def __init__(self):
self.bus = bus.Bus()
self.memory = memory.Memory(self.bus)
self.L2C0 = L2.L2(0, self.bus, self.memory)
self.L2C1 = L2.L2(1, self.bus, self.memory)
self.directoryController = directoryController.DirectoryController(
self.L2C0, self.L2C1, self.memory, self.bus)
self.C0 = chip.Chip(0, self.memory, self.bus, self.directoryController,
self.L2C0)
self.C1 = chip.Chip(1, self.memory, self.bus, self.directoryController,
self.L2C1)
self.window = Tk()
self.tableL1P0C0 = Treeview(self.window, height=2)
self.tableL1P1C0 = Treeview(self.window, height=2)
self.tableL1P0C1 = Treeview(self.window, height=2)
self.tableL1P1C1 = Treeview(self.window, height=2)
self.tableL2C0 = Treeview(self.window, height=4)
self.tableL2C1 = Treeview(self.window, height=4)
self.tableMemory = Treeview(self.window, height=16)
self.t1 = threading.Thread(target=self.C0.run)
self.t2 = threading.Thread(target=self.C1.run)
self.t3 = threading.Thread(target=self.directoryController.runL2toMem)
self.t4 = threading.Thread(target=self.directoryController.runMemtoL2)
self.t5 = threading.Thread(target=self.updateMemories)
def __init__(
self,
outputs,
memorySize,
discountFactor,
learningRate,
learnStart,
img_rows,
img_cols,
img_channels,
):
"""
Parameters:
- outputs: output size
- memorySize: size of the memory that will store each state
- discountFactor: the discount factor (gamma)
- learningRate: learning rate
- learnStart: steps to happen before for learning. Set to 128
"""
self.img_rows = img_rows
self.img_cols = img_cols
self.img_channels = img_channels
self.output_size = outputs
self.memory = memory.Memory(memorySize)
self.discountFactor = discountFactor
self.learnStart = learnStart
self.learningRate = learningRate
def __init__(self,
fm_infile,
seed_hv_infile,
is_early_fusion=True,
use_3_seed=False):
# Feature Memory
self.feature_memory = memory.FeatureMemory(fm_infile)
self.feature_memory.normalize()
self.feature_memory.downSample(self.downsample_interval)
self.feature_memory.discretize()
self.feature_memory.genTrainData(self.learning_rate)
# Seed HV progress over time
self.seed_hv = memory.Memory(seed_hv_infile)
# Associative Memory
memory.AssociativeMemory.dimension = self.dimension
self.associative_memory = memory.AssociativeMemory()
# Spatial Encoder
spatial_encoder.SpatialEncoder.dimension = self.dimension
self.spatial_encoder_GSR = spatial_encoder.SpatialEncoder(
self.num_channel_GSR)
self.spatial_encoder_ECG = spatial_encoder.SpatialEncoder(
self.num_channel_ECG)
self.spatial_encoder_EEG = spatial_encoder.SpatialEncoder(
self.num_channel_EEG)
# Temporal Encoder
temporal_encoder.TemporalEncoder.dimension = self.dimension
temporal_encoder.TemporalEncoder.ngram_size = self.ngram_size
self.temporal_encoder = temporal_encoder.TemporalEncoder(
) # used only when is_early_fusion is True
self.temporal_encoder_GSR = temporal_encoder.TemporalEncoder(
) # used only when is_early_fusion is False
self.temporal_encoder_ECG = temporal_encoder.TemporalEncoder(
) # used only when is_early_fusion is False
self.temporal_encoder_EEG = temporal_encoder.TemporalEncoder(
) # used only when is_early_fusion is False
self.is_early_fusion = is_early_fusion
self.use_3_seed = use_3_seed
# Prediction Statistics
self.correct_prediction = 0
self.wrong_prediction = 0
self.prediction_success_rate = 0
self.correct_v_prediction = 0
self.wrong_v_prediction = 0
self.prediction_v_success_rate = 0
self.correct_a_prediction = 0
self.wrong_a_prediction = 0
self.prediction_a_success_rate = 0
self.predicted_v_history = []
self.predicted_a_history = []
def __init__(self, inputs, outputs, memorySize, discountFactor,
learningRate, learnStart):
self.input_size = inputs
self.output_size = outputs
self.memory = memory.Memory(memorySize)
self.discountFactor = discountFactor
self.learnStart = learnStart
self.learningRate = learningRate
def __init__(self, duration): self.memory = memory.Memory() self.duration = duration for i in range(4): name = "CPU" + str(i) tmp_cpu = Cpu(system=self,name=name) self.cpus.append(tmp_cpu) pass
def __init__(self, txrx):
import memory
import gpio
import eder_logger
self._txrx = txrx
self.memory = memory.Memory()
