Skip to content

Output each char individually, avoid cost of appending to static string #138

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 2 commits into
base: master
Choose a base branch
from
Open

Output each char individually, avoid cost of appending to static string #138

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
e6279ee
Output each character individually to avoid the cost of appending to …
joelself Aug 20, 2016
cf9d80c
Write the start text to output
joelself Aug 28, 2016
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
20 changes: 6 additions & 14 deletions LanguageModel.lua
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -132,16 +132,8 @@ function LM:encode_string(s)
return encoded
end


function LM:decode_string(encoded)
assert(torch.isTensor(encoded) and encoded:dim() == 1)
local s = ''
for i = 1, encoded:size(1) do
local idx = encoded[i]
local token = self.idx_to_token[idx]
s = s .. token
end
return s
function LM:decode_char(encoded)
return self.idx_to_token[encoded[1][1]]
end


Expand All @@ -162,8 +154,8 @@ function LM:sample(kwargs)
local verbose = utils.get_kwarg(kwargs, 'verbose', 0)
local sample = utils.get_kwarg(kwargs, 'sample', 1)
local temperature = utils.get_kwarg(kwargs, 'temperature', 1)
local write = io.write

local sampled = torch.LongTensor(1, T)
self:resetStates()

local scores, first_t
Expand All @@ -173,7 +165,7 @@ function LM:sample(kwargs)
end
local x = self:encode_string(start_text):view(1, -1)
local T0 = x:size(2)
sampled[{{}, {1, T0}}]:copy(x)
write(start_text)
scores = self:forward(x)[{{}, {T0, T0}}]
first_t = T0 + 1
else
Expand All @@ -195,12 +187,12 @@ function LM:sample(kwargs)
probs:div(torch.sum(probs))
next_char = torch.multinomial(probs, 1):view(1, 1)
end
sampled[{{}, {t, t}}]:copy(next_char)
write(self:decode_char(next_char))
scores = self:forward(next_char)
end

self:resetStates()
return self:decode_string(sampled[1])
write('\n')
end


Expand Down
3 changes: 1 addition & 2 deletions sample.lua
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -38,5 +38,4 @@ if opt.verbose == 1 then print(msg) end

model:evaluate()

local sample = model:sample(opt)
print(sample)
model:sample(opt)