xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2025-12-10 05:55:55 +08:00
Code Issues Packages Projects Releases Wiki Activity

[TPU] Fix greedy decoding (#6933)

Browse Source
This commit is contained in:
Woosuk Kwon 2024-07-30 02:06:29 -07:00 committed by GitHub
parent af647fb8b3
commit 6e063ea35b
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 18 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

27
vllm/worker/tpu_model_runner.py
View File

@ -28,7 +28,9 @@ if TYPE_CHECKING:
logger = init_logger(__name__) logger = init_logger(__name__)
_PAD_SLOT_ID = -1 # NOTE(woosuk): In PyTorch XLA, index -1 is ignored. # Here we utilize the behavior that out-of-bound index is ignored.
# FIXME(woosuk): Find a more reliable way to prevent possible bugs.
_PAD_SLOT_ID = 1_000_000_000
# FIXME(woosuk): Temporarily disabled top-p sampling since it's too slow. # FIXME(woosuk): Temporarily disabled top-p sampling since it's too slow.
_ENABLE_TOP_P = False _ENABLE_TOP_P = False
# FIXME(woosuk): A temporary hack to support `n > 1`. # FIXME(woosuk): A temporary hack to support `n > 1`.
@ -414,10 +416,7 @@ class TPUModelRunner(ModelRunnerBase[ModelInputForTPU]):
best_of = [] best_of = []
for seq_group_metadata in seq_group_metadata_list: for seq_group_metadata in seq_group_metadata_list:
sampling_params = seq_group_metadata.sampling_params sampling_params = seq_group_metadata.sampling_params
# NOTE(woosuk): Here we mimic argmax sampling by applying a very t.append(sampling_params.temperature)
# low temperature. This is not accurate.
t.append(sampling_params.temperature
if sampling_params.temperature >= 1e-5 else 1e-5)
if sampling_params.top_p != 1 and not _ENABLE_TOP_P: if sampling_params.top_p != 1 and not _ENABLE_TOP_P:
raise NotImplementedError( raise NotImplementedError(
"Top-p sampling is currently disabled for the TPU backend " "Top-p sampling is currently disabled for the TPU backend "
@ -678,13 +677,23 @@ class ModelWrapper(nn.Module):
hidden_states = hidden_states.flatten(0, 1) hidden_states = hidden_states.flatten(0, 1)
logits = self.model.compute_logits(hidden_states, sampling_metadata) logits = self.model.compute_logits(hidden_states, sampling_metadata)
logits = logits / t.unsqueeze(dim=1) # Argmax sampling.
argmax_token_ids = torch.argmax(logits, dim=-1, keepdim=True)
argmax_token_ids = argmax_token_ids.repeat(1, num_samples)
# Zero temperature means greedy decoding. Avoid division by zero.
nonzero_t = torch.where(t != 0, t, 1.0)
logits = logits / nonzero_t.unsqueeze(dim=1)
if _ENABLE_TOP_P: if _ENABLE_TOP_P:
logits = _apply_top_p(logits, p.unsqueeze(dim=1)) logits = _apply_top_p(logits, p.unsqueeze(dim=1))
# Random sampling.
probs = torch.softmax(logits, dim=-1, dtype=torch.float32) probs = torch.softmax(logits, dim=-1, dtype=torch.float32)
next_token_ids = torch.multinomial(probs, sampled_token_ids = torch.multinomial(probs,
num_samples, num_samples,
replacement=True) replacement=True)
next_token_ids = torch.where(t != 0, sampled_token_ids,
argmax_token_ids)
return next_token_ids return next_token_ids