self.rpi = gpio.EderGpio()
self.logger = eder_logger.EderLogger()
def main():
# how far we want to be from the target
goal_distance = 1
# create the vrep environment
env = vrep_env.make(goal_distance,
rewarder=rewards.graduated(goal_distance))
mem = memory.Memory(MEMORY_CAPACITY,
dims=2 * env.state_dim + env.action_dim + 1)
# path to follow, just keep moving right
#path = ["R"] * 20 + ["exit"]
path = ["F"] * 7 + ["R"] * 7 + ["B"] * 7 + ["L"] * 7 + ["exit"]
# moves the target we are trying to fallow
mover = TargetMover(env.client_id,
target_handle=env.target_handle,
path=path)
# tells us what actions to take
actor = SimpleActor(goal_distance)
cumulative_reward = 0
for ep in range(MAX_EPISODES):
print("Episode: %d, Memory: %d, Average Reward: %f" %
(ep, mem.pointer, cumulative_reward / (ep + 1)))
# done loading sample points
if mem.pointer > MEMORY_CAPACITY:
break
# reset environment and mover
mover.reset()
s = env.reset()
# keep going tell mover or env say stop
while True:
done = mover.step()
if done:
break
a = actor.choose_action(s)
s_, r, done = env.step(a)
mem.store_transition(s, a, r, s_)
s = s_
cumulative_reward += r
print("X Pos: %f, Y Pos : %f, Velocity: L %f R %f, Reward: %f" %
(s[0], s[1], a[0], a[1], r))
if done:
break
#time.sleep(0.1)
print("Average Reward: " + str(cumulative_reward / MAX_EPISODES))
mem.save("/tmp/learn_to_follow_mem_10k")
env.stop()
def __init__(self,
actor,
critic,
memory_size=100000000,
minibatch_size=300):
self.actor = actor
self.critic = critic
self.memory = memory.Memory(memory_size)
self.minibatch_size = minibatch_size
def __init__(self,outputs,memorySize,discountFactor,learningRate,learnStart,img_rows,img_cols,img_channels):
self.output_size=outputs
self.memory=memory.Memory(memorySize)
self.discountFactor=discountFactor
self.learnStart=learnStart
self.learningRate=learningRate
self.img_rows=img_rows
self.img_cols=img_cols
self.img_channels=img_channels
def load(rom_path, config_path=None):
prog_rom = extract.extract(rom_path)
mem = memory.Memory()
mem.map_prog_rom(prog_rom)
del prog_rom
program = dasm_objects.Program(mem)
load_config(config_path, program.symbols, program.config)
return program
def test_empty_init(self):
"""Test default Memory creation"""
# 1. Create default Jumps object
mymem = memory.Memory()
# 2. Make sure it has the default values
self.assertEqual(mymem.part2, False)
self.assertEqual(mymem.banks, [])
def __init__(self, instructions, opcode, opcodeStr, dataval, address, arg1,
arg2, arg3, arg1Str, arg2Str, arg3Str, numInstructions,
destReg, src1Reg, src2Reg):
self.instructions = instructions
self.opcode = opcode
self.opcodeStr = opcodeStr
self.dataval = dataval
self.address = address
self.arg1 = arg1
self.arg2 = arg2
self.arg3 = arg3
self.arg1Str = arg1Str
self.arg2Str = arg2Str
self.arg3Str = arg3Str
self.numInstructions = numInstructions
self.destReg = destReg
self.src1Reg = src1Reg
self.src2Reg = src2Reg
self.PC = 96
self.cycle = 1
### LISTS ###
self.cycleList = [0]
self.R = [
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0
]
self.postMemBuff = [-1,
-1] # first number is value, second is instr index
self.postALUBuff = [-1, -1]
self.preMemBuff = [-1, -1]
self.preALUBuff = [-1, -1]
self.preIssueBuff = [-1, -1, -1, -1]
### OBJECTS ###
self.WB = writeBack.WriteBack(self.R, self.postMemBuff,
self.postALUBuff, destReg)
self.cache = cache.Cache(numInstructions, instructions, dataval,
address)
self.ALU = alu.ALU(self.R, self.postALUBuff, self.preALUBuff,
opcodeStr, arg1, arg2, arg3)
self.MEM = memory.Memory(self.R, self.postMemBuff, self.preMemBuff,
opcodeStr, arg1, arg2, arg3, dataval,
self.numInstructions, self.cache,
self.cycleList)
self.issue = issue.Issue(instructions, opcodeStr, dataval, address,
arg1, arg2, arg3, self.numInstructions,
destReg, src1Reg, src2Reg, self.R,
self.preIssueBuff, self.preMemBuff,
self.postMemBuff, self.preALUBuff,
self.postALUBuff)
self.fetch = fetch.Fetch(instructions, opcodeStr, dataval, address,
arg1, arg2, arg3, self.numInstructions,
destReg, src1Reg, src2Reg, self.R,
self.preIssueBuff, self.preMemBuff,
self.postMemBuff, self.preALUBuff,
self.postALUBuff, self.PC, self.cache)
