huggingface/text-generation-inference
5
0
Fork 0
mirror of https://github.com/huggingface/text-generation-inference.git synced 2025-09-10 20:04:52 +00:00
Code Issues Packages Projects Releases Wiki Activity

Apply suggestions from code review

Browse Source 
Co-authored-by: OlivierDehaene <olivier@huggingface.co>
This commit is contained in:
Nicolas Patry 2023-12-11 12:06:35 +01:00 committed by GitHub
parent b6519b5279
commit 0006fab5ab
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 0 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
server/text_generation_server/models/flash_causal_lm.py
View File

@ -422,8 +422,6 @@ class FlashCausalLMBatch(Batch):
block_tables_tensor = self.block_tables_tensor[indices]
input_lengths_tensor = self.input_lengths_tensor[indices]
slots = self.slots[slot_filtering_indices]
if slot_indices.max().item() > slots.shape[0]:
import ipdb;ipdb.set_trace()
next_token_chooser = self.next_token_chooser.filter(indices)
top_n_tokens_tensor = self.top_n_tokens_tensor[indices]
speculative_ids = self.speculative_ids[indices] if self.speculative_ids is not None else None

3
server/text_generation_server/utils/layers.py
View File

@ -742,9 +742,6 @@ try:
self._update_cos_sin_cache(dtype, position_ids.device, max_s)
if position_ids.max().item() >= max_s:
import ipdb;ipdb.set_trace()
cos = torch.index_select(self._cos_cached, 0, position_ids)
sin = torch.index_select(self._sin_cached, 0, position_ids)
# Note: this unsqueeze is not necessary on RoCm + VLLM ROPE implementation, but we leave it as is to avoid yet an other controlflow.