mirror of
https://github.com/huggingface/text-generation-inference.git
synced 2025-06-11 03:42:08 +00:00
connecting video to qwen2
This commit is contained in:
parent
5ced960f6e
commit
05464d26bf
@ -212,6 +212,9 @@ class VlmCausalLMBatch(FlashCausalLMBatch):
|
|||||||
processor, image_inputs, config, image_id
|
processor, image_inputs, config, image_id
|
||||||
)
|
)
|
||||||
image_id += 1
|
image_id += 1
|
||||||
|
elif chunk_type == "video" and config.model_type == "qwen2_vl":
|
||||||
|
# Based on Qwen2VL's video token format
|
||||||
|
full_text += f"<video>{chunk.video}</video>"
|
||||||
|
|
||||||
full_text = image_text_replacement_fixup(config, full_text)
|
full_text = image_text_replacement_fixup(config, full_text)
|
||||||
|
|
||||||
|
Loading…
Reference in New Issue
Block a user