huggingface/text-generation-inference
5
0
Fork 0
mirror of https://github.com/huggingface/text-generation-inference.git synced 2025-09-11 20:34:54 +00:00
Code Issues Packages Projects Releases Wiki Activity

hotfix for quantization

Browse Source
This commit is contained in:
fxmarty 2024-05-17 17:18:40 +00:00
parent f82ae76dff
commit 7a5f5d9757
2 changed files with 8 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
server/text_generation_server/models/custom_modeling/flash_llama_modeling.py
View File

@ -230,11 +230,15 @@ class LlamaMLP(nn.Module):
config.intermediate_size // weights.process_group.size()
)
# TODO: This is a hotfix to be removed & properly refactored.
self.quantize = config.quantize
def forward(self, hidden_states):
if (
SYSTEM == "rocm"
and self.hidden_act == "silu"
and hidden_states.shape[0] == 1
and not self.quantize
):
out = torch.empty(
hidden_states.shape[0],

4
server/text_generation_server/models/custom_modeling/flash_mistral_modeling.py
View File

@ -290,11 +290,15 @@ class MistralMLP(nn.Module):
config.intermediate_size // weights.process_group.size()
)
# TODO: This is a hotfix to be removed & properly refactored.
self.quantize = config.quantize
def forward(self, hidden_states):
if (
SYSTEM == "rocm"
and self.hidden_act == "silu"
and hidden_states.shape[0] == 1
and not self.quantize
):
out = torch.empty(
hidden_states.shape[0],