self.outputFileName = Setup.get_output_filename()
def __init__(self, endpoints, cert, ca_cert, interval, items, handlers,
defaults):
from urlparse import urlparse
from jinja2 import Environment
env = Environment()
self.items = {
n: (Template(n,
env.from_string(n)), Template(v, env.from_string(v)))
for (n, v) in items
}
self.handlers = {
n: (Template(n,
env.from_string(n)), Template(v, env.from_string(v)))
for (n, v) in handlers
}
self.interval = interval
parsed_endpoints = map(urlparse, endpoints.split(','))
if len(parsed_endpoints) <= 0:
raise ValueError("Endpoint parameter is empty or missing")
scheme = parsed_endpoints[0].scheme
if scheme == "stdout":
self.client = StdoutClient()
else:
self.client = EtcdClient(
tuple(
map(
lambda e: (e.hostname, e.port if e.port else 80
if e.scheme == "http" else 443),
parsed_endpoints)), scheme, cert, ca_cert)
self.defaults = TemplateDict(
{n: Template(v, env.from_string(v))
for (n, v) in defaults})
self.context = Context()
import cpu
import memory
import disk
import host
import systemd
import os_support
self.context.add_module(cpu.Cpu())
self.context.add_module(memory.Memory())
self.context.add_module(disk.Disk())
self.context.add_module(host.Host())
self.context.add_module(systemd.Systemd())
self.context.add_module(os_support.OS())
for n in self.handlers:
self.client.add_handler(self.context, self.defaults, n,
*(self.handlers[n]))
def __init__(self, inputs, outputs, memorySize, discountFactor, learningRate, learnStart, maxMargin):
self.sess = k_b.get_session()
self.input_size = inputs
self.output_size = outputs
self.memory = memory.Memory(memorySize)
self.discount_factor = discountFactor
self.learning_rate = learningRate
self.learn_start = learnStart
self.max_margin = maxMargin
def create_database(self):
if os.path.exists("../Database/episodes.npz"):
print("Database exists!")
self.__memory = memory.Memory()
else:
self.__memory = memory.Memory(DATABASE_SIZE)
for i in range(DATABASE_SIZE):
if i % 1000 == 0:
print("{}% completed".format(i / 100))
if (self.__done):
self.reset_environment()
action = self.__agent.action_space.sample()
next_obs, reward, done = self.execute_action(action)
# only include non-terminal states
# assuming game is never ending
if not done:
self.__memory.append(
[self.__obs, action, next_obs, reward])
self.__memory.save_database()
def __init__(self, obs_space, action_space,memorySize, discountFactor, \
learningRate,useTargetNetwork):
self.memory = memory.Memory(memorySize)
self.discountFactor = discountFactor
self.learningRate = learningRate
self.useTargetNetwork = useTargetNetwork
self.count_steps = 0
self.obs_space = obs_space
self.action_space = action_space
self.initNetworks()
def setup_computer(self, start_ip):
"""Build the computer."""
self.reg = cpu.Registers()
self.mem = memory.Memory()
self.alu = cpu.ArithmeticLogicUnit()
self.decoder_obj = decoder.Decoder(self.reg, self.mem, self.alu)
self.clock = cpu.Clock(self.reg, self.decoder_obj)
self.reg.ip = start_ip
def test_save_load():
mem = memory.Memory(10, dims=2 * STATE_DIM + ACTION_DIM + 1)
for i in range(100):
s = i % 4
a = [i % 3, i % 3]
r = i % 2 - 1
s_ = s + 1
mem.store_transition(s, a, r, s_)
saved_data = mem.data
mem.save("/tmp/test_mem_save")
mem = memory.Memory(12, dims=2 * STATE_DIM + ACTION_DIM + 1)
mem.store_transition(1, [1, 1], 1, 1)
mem.load("/tmp/test_mem_save")
assert np.array_equal(saved_data, mem.data)
# memory is full should be pointing to the end
assert mem.pointer == len(mem.data) - 1
assert mem.capacity == 10
def __init__(self, N_ACTIONS, memory_path=None):
if memory_path == None:
### MEMORY HYPERPARAMETERS
# Number of experiences the Memory can keep
self.memory = memory.Memory(1000000)
else:
self.memory = pkl.load(open(memory_path, 'rb'))
self.model = dqn.DQN()
# Set up tensorboard
self.model.set_up_board()
def setUp(self):
self.reg = cpu.Registers()
self.mem = memory.Memory()
self.alu = cpu.ArithmeticLogicUnit()
self.mem_if = decoder.MemoryInterface(self.mem)
self.instr = decoder.Instructions(reg=self.reg,
mem=self.mem,
alu=self.alu)
self.mem.write(ADDR1, VAL1)
def setup_computer():
"""Build the computer."""
