Search results

How do I set different compression versions?

Status: Open.

#29 In Cooperx521/PyramidDrop;

· liuyafeng0609 opened on Nov 10, 2025

Settings of Table 2

Status: Open.

#28 In Cooperx521/PyramidDrop;

· Wuzimeng opened on Oct 20, 2025

the "layer_list" and "image_token_ratio_list" in the Table 1

Status: Open.

#27 In Cooperx521/PyramidDrop;

· Maxlium opened on Sep 15, 2025

Have anyone tried evaluate PDrop under lmms-eval framework?

Status: Open.

#26 In Cooperx521/PyramidDrop;

· ghost opened on Jun 11, 2025

What does system token and instruction token mean?

Status: Open.

#25 In Cooperx521/PyramidDrop;

· jacksonmygo opened on May 19, 2025

Question about Last Token

Status: Open.

#23 In Cooperx521/PyramidDrop;

· HarrisonWu42 opened on May 8, 2025

Question about decoding stage.

Status: Open.

#21 In Cooperx521/PyramidDrop;

· chenghuaWang opened on Apr 27, 2025

During the model inference process, does LLaVA-NeXT-7B use the llava-v1.6-vicuna-7b model or the open-llava-next-vicuna-7b model?

Status: Open.

#18 In Cooperx521/PyramidDrop;

· lyl121 opened on Apr 7, 2025

Does pyramidDrop support multiple images?

Status: Open.

#16 In Cooperx521/PyramidDrop;

· Beniko95J opened on Mar 31, 2025

attention map visualization

Status: Open.

#14 In Cooperx521/PyramidDrop;

· qq1332427275 opened on Mar 21, 2025

[Author] Please place any models that need to be compatible with pdrop here.

Status: Open.

#12 In Cooperx521/PyramidDrop;

· Cooperx521 opened on Mar 6, 2025

Setting image_token_ratio_list for Specific Compression Rates

Status: Open.

#10 In Cooperx521/PyramidDrop;

· ShawnTan86 opened on Mar 1, 2025