huggingface/text-generation-inference
5
0
Fork 0
mirror of https://github.com/huggingface/text-generation-inference.git synced 2025-06-19 15:52:08 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix multi-modality concatenate

Browse Source 
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
This commit is contained in:
Wang, Yi A 2025-06-05 23:14:15 -07:00
parent acc02aeb3e
commit dbb24255c3
2 changed files with 4 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
backends/gaudi/server/text_generation_server/models/flash_vlm_causal_lm.py
View File

@ -262,8 +262,8 @@ class FlashVlmCausalLMBatch(FlashCausalLMBatch):
@classmethod @classmethod
@tracer.start_as_current_span("concatenate") @tracer.start_as_current_span("concatenate")
def concatenate(cls, batches): def concatenate(cls, batches, padded_total_bs: int = 0):
batch = super(FlashVlmCausalLMBatch, cls).concatenate(batches) batch = super(FlashVlmCausalLMBatch, cls).concatenate(batches, padded_total_bs)
batch.pixel_values = None batch.pixel_values = None
batch.pixel_attention_mask = None batch.pixel_attention_mask = None
batch.image_sizes = None batch.image_sizes = None

4
backends/gaudi/server/text_generation_server/models/mllama_causal_lm.py
View File

@ -48,8 +48,8 @@ class FlashMllamaCausalLMBatch(FlashVlmCausalLMBatch):
@classmethod @classmethod
@tracer.start_as_current_span("concatenate") @tracer.start_as_current_span("concatenate")
def concatenate(cls, batches): def concatenate(cls, batches, padded_total_bs: int = 0):
batch = super().concatenate(batches) batch = super().concatenate(batches, padded_total_bs)
batch.pixel_values = None batch.pixel_values = None
batch.pixel_attention_mask = None batch.pixel_attention_mask = None