reg = cpu.Registers()
mem = memory.Memory()
alu = cpu.ArithmeticLogicUnit()
decoder_obj = decoder.Decoder(reg, mem, alu)
clock = cpu.Clock(reg, decoder_obj)
reg.ip = ADDR(0x20)
return mem, clock
def interpret(ast : List[Node]) -> memory.Memory:
"""Interprets an AST Tree into a memory object
Args:
ast (List): An AST Tree
Returns:
memory.Memory: The memory object with the interpreted AST tree inside of it
"""
mem = memory.Memory()
return interpret_rec(ast, mem)
def test_empty_init(self):
"Test the default Memory creation"
# 1. Create default Memory object
myobj = memory.Memory()
# 2. Make sure it has the default values
self.assertEqual(myobj.part2, False)
self.assertEqual(myobj.text, None)
self.assertEqual(myobj.age, None)
self.assertEqual(myobj.turn, 0)
self.assertEqual(myobj.numbers, {})
def __init__(self):
self.logger = logging.getLogger('Borg.Brain.BodyController')
self.memory = memory.Memory()
self.emergency = False
self.use_marytts = USE_MARYTTS
self.use_ivona = IVONA
self.remote_speech = False
if self.remote_speech:
self.ivona_pub = rospy.Publisher('/ivona_speech', String)
print "[BodyController] using remote ivona speech."
elif self.use_ivona:
pygame.init()
def test_text_init(self):
"""Test Memory creation from text"""
# 1. Create Jumps object from text
mymem = memory.Memory(text=aoc_06.from_text(P1_EXAMPLES_TEXT)[0])
# 2. Make sure it has the specified values
self.assertEqual(mymem.part2, False)
self.assertEqual(len(mymem.banks), 4)
self.assertEqual(mymem.banks, [0, 2, 7, 0])
# 3. Check methods
self.assertEqual(mymem.cycle_until_seen(verbose=False, limit=9), 5)
def basic_test(address):
prog = extract.extract("smb.nes")
mem = memory.Memory()
mem.map_prog_rom(prog)
del prog
for _ in range(0, 50):
instruction = dasm.disassemble_instruction(mem, address)
assert instruction.address == address
address += instruction.size
print("0x{addr:04X}: {asm}".format(addr=instruction.address,
asm=instruction.assembly_string))
def train(session, graph, saver):
ri = RunInfo(
session=session,
graph=graph,
memory=m.Memory(),
)
for epoch_idx in range(1, config.NUM_EPOCHS + 1):
train_epoch(ri, epoch_idx)
print("Saving!")
saver.save(session, config.CHECKPOINT_FILENAME)
if epoch_idx % config.NUM_EPOCHS_PER_EVAL == 0:
play.evaluate_performance(session, graph, training_mode=False)
def __init__(self, factory=None, params=None, ds=None):
if self.__class__ == Workspace and not factory:
import memory
mem = memory.Memory()
self._store = mem._store
else :
if ds:
self._store = ds
elif factory:
self._store = factory.createDataStore(params)
else:
raise Exception('Workspace requires a data store or a factory')
self.factory = factory
self.params = params
def setUp(self):
params = configparse.ParameterDict()
self.IP = '192.168.0.1'
params.add_option('general', 'robot_ip', self.IP )
params.add_option('vision_controller', 'video_source_name', 'test')
params.add_option('general', 'starting_behavior', 'test1')
params.add_option('speech_controller', 'modules', '') #no speech modules, we'll add a test socket later on
params.add_option('speech_controller', 'start_speech', 'true') #no speech modules, we'll add a test socket later on
params.add_option('body','number_of_naos','0')
params.add_option('body','number_of_pioneers','0')
self.mem = memory.Memory()
self.mem.clear()
self.si = sensorintegrator.SensorIntegrator(params)
def part_two(args, input_lines):
"Process part two of the puzzle"
# 1. Create the puzzle solver
solver = memory.Memory(part2=True, text=input_lines[0])
# 2. Determine number of redistribution cycles
solution = solver.cycle_until_seen(verbose=args.verbose, limit=args.limit)
if solution is None:
print("No configuration repeats")
else:
print("A previous configuration was seen in %d redistribution cycles" %
(solution))
# 3. Return result
return solution is not None
