numpy - using specific forms of gradients of two neural network to minimize a loss w.r.t its parameters -


suppose have loss function f(self.theta, self.sigma) self.theta , self.sigma both defined 2 neural network, example, in tensorflow,

self.sigma = tf.contrib.layers.fully_connected(     inputs=tf.expand_dims(self.state, 0),     num_outputs=1,     activation_fn=none,     weights_initializer=tf.zeros_initializer)

if have derived forms of gradients of f(self.theta, self.sigma) w.r.t self.theta , self.sigma, denoted theta_gradient , sigma_gradient respectively, how can minimize loss function f(self.theta, self.sigma) using these given forms of gradients minimize loss function.

from this answer of similar question, seems can done using tf.apply_gradient when there set parameters neural network instead of two. have no idea how solve above problem. main difficulty here have 2 forms of gradients, theta_gradient theta, sigma_gradient sigma, how apply them optimize loss function unclear.

based on that, first attempt following,

self.optimizer = tf.train.adamoptimizer(learning_rate=learning_rate)           true_gradients_theta, true_gradients_eta = self.optimizer.compute_gradients(self.loss(theta, eta), [self.theta, self.eta])  my_own_gradients_theta = do_some_stuff_with(true_gradients_theta) my_own_gradients_eta = do_some_stuff_with(true_gradients_eta)  self.train_op = self.optimizer.apply_gradients(my_own_gradients_theta, my_own_gradients_eta)

however, there 2 main issues: first 1 tf not allow me self.train_op = self.optimizer.apply_gradients(my_own_gradients_theta, my_own_gradients_eta), second issue can not know mathematical form of true_gradients_theta, true_gradients_eta not know how edit them desirable mathematical form.

compute gradients returns list of pairs - variable , gradient, have create analogous structure , concatenate before passing apply_gradients.

to more specific, apply_gradients accepts argument grads_and_vals of form

grads_and_vars: list of (gradient, variable) pairs returned compute_gradients().

so need prepare list of form, example:

theta = tf.variable(...) sigma = tf.variable(...)  theta_grad = compute_my_gradient_for_theta(theta) # note these have tf ops sigma_grad = compute_my_sigma_gradient(sigma)  opt = tf.train.adamoptimizer(...) opt.apply_gradients([(theta_grad, theta), (sigma_grad, sigma)])

Comments

Post a Comment

Popular posts from this blog

Is there a better way to structure post methods in Class Based Views -

i'm using class based views in django , never specify url in ajax calls , omit action parameter in form because know these requests go through post method of class based view. notice view - profile page make many requests depending on user updating on page - post method become convoluted , if/else mess example: def post(self, request, *args, **kwargs): # handle user changing available date if request.post.get('availabledate') != none: self.updateavailabledate(request) return jsonresponse({'result':'success'}) if request.post.get('newprojectname') != none: creative_user = creativeuserprofile.objects.get(id = request.user.id) project = project.create_project(creative_user, request.post['newprojectname']) project.save() return jsonresponse({'projectid': project.id }) if request.post.get('projectimage') != none , request.files['file']: proj
Read more

performance - Why is XCHG reg, reg a 3 micro-op instruction on modern Intel architectures? -

i'm doing micro-optimization on performance critical part of code , came across sequence of instructions (in at&t syntax): add %rax, %rbx mov %rdx, %rax mov %rbx, %rdx i thought had use case xchg allow me shave instruction , write: add %rbx, %rax xchg %rax, %rdx however, dimay found agner fog's instruction tables , xchg 3 micro-op instruction 2 cycle latency on sandy bridge, ivy bridge, broadwell, haswell , skylake. 3 whole micro-ops , 2 cycles of latency! 3 micro-ops throws off 4-1-1-1 cadence , 2 cycle latency makes worse original in best case since last 2 instructions in original might execute in parallel. now... cpu might breaking instruction micro-ops equivalent to: mov %rax, %tmp mov %rdx, %rax mov %tmp, %rdx where tmp anonymous internal register , suppose last 2 micro-ops run in parallel latency 2 cycles. given register renaming occurs on these micro-architectures, though, doesn't make sense me done way. why wouldn't register renamer
Read more

jquery - Responsive Navbar with Sub Navbar -

good day, forgive me nav vocabulary not best. if question confusing apologize. i'm working on navigation bar needs centered on page @ 768px , above. items in nav open (or perhaps toggle) subnav lives directly below. these nav should appear centered on page. i have been able work through responsive portion of nav bar, having main items align vertically , sub nav menu of each item roll out below, , gave them indent separation. working well. however, issue not responsive side, behavior @ width greater 768px. happens toggling menu's sub navigation pushes other main items down page below sub nav. here's code: $(document).ready(function () { //stick in fixed 100% height behind navbar don't wrap $('#slide-nav.navbar-inverse').after($('<div class="inverse" id="navbar-height-col"></div>')); $('#slide-nav.navbar-default').after($('<div id="navbar-height-col"></d
Read more