restore step logging, remove adamant (for now)

ars.lua

@@ -51,30 +51,6 @@ local function kinda_lipschitz(dir, pos, neg, mid)
     return max(l0, l1) / (2 * dev)
 end
 
-local function amsgrad(step) -- in-place! -- TODO: fix this.
-    if mom1 == nil then mom1 = nn.zeros(#step) end
-    if mom2 == nil then mom2 = nn.zeros(#step) end
-    if mom2max == nil then mom2max = nn.zeros(#step) end
-
-    local b1_t = pow(cfg.adam_b1, epoch_i)
-    local b2_t = pow(cfg.adam_b2, epoch_i)
-
-    -- NOTE: with LuaJIT, splitting this loop would
-    --       almost certainly be faster.
-    for i, v in ipairs(step) do
-        mom1[i] = cfg.adam_b1 * mom1[i] + (1 - cfg.adam_b1) * v
-        mom2[i] = cfg.adam_b2 * mom2[i] + (1 - cfg.adam_b2) * v * v
-        mom2max[i] = max(mom2[i], mom2max[i])
-        if cfg.adam_debias then
-            local num = (mom1[i] / (1 - b1_t))
-            local den = sqrt(mom2max[i] / (1 - b2_t)) + cfg.adam_eps
-            step[i] = num / den
-        else
-            step[i] = mom1[i] / (sqrt(mom2max[i]) + cfg.adam_eps)
-        end
-    end
-end
-
 function Ars:init(dims, popsize, poptop, learning_rate, sigma, antithetic)
     self.dims = dims
     self.popsize = popsize or 4 + (3 * floor(log(dims)))
@@ -201,6 +177,8 @@ function Ars:tell(scored, unperturbed_score)
     end
 
     self.noise = nil
+
+    return step
 end
 
 return {

main.lua

@@ -17,9 +17,6 @@ local trial_neg = true
 local trial_params --= {}
 local trial_rewards = {}
 local trials_remaining = 0
-local mom1 -- first moments in AMSgrad.
-local mom2 -- second moments in AMSgrad.
-local mom2max -- running element-wise maximum of mom2.
 local es -- evolution strategy.
 
 local trial_frames = 0
@@ -114,10 +111,9 @@ local log_map = {
     delta_mean = 4,
     delta_std = 5,
     step_std = 6,
-    adam_std = 7,
-    weight_mean = 8,
-    weight_std = 9,
-    test_trial = 10,
+    weight_mean = 7,
+    weight_std = 8,
+    test_trial = 9,
 }
 
 local function log_csv(t)
@@ -263,28 +259,16 @@ local function learn_from_epoch()
         end
     end
 
+    local step
     if cfg.es == 'ars' and cfg.ars_lips then
-        es:tell(trial_rewards, current_cost)
+        step = es:tell(trial_rewards, current_cost)
     else
-        es:tell(trial_rewards)
+        step = es:tell(trial_rewards)
     end
 
-    local step_mean, step_dev = 0, 0
-    --[[ TODO
     local step_mean, step_dev = calc_mean_dev(step)
     print("step mean:", step_mean)
     print("step stddev:", step_dev)
-    --]]
-
-    local momstep_mean, momstep_dev = 0, 0
-    --[[ TODO
-    if cfg.adamant then
-        amsgrad(step)
-        momstep_mean, momstep_dev = calc_mean_dev(step)
-        print("amsgrad mean:", momstep_mean)
-        print("amsgrad stddev:", momstep_dev)
-    end
-    --]]
 
     base_params = es:params()
 
@@ -320,7 +304,6 @@ local function learn_from_epoch()
         delta_mean = delta_mean,
         delta_std = delta_std,
         step_std = step_dev,
-        adam_std = momstep_dev,
         weight_mean = weight_mean,
         weight_std = weight_std,
         test_trial = current_cost or 0,

snes.lua

@@ -244,8 +244,13 @@ function Snes:tell(scored)
         end
     end
 
+    local step = {}
+    for i, v in ipairs(g_mean) do
+        step[i] = self.std[i] * v
+    end
+
     for i, v in ipairs(self.mean) do
-        self.mean[i] = v + self.mean_adapt * self.std[i] * g_mean[i]
+        self.mean[i] = v + self.mean_adapt * step[i]
     end
 
     local otherwise = {}
@@ -257,6 +262,8 @@ function Snes:tell(scored)
     end
 
     self:adapt(asked, otherwise, utility)
+
+    return step
 end
 
 function Snes:adapt(asked, otherwise, qualities)

xnes.lua

@@ -179,9 +179,14 @@ function Xnes:tell(scored, noise)
 
     -- finally, update according to the gradients.
 
+    local step = {}
+    for i, v in ipairs(dotted) do
+        step[i] = self.sigma * v
+    end
+
     local dotted = dot_mv(self.covars, g_delta)
     for i, v in ipairs(self.mean) do
-        self.mean[i] = v + self.mean_adapt * self.sigma * dotted[i]
+        self.mean[i] = v + self.mean_adapt * step[i]
     end
 
     --[[
@@ -201,6 +206,8 @@ function Xnes:tell(scored, noise)
     --self.sigma = exp(self.log_sigma)
     --for i, v in ipairs(self.log_covars) do self.covars[i] = exp(v) end
     self.noise = nil
+
+    return step
 end
 
 return {