Apply temperature mods to all SRG models
(Honestly this needs to be base classed at this point)
This commit is contained in:
parent
7659bd6818
commit
1e4083a35b
|
@ -204,12 +204,14 @@ class ConfigurableSwitchedResidualGenerator2(nn.Module):
|
|||
|
||||
def update_for_step(self, step, experiments_path='.'):
|
||||
if self.attentions:
|
||||
temp = max(1, int(self.init_temperature * (self.final_temperature_step - step) / self.final_temperature_step))
|
||||
if temp == 1 and self.heightened_final_step and self.heightened_final_step != 1:
|
||||
temp = max(1,
|
||||
1 + self.init_temperature * (self.final_temperature_step - step) / self.final_temperature_step)
|
||||
if temp == 1 and self.heightened_final_step and step > self.final_temperature_step and \
|
||||
self.heightened_final_step != 1:
|
||||
# Once the temperature passes (1) it enters an inverted curve to match the linear curve from above.
|
||||
# without this, the attention specificity "spikes" incredibly fast in the last few iterations.
|
||||
h_steps_total = self.heightened_final_step - self.final_temperature_step
|
||||
h_steps_current = max(min(step - self.final_temperature_step, h_steps_total), 1)
|
||||
h_steps_current = min(step - self.final_temperature_step, h_steps_total)
|
||||
# The "gap" will represent the steps that need to be traveled as a linear function.
|
||||
h_gap = 1 / self.heightened_temp_min
|
||||
temp = h_gap * h_steps_current / h_steps_total
|
||||
|
@ -217,7 +219,7 @@ class ConfigurableSwitchedResidualGenerator2(nn.Module):
|
|||
temp = 1 / temp
|
||||
self.set_temperature(temp)
|
||||
if step % 50 == 0:
|
||||
[save_attention_to_image(experiments_path, self.attentions[i], self.transformation_counts, step, "a%i" % (i+1,)) for i in range(len(self.switches))]
|
||||
save_attention_to_image(experiments_path, self.attentions[0], self.transformation_counts, step, "a%i" % (1,), l_mult=10)
|
||||
|
||||
def get_debug_values(self, step):
|
||||
temp = self.switches[0].switch.temperature
|
||||
|
@ -231,6 +233,17 @@ class ConfigurableSwitchedResidualGenerator2(nn.Module):
|
|||
return val
|
||||
|
||||
|
||||
def load_state_dict(self, state_dict, strict=True):
|
||||
# Support backwards compatibility where accumulator_index and accumulator_filled are not in this state_dict
|
||||
t_state = self.state_dict()
|
||||
if 'switches.0.switch.attention_norm.accumulator_index' not in state_dict.keys():
|
||||
for i in range(4):
|
||||
state_dict['switches.%i.switch.attention_norm.accumulator' % (i,)] = t_state['switches.%i.switch.attention_norm.accumulator' % (i,)]
|
||||
state_dict['switches.%i.switch.attention_norm.accumulator_index' % (i,)] = t_state['switches.%i.switch.attention_norm.accumulator_index' % (i,)]
|
||||
state_dict['switches.%i.switch.attention_norm.accumulator_filled' % (i,)] = t_state['switches.%i.switch.attention_norm.accumulator_filled' % (i,)]
|
||||
super(DualOutputSRG, self).load_state_dict(state_dict, strict)
|
||||
|
||||
|
||||
class Interpolate(nn.Module):
|
||||
def __init__(self, factor):
|
||||
super(Interpolate, self).__init__()
|
||||
|
@ -323,17 +336,6 @@ class ConfigurableSwitchedResidualGenerator3(nn.Module):
|
|||
return val
|
||||
|
||||
|
||||
def load_state_dict(self, state_dict, strict=True):
|
||||
# Support backwards compatibility where accumulator_index and accumulator_filled are not in this state_dict
|
||||
t_state = self.state_dict()
|
||||
if 'switches.0.switch.attention_norm.accumulator_index' not in state_dict.keys():
|
||||
for i in range(4):
|
||||
state_dict['switches.%i.switch.attention_norm.accumulator' % (i,)] = t_state['switches.%i.switch.attention_norm.accumulator' % (i,)]
|
||||
state_dict['switches.%i.switch.attention_norm.accumulator_index' % (i,)] = t_state['switches.%i.switch.attention_norm.accumulator_index' % (i,)]
|
||||
state_dict['switches.%i.switch.attention_norm.accumulator_filled' % (i,)] = t_state['switches.%i.switch.attention_norm.accumulator_filled' % (i,)]
|
||||
super(DualOutputSRG, self).load_state_dict(state_dict, strict)
|
||||
|
||||
|
||||
class DualOutputSRG(nn.Module):
|
||||
def __init__(self, switch_depth, switch_filters, switch_reductions, switch_processing_layers, trans_counts, trans_kernel_sizes,
|
||||
trans_layers, transformation_filters, initial_temp=20, final_temperature_step=50000, heightened_temp_min=1,
|
||||
|
@ -398,12 +400,14 @@ class DualOutputSRG(nn.Module):
|
|||
|
||||
def update_for_step(self, step, experiments_path='.'):
|
||||
if self.attentions:
|
||||
temp = max(1, int(self.init_temperature * (self.final_temperature_step - step) / self.final_temperature_step))
|
||||
if temp == 1 and self.heightened_final_step and self.heightened_final_step != 1:
|
||||
temp = max(1,
|
||||
1 + self.init_temperature * (self.final_temperature_step - step) / self.final_temperature_step)
|
||||
if temp == 1 and self.heightened_final_step and step > self.final_temperature_step and \
|
||||
self.heightened_final_step != 1:
|
||||
# Once the temperature passes (1) it enters an inverted curve to match the linear curve from above.
|
||||
# without this, the attention specificity "spikes" incredibly fast in the last few iterations.
|
||||
h_steps_total = self.heightened_final_step - self.final_temperature_step
|
||||
h_steps_current = max(min(step - self.final_temperature_step, h_steps_total), 1)
|
||||
h_steps_current = min(step - self.final_temperature_step, h_steps_total)
|
||||
# The "gap" will represent the steps that need to be traveled as a linear function.
|
||||
h_gap = 1 / self.heightened_temp_min
|
||||
temp = h_gap * h_steps_current / h_steps_total
|
||||
|
@ -411,7 +415,7 @@ class DualOutputSRG(nn.Module):
|
|||
temp = 1 / temp
|
||||
self.set_temperature(temp)
|
||||
if step % 50 == 0:
|
||||
[save_attention_to_image(experiments_path, self.attentions[i], self.transformation_counts, step, "a%i" % (i+1,)) for i in range(len(self.switches))]
|
||||
save_attention_to_image(experiments_path, self.attentions[0], self.transformation_counts, step, "a%i" % (1,), l_mult=10)
|
||||
|
||||
def get_debug_values(self, step):
|
||||
temp = self.switches[0].switch.temperature
|
||||
|
|
Loading…
Reference in New Issue
Block a user