tfrac{0.04+0.16+0.09}{3} \end{bmatrix} = \begin{bmatrix} 0.0967\\ 0.2067\\0.0967\end{bmatrix} }
Step 3.2: Compute column moments (Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle C_t} )
The process is same as row moments
Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle C_t = \hat{\beta}\cdot C_{{t-1}} + (1-\hat{\beta})\cdot (\tfrac{1}{n}\textstyle \sum_{j=1}^n \displaystyle G^{2}_t[i,j]+\epsilon_1) }
Column-wise mean (Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle C_t} ):
Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle C_1 = \begin{bmatrix} \tfrac{0.09+025+0.04}{3} \\ \tfrac{0.04+0.36+0.16}{3}\\\tfrac{0.16+0.01+0.09}{3} \end{bmatrix} = \begin{bmatrix} 0.1267\\ 0.1867\\0.0867\end{bmatrix} }
Step 3.3: Second Moment Estimate (Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle V_t} )
The Second Moment Estimate is calculated as the outer product of the row moments (Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle R_t} ) and column moments (Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle C_t} ).
Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle \hat{V}_t = R_t \otimes C_t}
Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle \hat{V}_1 = \begin{bmatrix} 0.0967\\0.2067\\0.0967 \end{bmatrix} \otimes \begin{bmatrix} 0.1267&0.1867&0.0867\\ \end{bmatrix} }
Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle \hat{V}_1 = \begin{bmatrix} 0.0122&0.0180&0.0084\\ 0.0262&0.0386&0.0179\\ 0.0122&0.0180&0.0084\end{bmatrix} }
Step 4: Update the vector (Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle U_t } )
step 4.1: Find the vector value of Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle U_t }
Formula of Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle U_t }
Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle U_t = \frac{G_t}{\sqrt{V_t+\epsilon_1}} }
Substitute Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle C_t} and Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle V_t}
Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle U_1 = \frac{\begin{bmatrix}0.3&-0.2&0.4 \\ -0.5&0.6&-0.1\\0.2&-0.4&0.3 \end{bmatrix}}{\sqrt{\begin{bmatrix} 0.0122&0.0180&0.0084\\ 0.0262&0.0386&0.0179\\0.0122&0.0180&0.0084 \end{bmatrix}}} }
Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle U_1 = \begin{bmatrix} 2.711&-1.489&4.370\\-3.090&3.055&-0.747\\1.807&-2.978&3.278 \end{bmatrix} }
step 4.2: Clipped Update Vector Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle \hat{U_t} }
Formula of Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle \hat{U_t} }
Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle \hat{U_t} = \frac{U_t}{max(1,\tfrac{RMS(U_t)}{d}) } }
Compute RMS of Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle U_t }
Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle RMS(U_t) = \sqrt{\tfrac{1}{9} \sum_{i=1}^9 U_t[i]^2} \approx 3.303 }
Since RMS(Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle U_t } )>d, scale Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle U_t } by Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle \tfrac{1}{3.303} }
Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle \hat{U_t} = \begin{bmatrix} 0.965&-0.53&1.556 \\-1.1&1.088&-0.266\\0.664&-1.06&1.167 \end{bmatrix} }
Step 4: Weight Update (Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle X_1 } )
Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle X_1 = X_0 - \alpha \cdot \hat{U_t}}
The result for first iteration
Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle X_1 = \begin{bmatrix} 0.7 &-0.5& 0.9\\ -1.1 & 0.8& -1.6\\1.2&-0.7& 0.4 \end{bmatrix} - 0.00806 \cdot \begin{bmatrix} 0.965&-0.53&1.556 \\-1.1&1.088&-0.266\\0.664&-1.06&1.167 \end{bmatrix} }
Failed to parse (SVG (MathML can be enabled via browser plugin): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle X_1 = \begin{bmatrix} 0.692&-0.496&0.887 \\-1.091&0.791&-0.596\\ 1.195&-0.691&0.391\end{bmatrix} }