Skip to content
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

Enable generation tests #407

Merged
merged 10 commits into from
Jun 8, 2023
Merged

Enable generation tests #407

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
10 commits
Select commit Hold shift + click to select a range
ddde283
WIP: enable generation tests
kundaMwiza Jun 2, 2023
dbebc55
Refactor generation tests based on transformers 4.29.2
kundaMwiza Jun 5, 2023
121204d
Clean generation tests
kundaMwiza Jun 6, 2023
990e52a
Generation utils: support multiple eos tokens
kundaMwiza Jun 6, 2023
41c4cd0
Check style with ruff
kundaMwiza Jun 6, 2023
3e3f5ea
Run ruff formatter
kundaMwiza Jun 6, 2023
f462bb7
Ensure each test detaches from IPUs after execution
kundaMwiza Jun 6, 2023
3c33d46
Fix paralellization when using KV cache
kundaMwiza Jun 8, 2023
f1ae2ce
make stle and isort
kundaMwiza Jun 8, 2023
a3e124d
make style, one more time
kundaMwiza Jun 8, 2023
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
26 changes: 22 additions & 4 deletions optimum/graphcore/generation/utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -283,6 +283,14 @@ def detachFromDevice(self):
if hasattr(self, "poptorch_decoder"):
self.poptorch_decoder.detachFromDevice()

def destroy(self):
if hasattr(self, "poptorch_encoder"):
self.poptorch_encoder.destroy()
delattr(self, "poptorch_encoder")
if hasattr(self, "poptorch_decoder"):
self.poptorch_decoder.destroy()
delattr(self, "poptorch_decoder")

def _get_generation_step_tensor(self, generation_step, ascending=False):
# Returns a 1 dimensional tensor of the form [device_iterations * replication factor]
# with all elements equal to generation_step.
Expand Down Expand Up @@ -433,6 +441,9 @@ def greedy_search(
max_length = stopping_criteria.max_length
pad_token_id = pad_token_id if pad_token_id is not None else self.config.pad_token_id
eos_token_id = eos_token_id if eos_token_id is not None else self.config.eos_token_id
if isinstance(eos_token_id, int):
eos_token_id = [eos_token_id]
eos_token_id_tensor = torch.tensor(eos_token_id).to(input_ids.device) if eos_token_id is not None else None
output_scores = output_scores if output_scores is not None else self.config.output_scores
output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
output_hidden_states = (
Expand Down Expand Up @@ -547,8 +558,10 @@ def greedy_search(
cur_len = cur_len + 1

# if eos_token was found in one sentence, set sentence to finished
if eos_token_id is not None:
unfinished_sequences = unfinished_sequences.mul((next_tokens != eos_token_id).long())
if eos_token_id_tensor is not None:
unfinished_sequences = unfinished_sequences.mul(
next_tokens.tile(eos_token_id_tensor.shape[0], 1).ne(eos_token_id_tensor.unsqueeze(1)).prod(dim=0)
)

# stop when each sentence is finished, or if we exceed the maximum length
if unfinished_sequences.max() == 0 or stopping_criteria(input_ids, scores):
Expand Down Expand Up @@ -1030,6 +1043,9 @@ def sample(
logits_warper = logits_warper if logits_warper is not None else LogitsProcessorList()
pad_token_id = pad_token_id if pad_token_id is not None else self.config.pad_token_id
eos_token_id = eos_token_id if eos_token_id is not None else self.config.eos_token_id
if isinstance(eos_token_id, int):
eos_token_id = [eos_token_id]
eos_token_id_tensor = torch.tensor(eos_token_id).to(input_ids.device) if eos_token_id is not None else None
output_scores = output_scores if output_scores is not None else self.config.output_scores
output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
output_hidden_states = (
Expand Down Expand Up @@ -1139,8 +1155,10 @@ def sample(
cur_len = cur_len + 1

# if eos_token was found in one sentence, set sentence to finished
if eos_token_id is not None:
unfinished_sequences = unfinished_sequences.mul((next_tokens != eos_token_id).long())
if eos_token_id_tensor is not None:
unfinished_sequences = unfinished_sequences.mul(
next_tokens.tile(eos_token_id_tensor.shape[0], 1).ne(eos_token_id_tensor.unsqueeze(1)).prod(dim=0)
)

# stop when each sentence is finished, or if we exceed the maximum length
if unfinished_sequences.max() == 0 or stopping_criteria(input_ids, scores):
Expand Down
Loading