Search results

The version of transformers

Status: Open.

#57 In pkunlp-icler/FastV;

· jiuyixu25 opened on Jan 26, 2026

开源qwen的Demo

Status: Open.

#56 In pkunlp-icler/FastV;

· xiaonengmiao opened on Nov 18, 2025

AttributeError

Status: Open.

#55 In pkunlp-icler/FastV;

· nagexiaochengzi opened on Sep 11, 2025

qwen-kv

Status: Open.

#54 In pkunlp-icler/FastV;

· xyfore opened on Jul 15, 2025

SYS_LENGTH值怎么确定

Status: Open.

#53 In pkunlp-icler/FastV;

· xyfore opened on Jul 7, 2025

Could you please clarify whether the results reported for FastV were obtained with use_cache set to False and using the in-place computation strategy?

Status: Open.

#52 In pkunlp-icler/FastV;

· 924973292 opened on Jul 1, 2025

How to reproduce the latency and GPU memory results?

Status: Open.

#50 In pkunlp-icler/FastV;

· surendrapathak01 opened on Jun 24, 2025

Potential Mistake in kvcache implementation--Why always use the same-layer kvcache?

Status: Open.

#49 In pkunlp-icler/FastV;

· lxt160980 opened on May 11, 2025

Criteria for selecting tokens

Status: Open.

#48 In pkunlp-icler/FastV;

· EvelynZhang-epiclab opened on Apr 21, 2025

Problem with lmms-eval: Code modification instructions don't match my conda env file

Status: Open.

#47 In pkunlp-icler/FastV;

· zychen8 opened on Apr 10, 2025

Is there a bug in the <code>fastv_kvcache.py</code> ?

Status: Open.

#46 In pkunlp-icler/FastV;

· Fanziyang-v opened on Apr 7, 2025

Setup question

Status: Open.

#45 In pkunlp-icler/FastV;

· Schen-sony opened on Mar 17, 2025