Search results

evaluate(): DDP unwrap is nested under 'val' branch; zero_shot_eval still wrapped

Status: Open.

#1152 In mlfoundations/open_clip;

· shaun0927 opened on Apr 17, 2026

text_global_pool(pool_type='eos'): silent fallback to index 0 on missing EOS

Status: Open.

#1151 In mlfoundations/open_clip;

· shaun0927 opened on Apr 17, 2026

pt_load: training --resume path loads pickle without weights_only

Status: Open.

#1150 In mlfoundations/open_clip;

· shaun0927 opened on Apr 17, 2026

openai.py: torch.load fallback missing weights_only=True

Status: Open.

#1149 In mlfoundations/open_clip;

· shaun0927 opened on Apr 17, 2026

factory.load_state_dict: dead except TypeError fallback drops weights_only

Status: Open.

#1148 In mlfoundations/open_clip;

· shaun0927 opened on Apr 17, 2026

Stealthy two-repository attack when loading remote tokenizer on HuggingFace

Status: Open.

#1146 In mlfoundations/open_clip;

· Vancir opened on Apr 10, 2026

[Bug/Question] Unexpected behavior and redundant computation with gather_with_grad and local_loss flag combinations

Status: Open.

#1144 In mlfoundations/open_clip;

· rubenfigge opened on Mar 31, 2026

Consider not using the root logger

Status: Open.

#1138 In mlfoundations/open_clip;

· spezialspezial opened on Feb 20, 2026

Request for SigLIP2 Training Code and Pretrained Weights

Status: Open.

#1041 In mlfoundations/open_clip;

· eitamarSaraf opened on Mar 4, 2025

error when use try to use int8 operations with OpenCLIP.

Status: Open.

#1012 In mlfoundations/open_clip;

· xiaohoua opened on Dec 26, 2024

Apply scale loss when performing accum_freq

Status: Open.

#957 In mlfoundations/open_clip;

· AshStuff opened on Oct 14, 2024

SigLip memory consumption increases as we scale number of GPUs

Status: Open.

#942 In mlfoundations/open_clip;

· khalidsaifullaah opened on Sep 26, 2024