4 Pull requests merged by 1 person

• FIX: Failing single GPU tests related to hotswapping

  #2385 merged Feb 19, 2025

• SEC Bump transformers version used in examples

  #2374 merged Feb 19, 2025

• CI Skip audio test on single GPU CI

  #2380 merged Feb 18, 2025

• FIX: Avoid caching in X-LoRA generate

  #2384 merged Feb 18, 2025

4 Pull requests opened by 2 people

• ENH Allow rank/alpha keys to be "fully qualified"

  #2382 opened Feb 17, 2025

• orthogonal lora layer init

  #2389 opened Feb 20, 2025

• FIX Model with nested all-linear target modules

  #2391 opened Feb 20, 2025

• ENH Make hotswap error on compile optional

  #2393 opened Feb 21, 2025

3 Issues opened by 2 people

• Bug: Using 2 LoRA configs with `target_modules='all-linear'` leads to nested LoRA layers

  #2390 opened Feb 20, 2025

• ValueError: Target module Qwen2_5_VisionTransformerPretrainedModel is not supported.

  #2388 opened Feb 19, 2025

• Bug when deleting adapters of a model with modules_to_save

  #2381 opened Feb 17, 2025

7 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• Standalone Custom Tokens Tuner and integrated into LoRA

  #2376 commented on Feb 21, 2025 • 39 new comments

• Implementation of adapter?

  #244 commented on Feb 15, 2025 • 0 new comments

• Comparison of Different Fine-Tuning Techniques for Conversational AI

  #2310 commented on Feb 17, 2025 • 0 new comments

• Prefix Tuning dimension error with Qwen2 and missing vocab_size for PaliGemma2

  #2315 commented on Feb 17, 2025 • 0 new comments

• prompt_tuning_peft tutorial raises cache layer error

  #2379 commented on Feb 19, 2025 • 0 new comments

• Peft version upgrade from 0.4.0 to 0.14.0 results in "No module named \u0027peft.utils.config\u0027" error

  #2339 commented on Feb 21, 2025 • 0 new comments

• [FEAT] Add support for optimum-quanto

  #2000 commented on Feb 20, 2025 • 0 new comments