add option 'random_length'
Browse files- automata.py +21 -7
automata.py
CHANGED
@@ -117,7 +117,12 @@ class AutomatonSampler:
|
|
117 |
data_config['length'] = 20
|
118 |
self.T = self.data_config['length']
|
119 |
|
120 |
-
|
|
|
|
|
|
|
|
|
|
|
121 |
|
122 |
def f(self, x):
|
123 |
"""
|
@@ -128,6 +133,11 @@ class AutomatonSampler:
|
|
128 |
def sample(self):
|
129 |
raise NotImplementedError()
|
130 |
|
|
|
|
|
|
|
|
|
|
|
131 |
def help(self):
|
132 |
print(self.__info__)
|
133 |
|
@@ -146,7 +156,8 @@ class BinaryInputSampler(AutomatonSampler):
|
|
146 |
raise NotImplementedError()
|
147 |
|
148 |
def sample(self):
|
149 |
-
|
|
|
150 |
return x, self.f(x)
|
151 |
|
152 |
class ParitySampler(BinaryInputSampler):
|
@@ -258,8 +269,9 @@ class ABABSampler(BinaryInputSampler):
|
|
258 |
def sample(self):
|
259 |
pos_sample = np.random.random() < self.prob_abab_pos_sample
|
260 |
if pos_sample:
|
261 |
-
|
262 |
-
x
|
|
|
263 |
x = np.array(x)
|
264 |
return x, self.f(x)
|
265 |
else:
|
@@ -295,9 +307,10 @@ class FlipFlopSampler(AutomatonSampler):
|
|
295 |
return np.array(states)
|
296 |
|
297 |
def sample(self):
|
298 |
-
|
|
|
299 |
nonzero_pos = (rand < 0.5).astype(np.int64)
|
300 |
-
writes = np.random.choice(range(1, self.n_states+1), size=
|
301 |
x = writes * nonzero_pos
|
302 |
return x, self.f(x)
|
303 |
|
@@ -389,7 +402,8 @@ class SymmetricSampler(AutomatonSampler):
|
|
389 |
return np.array(labels)
|
390 |
|
391 |
def sample(self):
|
392 |
-
|
|
|
393 |
|
394 |
return x, self.f(x)
|
395 |
|
|
|
117 |
data_config['length'] = 20
|
118 |
self.T = self.data_config['length']
|
119 |
|
120 |
+
if 'random_length' not in data_config:
|
121 |
+
data_config['random_length'] = 0
|
122 |
+
self.random_length = data_config['random_length']
|
123 |
+
|
124 |
+
self.__info__ = " - T (int): sequence length.\n" \
|
125 |
+
+ " - random_length (int in {0, 1}): whether to randomly sample a length per sample.\n"
|
126 |
|
127 |
def f(self, x):
|
128 |
"""
|
|
|
133 |
def sample(self):
|
134 |
raise NotImplementedError()
|
135 |
|
136 |
+
def sample_length(self):
|
137 |
+
if self.random_length:
|
138 |
+
return self.np_rng.choice(range(1, self.T+1))
|
139 |
+
return self.T
|
140 |
+
|
141 |
def help(self):
|
142 |
print(self.__info__)
|
143 |
|
|
|
156 |
raise NotImplementedError()
|
157 |
|
158 |
def sample(self):
|
159 |
+
T = self.sample_length()
|
160 |
+
x = self.np_rng.binomial(1, self.prob1, size=T)
|
161 |
return x, self.f(x)
|
162 |
|
163 |
class ParitySampler(BinaryInputSampler):
|
|
|
269 |
def sample(self):
|
270 |
pos_sample = np.random.random() < self.prob_abab_pos_sample
|
271 |
if pos_sample:
|
272 |
+
T = self.sample_length()
|
273 |
+
x = [0,1,0,1] * (T//4)
|
274 |
+
x += [0,1,0,1][:(T%4)]
|
275 |
x = np.array(x)
|
276 |
return x, self.f(x)
|
277 |
else:
|
|
|
307 |
return np.array(states)
|
308 |
|
309 |
def sample(self):
|
310 |
+
T = self.sample_length()
|
311 |
+
rand = np.random.uniform(size=T)
|
312 |
nonzero_pos = (rand < 0.5).astype(np.int64)
|
313 |
+
writes = np.random.choice(range(1, self.n_states+1), size=T)
|
314 |
x = writes * nonzero_pos
|
315 |
return x, self.f(x)
|
316 |
|
|
|
402 |
return np.array(labels)
|
403 |
|
404 |
def sample(self):
|
405 |
+
T = self.sample_length()
|
406 |
+
x = np.random.choice(range(self.n_actions), replace=True, size=T)
|
407 |
|
408 |
return x, self.f(x)
|
409 |
|